P Parametric Sample Selection Models

Document technical information

Format pdf
Size 7.9 MB
First found Jun 9, 2017

Document content analysis

Language
English
Type
not defined
Concepts
no text concepts found

Persons

Allen Lane
Allen Lane

wikipedia, lookup

David L. Bazelon
David L. Bazelon

wikipedia, lookup

Organizations

Places

Transcript

P
Parametric Sample Selection Models regression and are distinguished thereby from
William Rhodes
Abt Associates Inc., Cambridge, MA, USA
Overview
James Heckman’s work (1976, 1979) regarding
the consequences and solutions for sample selection bias has influenced three decades of social
science research. For this and related contributions, he won a Nobel Prize in economics.
Heckman’s original solution requires strong
parametric assumptions. When wrong, those
assumptions can lead to parameter estimates
with mean-squared errors worse than those of
estimators that ignore selection bias.
Nevertheless, the parametric approach
remains popular. Discussions appear in econometrics textbooks (Davidson and MacKinnon
1993, pp. 542–545; Cameron and Trivedi 2005,
pp. 546–569; Greene 2008, pp. 863–903) and
technical papers; estimation routines appear in
standard econometrics software; and the
approach is frequently used in criminal justice
research (Bushway et al. 2007). Criminologist
and criminal justice researchers cannot disregard
the parametric approach.
All estimators require some parametric
assumptions, so what distinguishes estimators
that are parametric from another class? Estimators considered in this entry require strong distributional assumptions about error terms in a
estimators that impose weaker assumptions.
Without being comprehensive, this entry identifies six cases where selection bias occurs in
criminal justice research:
• Sample selection due to unobserved variables
• Estimating treatment effects using observational data
• Breakdowns in randomized controlled trials
• Survival models with censoring and competing events
• Endogenous stratification/choice-based sampling
• Missing data
This entry shows how parametric assumptions
can overcome bias but also how the approach
can fail. It illustrates points with Monte Carlo
experiments.
Theoretical Background
To motivate a discussion, consider a data generation process (DGP) comprising two equations.
The first is the outcome equation: It explains
the outcome of interest in terms of explanatory
variables. The objective is to estimate one or
more parameters of this outcome equation.
The second is the selection equation: It explains
how data get sampled for analysis. Selection bias
may occur when the same factors affect both
sampling and outcome (Rosenbaum 2002).
After introducing notation, this entry explains
when selection bias occurs and how parametric
models correct for bias.
G. Bruinsma, D. Weisburd (eds.), Encyclopedia of Criminology and Criminal Justice,
DOI 10.1007/978-1-4614-5690-2, # Springer Science+Business Media New York 2014
P
3430
When writing the two equations, use the
notation:
Y Denotes the outcome measure: the dependent variable in the first equation.
A subscript denoting the ith observation is
implicit.
S
Denotes the selection measure: the dependent variable in the second equation. Its
interpretation will differ across the six
cases.
X This is an observed row vector of variables
that affect both Y and S. Observed means
that the researcher has measured these variables and uses them in the analysis.
X typically includes a constant.
W This is an unobserved row vector of variables that affect both Y and S. The
researcher has not measured these variables
or else omits them from the analysis.
Assume that unobserved variables are distributed independent of other explanatory
variables appearing in the outcome equation; consequences of assuming otherwise
are discussed subsequently.
V This is a vector of observed variables that
affect Y but not S.
Z
This is a vector of variables that affect S but
not Y. When observed, the vector comprises
instrumental variables used to impose identification restrictions.
b
This represents a conformable column
vector of parameters appearing in the outcome equation. Subscripts will denote
whether these pertain to X, W, or V. Conformable means the column vector b has as
many elements as the row vectors X, W,
and V.
a
This represents a conformable column vector of parameters appearing in the selection
equation. Subscript will denote whether
these parameters pertain to X, W, or Z.
e
This represents an error term pertaining to
the first equation.
u
This represents an error term pertaining to
the second equation.
For simplicity, assume the error terms are
mean-zero, identically and independently
Parametric Sample Selection Models
distributed within each equation, independently
distributed across equations, and independent of
X, W, V, and Z. Assuming otherwise would add
complications without enhancing insight.
Given this notation and limiting the immediate
discussion to linear models, a general data generation process (DGP) is written:
Y ¼ XbX þ WbW þ VbV þ e
(1)
S ¼ XaX þ WaW þ ZaZ þ u
(2)
A DGP is the actual stochastic process generating the data. The researcher necessarily omits
unobserved variables from the analysis, so
a model of the DGP is expressed as:
Y ¼ XbX þ VbV þ e
(3)
S ¼ XaX þ ZaZ þ u
(4)
The error terms change to e* and u*:
e ¼ e þ WbW
u ¼ u þ WaW
(5)
Some additional notation will be useful:
This is the variance of e*.
This is the variance of u*.
se u This is the covariance of e* and u*.
Although e is independent of X in the population, the following section will show that e* is not
independent of X in the sample used for estimation. Consequently, a regression based on model
(3) will violate a fundamental assumption: errors
must be independent of explanatory variables.
Standard regression procedures yield biased and
inconsistent estimates of bX.
Is the DGP encapsulated by Eqs. 1 and 2
plausible or just abstract mumbo-jumbo? In fact
behavioral models often translate into DGPs with
counterparts to Eqs. 1 and 2. Based on perceptions of self-interest, individuals elect to participate in activities (Eq. 2) studied by researchers
(Eq. 1). Alternatively, based on perceptions of
community interests, agents (prosecutors, judges,
probation officers, etc.) perform selection to
s2e
s2u
Parametric Sample Selection Models
3431
enhance community safety. Thus, the DGP and
statistical models cover situations common to
criminology/criminal justice.
Correctly specified parametric models eliminate selection bias. One approach uses maximum
likelihood or partial likelihood to estimate
parameters of the joint distribution of e* and u*
in the sample. The joint distribution is written as
fb ðe ; u jsampleÞ
(6)
The a, b, and s parameters are implicit arguments in this distribution. The notation indicates
that the distribution depends on the sampling
procedure, and different sampling procedures
lead to various distributions (Maddala 1983),
only a few of which are discussed in this entry.
A second approach is to estimate the expected
value of e* in the sample and introduce this
expected value into the model. Equation 3 now
becomes
Y ¼ XbX þ VbV þ BE½e jselection þ v
(7)
The new error term v is mean-zero in expectation but heteroscedastic, so computed standard
errors require adjustment (Cameron and Trivedi
2005, p. 567; Greene 2008, p. 888). Modern software applies these corrections.
Some parametric assumptions are required.
Although other distributions might be used, typically analysts assume that e* and u* are distributed as bivariate normal, perhaps after
transformations to increase model flexibility and
perhaps after converting the s from scalars to
functions of explanatory variables. Inferences
are asymptotic, meaning that valid inferences
require large samples.
Econometricians have introduced semiparametric
approaches
for
estimating
E½e jselection (Vella 1998). These approaches
are not commonly used; they do not appear in
standard computing software; and (Greene 2002,
pp. E23–2) for one is skeptical that semiparametric approach will reach mainstream
application. This entry does not discuss semiparametric approaches.
P
In practice Eqs. 6 and 7 will not produce the
same results. Results may be very different if the
distributional assumptions are wrong. Furthermore, Eq. 6 can be adapted for models that are
nonlinear in the outcome equation. This is
unclear for Eq. 7 (Cameron and Trivedi 2005,
p. 192). Many models are nonlinear, so Eq. 6
seems preferable, but Eq. 7 has heuristic and
practical value.
The form provided by Eq. 7 is especially useful for interpreting results. Let x represent
a single component of X and let bx represent the
corresponding parameter. If the researcher’s
interest is how x changes Y in the population,
then the answer is bx. But often the researcher’s
interest is how Y changes with x given a selection
mechanism that has behavioral interpretations.
Then the answer is
@Y
@E½e jselection
¼ bx þ B
@x
@x
(8)
The correct interpretation depends on a wellspecific research question (Heckman and Vytlacil
2007a, b).
The difficulty of applying these solutions
depends on an as yet unspecified assumption
about the selection process. When population
data are available for estimating Eq. 4 but only
selected data are available for estimating Eq. 3,
the data are censored. When only sampled data
are available for estimating both Eqs. 3 and 4, the
data are truncated. This entry deals exclusively
with censored data although similar concepts
apply to truncated data. Unfortunately there is
room for confusion, because the censoring problem is sometimes referenced as incidental truncation to mean that the data for estimating Eq. 3
are truncated by the process described by Eq. 2.
This entry is focused on selection rules that
bifurcate the population, but other selection processes occur. Especially in criminal justice applications, where justice administration acts as
a sieve, selection may involve a sequence of
decisions (Vella 1998, p. 153). Also in criminal
justice applications, the selection rules may put
individuals into ordered categories (low risk,
medium risk, high risk) or into nominal
P
P
3432
categories. There exist suitable econometric
models (Vella 1998), some of which have been
programmed into conventional computing software, but discussion of alternative selection rules
goes beyond the scope of this entry.
Six Cases Illustrating Selection Bias
Sample Selection Due to Unobserved
Variables
Assume that the selection equation determines
which data enter into estimation of the outcome
equation. Specifically, a data point is selected
when S is larger than a threshold value; the
threshold is 0 here. Elaborate the data generation
process:
S 0 implies selection into the sample used to
estimate the outcome equation.
S < 0 implies exclusion from the sample.
From Eq. 2, sample selection occurs when
S ¼ XaX þ WaW þ ZaZ þ u 0 or when
WaW XaX ZaZ u
When the linear combination XaX is very large
in the population, any value of the linear combination WaW will likely pass the test for inclusion
in the sample; when XaX is very small, only large
values of WaW will likely pass the test. This
induces a negative correlation between XaX and
WaW in the sample. The reason is that the linear
combination WbW is an element of e* in the model
represented by Eq. 3. Given their common elements, WaW and WbW are likely correlated, in
which case elements of X will be correlated with
the error term e* in the sample.
The same argument establishes that X will be
correlated with Z in the sample. This does not
matter, however, because Z is not part of the
outcomes equation. V appears in the outcomes
equation, but V does not affect e*. The entire
explanation for selection arises because the
unobserved W appears in both the outcome equation and the selection equation (Rosenbaum
2002). The classification of variables as X, W,
V, and Z depends on the model, an observation
important for model building (Pearl 2000;
Gerring 2012), but those implications are not
discussed here.
Parametric Sample Selection Models
What are the solutions? An obvious solution is
to assure that X comprises variables that affect
both the selection and outcome equations so there
are no W variables. That may be impossible and
the solution is uncertain because there is seldom
reason to justify that X is inclusive. A researcher
who relies on this reasoning will typically face
validity challenges. Nevertheless, this assumption underlies all estimation procedures that
assume selection on the observables.
In contrast, Heckman’s solution assumes selection on the unobserved W. As noted, estimation
requires assuming a joint distribution of e* and u*.
Assuming that e* and u* are bivariate normal is
problematic. Even if e and u are independent normal, bivariate normality of e* and u* depends on
WaW and WbW being distributed as normal in the
population. Although possible, it is a strong
^ will be inconsistent.
assumption, and if wrong, b
Criminologists and criminal justice researchers
have used the parametric model to deal with selection bias (Bushway et al. 2007). For example, in an
early application in criminal justice research, the
author used Heckman’s approach to estimate how
risk factors predict pretrial misconduct; the sample
selection mechanism entered because the worst
pretrial risks were detained pending trial so they
were omitted from the sample for estimating the
outcome equation. That application illustrates
a point: Eq. 4 can be estimated consistently
because the analyst observes release/detention
decisions for every offender. Selection bias affects
the estimation of Eq. 3 because the outcomes are
observed for only those offenders released pending trial, but the inference is about risk in the
population of offenders. The application motivates
two-step estimation: Estimate the pretrial release
decision first and then estimate the pretrial
misconduct equation.
Estimating Treatment Effects Using
Observational Data
In the second case, the selection equation
determines who enters treatment. An explanation
requires modest changes to Eq. 1 through
Eq. 4. Let
T is a dummy variable that denotes treatment
when T ¼ 1 and no treatment when T ¼ 0
Parametric Sample Selection Models
d a vector of parameters (the treatment effect),
which is the objective of estimation
Then
S 0 implies selection into treatment (T ¼ 1).
S < 0 implies no treatment (T ¼ 0).
Treatment potentially changes all the parameters in Eq. 1, so rewrite Eq. 1 as
Y ¼ Xb0X þ Wb0W þ Vb0V þ e if treated
Y ¼ XbX þ WbW þ VbV þ e otherwise
Equivalently the DGP can be rewritten as
a single equation with interactions
Y ¼XbX þ WbW þ VbV
þ ½XdX þ WdW þ VdV T þ e
The model becomes
Y ¼XbX þ VbV þ ½XdX þ VdV T
þ ½WbW þ WTdW þ e
¼ XbX þ VbV þ ½XdX þ VdV T þ e
Where
e ¼ WbW þ WTdW þ e
Often selection into treatment depends on
whether the person being treated his or her
agent perceives treatment to be beneficial. This
theory motivates specification of the selection
equation. It is common to assume that the d are
zero except for a constant—that is, XdX +
VdV ¼ d and WTdW ¼ 0. This assumption has
negative consequences if wrong. Nevertheless,
for simplicity, treat d as a scalar.
T is not orthogonal to e**. The reasoning is the
same as before. T is partly determined by W; also,
e** is correlated with W. Hence, T is correlated
with e**. An analyst could not estimate d consistently in a regression that fails to account for
selection.
Simple solutions follow when selection on
observables holds (Imbens and Wooldridge
2009). Otherwise, estimation rests on strong
parametric assumption about the joint distribution for e** and u**. As a practical matter,
parameter identification requires that Z be
observed and included in the model. Identification is discussed later. For example, in the
3433
P
criminal justice literature, the author and his or
her colleagues (Rhodes et al. 2001) used the
parametric approach to evaluate within prison
substance abuse treatment.
Breakdowns in Randomized Controlled Trials
RCT imposes selection by design. Unless the
design is perverse, random assignment typically
solves the selection bias problem. There are
three situations. First, selection is purely random: The selection mechanism contains no variables in common with the outcome equation.
Estimation is straightforward: the mean difference between the treated and control group is an
unbiased estimate of the average treatment
effect; a regression might be used to increase
efficiency or to learn about heterogeneous treatment effects.
Second, selection is random conditional on
X (exogenous sampling). Stratified sampling
obviously qualifies, as do most complex survey
designs. There is no sampling bias provided the
evaluator uses an estimator that accounts for X.
This might be done with weighting, but that topic
goes beyond the scope of this entry.
Third, selection is random conditional on W.
This perverse case happens. Consider a multisite
study where the researcher samples randomly
from within a site but does not select randomly
across sites – perhaps because many sites decline
to participate in the study. Or, consider the problem where some subjects assigned to treatment
refuse the treatment or some subjects assigned to
the control condition cross over to the treatment
condition, or both. Random assignment involves
selection bias in these common cases. Evaluators
sometimes use the terms internal and external
validity to describe this perverse case.
Thus, well-crafted and implemented RCTs are
free of selection bias. However, true RCTs are
difficult to design and implement, and selection
bias occurs when design/implementation is
faulty. Selection may be present in RCT (Berk
2005; Sampson 2010), and furthermore, the RCT
may be unable to answer important policy questions about the distribution of treatment effects
(Heckman and Vytlacil 2007a; Heckman et al.
1997). Meta-analysis (Borenstein et al. 2009)
P
P
3434
often integrates experimental and quasiexperimental studies to understand selection
bias and better estimate treatment effects.
Survival Models with Censoring
and Competing Events
Survival analysis has a long history in the study
of criminal recidivism. Advances have incorporated competing events into the study of criminal
recidivism. Selection bias can occur when estimating survival models for the reasons discussed
in 2.2 through 2.4, but survival models sometimes raise special selection problems.
Equations 1 and 2 will again be useful but now
recast those equations by assuming there are two
events, A and B. Event A might represent the
occurrence of criminal recidivism, defined as an
arrest for a new crime while under community
supervision. Event B might represent the occurrence of a revocation for a technical violation of
the conditions of supervision, treated here as
a competing event. Y and S represent the logarithm of the time until events A and B, respectively. The error term e is the distribution of
log-time until an arrest conditional on X, W, and
V; the error term u is the distribution of log-time
until a revocation conditional on X, W, and Z. The
analyst can observe the timing for whichever event
occurs first: A or B. Consequently, one can treat
B as “selecting” what is observable about A.
Almost always, analysts assume that the
timing of event A (the outcome of interest) is
conditionally independent of the timing of event
B (competing events). This is equivalent to
assuming selection on observables. If selection
on observables is wrong, traditional survival
analysis will result in biased parameter estimates.
Given censoring because of unobservables,
the solution is also familiar. The analyst needs
to model the joint distribution of e* and u*, but
accounting for the error structure is complicated
(van den Berg 2008). Moreover, commonly
available software lacks estimation routines.
Nevertheless, the maximum likelihood approach
introduced earlier provides a solution requiring
parametric assumptions; the author used this
approach to study criminal recidivism under
community supervision.
Parametric Sample Selection Models
To extend this observation, the reader might
note that survival models are simply censored
regressions. Therefore, the same selection problems that affect survival models also affect Tobit
models, which are frequently used in criminal
justice research.
Endogenous Stratification/Choice-Based
Sampling
Many samples depend on the occurrence of some
event. For example, data from the Arrestee Drug
Abuse Monitoring (ADAM) survey requires
a subject to be arrested and booked. Samples
used in the criminal careers literature require
a subject to acquire a criminal history. Such samples involve endogenous stratification or choicebased sampling.
To illustrate the problem, return to Eq. 1 and
think of Y as a count variable – zero arrests, one
arrest, two arrests, and so on occurring from
a Poisson process. The research question is to
estimate the arrest rate during a period of interest.
Return to Eq. 2 and think of S as being the
event: arrested and booked on the day that
ADAM arrestees were sampled. Selection bias
occurs because offenders with high crime commission rates have relatively high sampling probabilities. The sample will not represent the
general population of offenders.
Notice a subtle difference between endogenous stratification and the other selection problems: Selection into the sample depends on the
process determining the outcome. To illustrate,
assume a subject is sampled into the study
depending on the outcome, so:
Y > 0 implies selection into the sample
Y 0 implies no selection into the sample
The outcome equation is written as
Y ¼ XbX þ VbV þ e
The expected value for e*** in the sample is
E½e ¼ E½eje > XbX WbW VbV The error term e*** will be correlated with
X and V in the sample.
Endogenous stratification raises complex estimation issues (Cameron and Trivedi 2005,
pp. 822–829). Solutions may be best developed
for Poisson and negative binomial models
(Cameron and Trivedi 1998), which is fortunate
Parametric Sample Selection Models
3435
given the ubiquitous role of Poisson models in
criminal justice modeling. The criminal careers
literature has sometimes dealt with the issue;
early illustrations come from the Rand Corporation and others. The author has used a negative
binomial model with endogenous stratification to
estimate the arrest rate for drug-involved
offenders in the community based on a sample
of arrestees (Rhodes et al. 2007).
Missing Data
Analysts often deal with missing data (Schaefer
1997; Little and Rubin 2002). For example,
within a data set, sex may be missing for some
subjects, age may be missing for others, and both
age and sex may be missing for still others. The
variable is observed (it is an X variable), but its
value is missing for some observations.
One way to deal with missing data is to
exclude cases where any data are missing. This
is called list-wise or case-wise deletion. This
depletes the sample and thereby reduces power;
it can bias parameter estimation.
An alternative approach is to impute values for
the missing data, and imputation software is
available for common problems. Imputation and
estimation with imputed data is complicated.
Consider a simple problem: A single variable
x from the vector X has missing values. Assume
that x is the Kth element of X and that the other
K-1 elements are never missing. To focus attention on x, rewrite Eq. 1 equivalently as
Y ¼ X1:K1 bX1:K1 þ xbXK þ WbW þ VbV þ e
(9)
Interpret the selection equation as determining
when x is observed (S > 0) and missing (S < 0).
To emphasize the role of x, rescale so the aK ¼ 1
and rewrite Eq. 2 as
S ¼ X1:K1 aX1:K1 þ x þ WaW þ ZaW þ u
(10)
This implies that x is observed when
x X1:K1 a1:K1 ZaZ WaW u
(11)
P
As before, the motivation is to estimate the
parameters in the model:
Y ¼ X1:K1 bX1:K1 þ xK aK þ VbV þ e
(12)
Consider three cases. The first is when elements of a in Eq. 11 are 0 except perhaps for
the constant. Selection, which depends only on u,
is called missing completely at random. Missing
data will not bias the parameter estimates for
Eq. 12 because the error term e* will not be
correlated with X, x, or V. The second situation
is when aW is zero but other a parameters are not
zero. Selection, which now depends on one or
more of the observed variables, is called missing
at random. Missing data will not bias parameter
estimates because the error term e* will be
uncorrelated with X, x, and V. The third situation
is where aW 6¼ 0. Selection, which now depends
on unobservable variables, is called missing not
completely at random. This leads to biased
parameter estimates because the missing data
will be correlated with the error term e*.
There are parametric solutions to the first two
cases and approximate solutions to the third case.
The solution requires specifying x as a function of
observables. A specification might be
x ¼ YBY þ X1:K1 BX1:K1 þ VBV þ ZBZ þ RBR þ v
(13)
There are some surprises in Eq. 13. One is that
the outcome variable Y appears in the specification. The intuition is that x predicts Y, so
Y predicts x, and predicting x is the purpose of
Eq. 13. A second surprise is inclusion of a vector
of variables R that did not appear elsewhere in the
DGP. Again, the instrumental objective is to predict x, and it seems plausible that some set of
variables R might predict x yet not explain Y.
Finally, v is an error term whose distribution is
known except for some unspecified parameters.
The solution is parametric because of the linear
specification and the assumed distribution for v.
Estimation of Eq. 13, including the parameters of
the distribution of v, leads to a predicted (posterior) distribution for missing values for x.
P
P
3436
Estimation of Eq. 9 proceeds by sampling values
of x from this posterior distribution and using
those sampled values to estimate the parameter
appearing in Eq. 9. The approach is conceptually
simple but estimation rests on the specified distributional assumptions, is computer intensive,
and requires specialized software.
In the first two cases—missing completely at
random and missing at random—Bayesian logic
leads to consistent estimates for the parameters in
Eq. 9. This is not true in the third case. There the
analyst has two choices. The first is to construct
an imputation model that is explicit about the
selection mechanism; this is not frequently
done. The second is to use imputations from
Eq. 13 with an expectation that these will reduce
mean-squared errors in the b. This is plausible
because the imputations are based on additional
information that would otherwise be ignored by
case-wise deletion.
The single imputation problem is simple compared with a multiple imputation problem where
all of the X and V variables may have missing
values. In that case, the starting assumption is that
X and V are distributed as multivariate normal
perhaps after transformations and after conditioning on other variables such as Y and R. This
solution is highly parametric and probably
unrealistic. Nevertheless, experience shows that
estimation is insensitive to distributional assumptions although one would be dubious of applying
this approach to data that have a high proportion
of missing to known values.
As an illustration, the ADAM survey randomly samples arrestees, questions them about
drug use, and requests a urine specimen to test for
recent drug use. Not all arrestees are able or
willing to comply with the request, leading to
missing results for urine tests. ADAM analysts
have developed procedures used to impute urine
test results conditional on self-reports about
recent drug use.
Identification
The first five cases share a common problem:
The arguments assume that W is orthogonal to
X and V. This assumption was convenient
because it meant that in the absence of any
Parametric Sample Selection Models
selection bias, an analyst could consistently
estimate the b parameters.
What if the distribution of W is not orthogonal
to X and V? Then when W is omitted from the
outcome equation, the b parameters would not be
identified even absent selection bias. The problem is omitted variable bias, which is discussed in
introductory textbooks and is widely appreciated
by researchers. One can nevertheless think of
Eq. 3 as estimating a conditional mean for the
population (Angrist and Pischke 2009). Provided
the analyst considers his or her estimates to pertain to a conditional mean function, the argument
regarding selection bias does not require any
assumptions about independence between
W and either X or V, so one can focus on selection bias as the problem.
Putting this discussion about orthogonality
aside, the parametric solution to selection bias
requires an assumption about the joint distribution between e* and u*. Suppose the assumption
is correct. Is this sufficient to identify the b? The
answer is a qualified yes, but asymptotic justification for the parameter estimates and their sampling distributions may require very large
samples.
Equation 7 provides some intuition. Absent Z,
E[e*|sample] will be nearly collinear with X. The
relationship will not be perfectly collinear because
the nature of the joint distribution fb will induce
nonlinearities, but the departure from collinearity
will be slight in most real-world applications. Consequently, the standard errors for bX would be
large and inference tenuous. Few analysts would
feel comfortable relying on identification resting
on nonlinearities induced by correct specification
of the joint distribution of e* and u*.
A more compelling case for identification
requires the presence of Z. Intuition is again
useful. The larger is the explanatory power of
Z in Eq. 4, the smaller is the correlation between
X and E[e*|sample]. The availability of
Z mitigates the collinearity problem. There is an
extensive literature on instrumental variables
(Hahn and Hausman 2003; Stock et al. 2002).
This entry is not the place to summarize that
literature except to review some properties of
a good instrument.
Parametric Sample Selection Models
First, to be an instrument, Z must appear in the
selection equation but not in the outcome equation. Second, to be a useful instrument, Z must
account for a good deal of the variance in S.
A weak instrument (e.g., one that accounts for
a negligible amount of the variance in S
conditional on X) can provide an estimate with
worse mean-squared error properties than estimates from models that ignore selection bias.
Unfortunately, good instruments are difficult to
find and to justify, a fact that greatly limits the
appeal of the Heckman estimation procedures.
This is a good point to pose the question: If
correcting for selection bias can worsen estimates, can an analyst tell if correcting for selection bias is necessary? The answer is a tentative
yes: The test is equivalent to testing whether
se*u* ¼ 0. The practical problem is that without
a strong instrument, this test will lack power. An
analyst would frequently fail to reject the null
hypothesis of no selection bias when, in fact,
selection bias is serious. There are tests of
whether the parameters are identified by nonlinearities or identification restrictions, but these
tests require correct parametric assumptions
and, anyway, do not answer the question of
whether selection bias exists.
A final point closes this section: Much of the
argument regarding instrumental variables pertains to linear regression models and does not
translate readily into the nonlinear models often
used in criminology/criminal justice research.
For useful discussions, see Davidson and
MacKinnon (1993), pp. 224–226 and Cameron
and Trivedi (2005), pp. 192–199. This is
a difficult problem. Angrist and Pischke (2009)
advise researchers to assess if the research question is answerable without actually estimating
the b. The author (Rhodes 2010) adopted that
recommendation in a justice context.
Monte Carlo Illustrations
A Monte Carlo experiment illustrates points
based on the first case (section “Sample Selection
Due to Unobserved Variables”). There are single
X, W, V, and Z variables. X, V, and Z are
3437
P
distributed as independent normal with mean of
zero and standard deviation of 1; W has different
distributions across the experiments. The a and b
parameters all equal 1. The error terms e and
u are distributed as normal with mean of zero
and standard deviation of 2. The simulation
generates 1,000 observations before selection
(about 500 after selection), and the experiment
is replicated 5,000 times.
In the first experiment (see Table 1), W is
distributed as normal with mean of 0 and
standard deviation of 2. This is ideal for the
Heckman estimator because e* and u* are
distributed as joint normal. The table shows
parameter estimates using an OLS regression,
Heckman’s
maximum
likelihood,
and
Heckman’s two-step procedure. The table
reports the average of the three b parameters
across 5,000 replications (bC is the constant),
the “true” standard error across those 5,000 replications (based on the 5,000 replications) and
the average of the estimated standard errors
across those 5,000 replications.
OLS with a robust covariance estimator yields
biased and inconsistent parameter estimates for
bX and bC. The estimate for bV is not biased. The
maximum likelihood Heckman procedure yields
consistent parameter estimates at the expense of
inflating standard errors. The two-step procedure
yields consistent parameter estimates although
the standard errors are even larger. The Heckman
estimators perform well in this ideal case.
The second exercise changes the distribution
for the instrument Z. Although still normally
distributed with mean of zero, it now has a standard deviation of 0.01. This simulates a weak
instrument. Everything else in the simulations
remains the same. Results are reported as exercise 2 in Table 1.
The table shows no results for the maximum
likelihood procedure, which often fails to converge
on a solution. The two-step estimator is less
demanding and makes the intended points. First,
the two-step estimator yields seriously biased
parameter estimates, which should give pause to
those who seek to rely upon nonlinearity to identify parameters. Second, the mean-squared error
properties of the two-step estimator are worse
P
P
3438
Parametric Sample Selection Models
Parametric Sample Selection Models, Table 1 A Monte Carlo demonstration of selection bias
Experiment 1 bX
bV
bC
Experiment 2 bX
bV
bC
Experiment 3 bX
bV
bC
Experiment 4 bX
bV
bC
Experiment 5 bX
bV
bC
Exercises:
Least squares
True
standard
Estimate error
0.728 0.123
0.999 0.118
2.077 0.121
0.694 0.122
1.001 0.118
2.145 0.120
0.872 0.112
1.000 0.107
1.306 0.110
0.034 0.397
1.002 0.375
3.432 0.416
0.973 0.090
0.999 0.086
1.576 0.088
Estimated
standard
error
0.122
0.118
0.121
0.121
0.117
0.120
0.112
0.106
0.113
0.373
0.351
0.403
0.090
0.086
0.089
Maximum likelihood
Two-step
True
Estimated
True
standard standard
standard
Estimate error
error
Estimate error
0.993
0.163
0.160
1.002
0.176
0.999
0.117
0.117
0.999
0.117
1.024
0.416
0.407
0.990
0.487
0.859
1.023
1.001
0.118
1.522
3.925
1.025
0.175
0.153
1.006
0.151
1.000
0.106
0.105
1.000
0.107
0.806
0.436
0.353
0.867
0.351
1.638
3.941
1.006
0.374
3.744 17.346
0.999
0.121
0.119
0.999
0.121
0.999
0.086
0.086
0.999
0.086
1.499
0.244
0.240
1.498
0.242
Estimated
standard
error
0.174
0.117
0.485
1.406
0.146
5.365
0.150
0.105
0.343
5.708
0.499
24.586
0.120
0.086
0.241
1. W has a normal distribution and the instrument is strong
2. W has a normal distribution and the instrument is weak
3. W has a lognormal distribution with variance of 1.02 and the instrument is strong
4. W has a lognormal distribution with variance of 1.02 and the instrument is weak
5. W has a binomial distribution with probability equal to 0.5
than those of the OLS estimator: 0.329 for OLS
and 1.033 for Heckman.
The third exercise is the same as the first but
changes the distribution for W. Previously distributed as normal, it is now distributed as lognormal. This is easily accomplished by taking the
exponential of the original W multiplied by 0.7.
The distribution of W now has a mean of 1.278
and a standard deviation of 1.02. The exercise
subtracts 1.278 from the simulated values of
W to center the new distribution of W on zero.
The assumption of a lognormal distribution for
W violates the assumptions that e* and u* are
distributed as bivariate normal. Nevertheless, the
maximum likelihood and two-step procedures
perform tolerably given a strong instrument.
The fourth exercise is the same as the third but
weakens the instrument by changing Z to have
a normal distribution with standard deviation of
0.01. As before, convergence is a problem for the
maximum likelihood solution, but the two-step
approach provides a solution. Estimates are
biased with large sampling variances.
The fifth and final exercise uses the original
W but now changes it to a dummy variable coded
1 if W exceeds 0 and coded 0 otherwise. The
instrument is strong. As before, the Heckman
procedure provides tolerable solutions especially
for the slope coefficient despite the faulty distributional assumptions.
Simulations cannot cover all situations, but
generalizing based on the above leads to two
conclusions: First, the Heckman procedure is
fairly robust to departures from bivariate normality provided the instrument is strong. Second, the
Heckman procedure provides poor estimates
when instruments are weak regardless of whether
distributional assumptions are correct.
Conclusions
Application of the parametric approach is
demanding, but this is equally true of partially
parametric approaches. An applied researcher
should provide a sober, informed consideration
Parametric Sample Selection Models
of whether his or her data can answer the research
question. The following observations might
inform that consideration.
First, when the researcher has access to a rich
set of data, which includes all variables relevant
to the outcome equation according to substantive
theory, the researcher might assume selection on
the observables. In general, assuming selection
on the observables is difficult to justify, but justification is advanced when the setting allows
repeated observations (difference-in-differences)
and exogenous interruptions (regression discontinuity) provided the researcher is content with
estimating a subset of the b identified by these
designs.
Second, when selection on the observables is
untenable, the researcher needs strong instruments to impose identifying restrictions. Good
instruments are difficult to find and justify, but
examples exist in criminal justice research.
Although nonlinearities can provide identifying
restrictions, these are not reliable.
Weak instruments are a problem. All the estimators discussed in this entry provide estimates
of b that are biased but consistent provided
distributional assumptions hold. The weaker the
instrument, the larger the sample required
to reduce the bias to an acceptable level. Asymptotic justification can require samples in the tens
of thousands when instruments are weak.
Third, it is common for a researcher to estimate two competing models that (1) do not adjust
for selection bias and (2) adjust for selection bias
using parametric assumption. If the two
approaches provide different results, the
researcher might conclude that the second
model is preferred and even necessary. Exercise
2 demonstrated that this is a poor test that provides no basis for concluding that the Heckman
model is an improvement.
Along this line, this entry does not cover an
important body of work that uses RCT to estimate
a treatment effect (case 2 from above) and then
applies the Heckman estimator to observational
data to estimate ostensibly the same treatment
effect. Researchers interpret the difference
between the two estimates as demonstrating
failure of the Heckman approach, but this
3439
P
comparison does not accurately quantify the
bias. Rather, observational data have different
coverage and consequently provide estimates
for a different population (Heckman et al.
1998). Thus, the comparison does not demonstrate failure of the parametric approach to produce useful estimates of treatment effectiveness.
One final point ends this entry: This entry has
treated estimation of the outcome equation as
providing the estimates of scientific interest.
This is not always the case. Especially when
making policy recommendations, the researcher
may be interested in estimation and interpretation
of the selection equation, or the researcher might
be interested in the distribution of treatment
effects across the population, or both (Heckman
and Vytlacil 2007a, b). Even when the selection
equation has little scientific interest, careful
thought about the selection equation is essential
for applying the parametric approach. This
thought exercise is a design consideration driven
by substantive theory. Statistical theory alone
cannot provide a magic solution.
Related Entries
▶ Causes
▶ Econometrics of Crime
▶ Interrupted Time Series Models
▶ Longitudinal Studies in Criminology
▶ Propensity Score Matching
▶ Randomized Experiments in Criminology and
Criminal Justice
▶ Recent Perspectives on the Regression
Discontinuity Design
▶ Sample Selection Models
▶ Sample Selection Problems
Recommended Reading and References
Angrist J, Pischke J (2009) Mostly harmless econometrics: an empiricist’s companion. Princeton University
Press, Princeton
Berk R (2005) Randomized experiments as the bronze
standard. J Exp Criminol 1:417–433
Borenstein M, Hedges L, Higgins J, Rothstein H (2009)
Introduction to meta-analysis. Wiley, West Sussex
P
P
3440
Bushway S, Johnson B, Slocum L (2007) Is the magic still
there? The use of the Heckman two-step correction for
selection bias in criminology. J Quant Criminol
23:151–178
Cameron A, Trivedi P (1998) Regression analysis for
count data. Cambridge University Press, Cambridge
Cameron A, Trivedi P (2005) Microeconometrics:
methods and applications. Cambridge University
Press, Cambridge
Davidson R, MacKinnon JG (1993) Estimation and inference in econometrics. Oxford University Press, New
York
Gerring J (2012) Social science methodology: a unified
framework, 2nd edn. Cambridge University Press,
Cambridge
Greene W (2002) LIMDEP version 8.0: econometric
modeling guide, volume 2. Econometric Software,
Plainview
Greene W (2008) Econometric analysis, 6th edn. Prentice
Hall, Upper Saddle River
Hahn J, Hausman J (2003) Weak instruments: diagnostics
and cures in empirical econometrics. Am Econ Rev
93:118–125
Heckman J (1976) The common structure of statistical
models of truncation, sample selection and limited
dependent variables and a simple estimator for such
models. Ann Econ Soc Meas 5:475–492
Heckman J (1979) Sample selection bias as a specification
error. Econometrica 47:153–161
Heckman J, Vytlacil E (2007a) Econometric evaluation of
social programs, part I: causal models, structural
models and econometric policy evaluations. In: Heckman J, Leamer E (eds) Handbook of econometrics
volume 6B. North Holland Press, Amsterdam
Heckman J, Vytlacil E (2007b) Econometric evaluations
of social programs, part II: using the marginal treatment effect to evaluate social programs, and to forecast
their effects in new environments. In: Heckman J,
Leamer E, Leamer E (eds) Handbook of econometrics
volume 6B. North-Holland Press, Amsterdam
Heckman J, Smith J, Clements N (1997) Making the most
out of programme evaluations and social experiments:
accounting for heterogeneity in programme impacts.
Rev Econ Stud 64:487–535
Heckman J, Ichimura H, Smith J, Todd P (1998) Characterizing selection bias using experimental data.
Econometrica 66(5):1017–1098
Imbens G, Wooldridge J (2009) Recent developments in
the econometrics of program evaluation. J Econ Lit
47(1):5–86
Little R, Rubin D (2002) Statistical analysis with missing
data, 2nd edn. Wiley, Hobokan
Maddala G (1983) Limited-dependent and qualitative
variables in econometrics. Cambridge University
Press, Cambridge
Pearl J (2000) Causality: models, reasoning and inferences. Cambridge University Press, Cambridge
Rhodes W (2010) Estimating treatment effects and
predicting recidivism for community supervision
Parole
using survival analysis with instrumental variables.
J Quant Criminol 26:391–413
Rhodes W, Pelissier B, Gaes G, Saylor B, Camp S,
Wallace S (2001) Alternative solutions to the problem
of selection bias in an analysis of federal residential
drug treatment programs. Eval Rev 25:331–369
Rhodes W, Kling R, Johnson P (2007) Using booking data
to model drug use arrest rates: a preliminary to estimating the prevalence of chronic drug users. J Quant
Criminol 23(1):1–22
Rosenbaum P (2002) Observational studies, 2nd edn.
Springer, New York
Sampson R (2010) Gold standard myths: observations on
the experimental turn in quantitative criminology.
J Quant Criminol 26:489–500
Schaefer J (1997) Analysis of incomplete missing data.
Chapman & HALL/CRC, Boca Raton
Stock J, Wright J, Yogo M (2002) A survey of weak
instruments and weak identification in generalized
methods of moments. J Bus Econ Stud 20:518–529
van den Berg G (2008) Competing risks models. In:
Durlauf S, Blume L (eds) The new palgrave dictionary
of economics, 2nd edn. Palgrave, MacMillan
Vella F (1998) Estimating models with sample selection
bias: a survey. J Hum Resour 33(1):127–169
Parole
▶ Desistance and Supervision
▶ Early Release from Prison
▶ Front Door and Backdoor Sentencing
▶ Mentally Disordered Offenders
Community Supervision
▶ Probation and Community Sanctions
Under
Parole and Prisoner Reentry in the
United States
Joan Petersilia
Stanford Law School, Stanford University,
Stanford, CA, USA
Overview
The US prison and parole population declined in
2009 for the first time in 30 years, perhaps
indicating the beginning of the end of America’s
Parole and Prisoner Reentry in the United States
long commitment to mass incarceration. But current prison population declines may well be
reversed if we do not do a better job of planning
for the reentry of prisoners who have finished
their sentences. At release, many prisoners are
unable to find jobs and suitable housing. Some
will be legally barred from voting, receiving
public assistance, obtaining a driver’s license, or
retaining custody of their children. If the past is
any lesson, many (upward of two-thirds) eventually return to crime and prison, where the cycle
begins again.
Faced with these realities, what can we do?
Fortunately, there is much that can be done. We
must focus attention on parole policies and
practices. Parole – which refers to both
a release mechanism and a method of community
supervision – is central to crime control and
prisoner reintegration. If parole is effective, dangerous offenders remain in prison, and those who
are released are better prepared for reentry. If
parole resources are misdirected, community
safety is threatened as prisoners return home
with few resources and little surveillance.
Which scenario proves true will depend on realistic expectations and an understanding of
parole’s history, current operations, needs of
the parole population, and “what works” in reentry programming.
Fundamentals of Parole and Reentry
Definition and Functions of Parole
People often confuse probation and parole.
Probation is a judge’s sentence that allows
a convicted offender to continue to live in the
community after criminal conviction, with
restrictions on activities and with supervision
for the duration of the sentence. Parole refers to
offenders who have spent time in prison and are
released to complete the remainder of their
sentence under community supervision. Parole
is usually granted from authorities in the correctional system (i.e., a parole board), since responsibility for offenders passes from the judicial
system to the correctional system upon imprisonment. Parolees are technically still in state
custody; they have merely been granted the privilege of living in the community instead of prison.
3441
P
If parolees or probationers violate the rules of
their release, they can be returned to
incarceration.
Inmates are released from prison mandatorily
or discretionarily. Mandatory release is release
after a specified period of time, as required by
law, and occurs in jurisdictions using determinate
sentencing. In determinate sentencing, the
offender is given a set amount of time to serve
by the court, although these sentencing structures
may still incorporate a degree of discretion.
Determinate sentencing eliminates parole boards,
although the exact requirements vary by state.
Discretionary release is at the paroling
authority’s discretion, within boundaries
established by the sentence and by law.
In those states that permit discretionary
release, state laws give parole boards the authority to change, within certain limits, the length of
a sentence that is actually served. Parole officials
may also change the conditions under which
convicted offenders are supervised and may
release offenders from prison to supervision in
the community or to an outside facility. Parole
authorities can also issue warrants revoking
parole and reincarcerating offenders who violate
parole conditions. For jurisdictions with determinate sentencing and no discretion for the timing
of release, the paroling authority may still
determine conditions of release. They thus can
have a direct effect on prison management. For
example, they can increase the number of
prisoners required to be on post-prison supervision, or they can decrease, by policy, the number
of parole revocations returned to prison. It is this
gatekeeper role that makes paroling authorities so
central to current debates about prisoner reentry
and prison crowding.
With parole, like probation, information about
an offender is gathered and presented to
a decision-making authority, and that authority
has the power to release the offender under
specific conditions, which are articulated in
a contract signed by the offender. Contractual
conditions may be standard (applicable to all
parolees) or tailored to particular offenders.
Standard parole conditions are similar throughout
most jurisdictions and include payment of
P
P
3442
supervision fees, finding employment, not carrying weapons, reporting changes of address and
employment, not committing crimes, and submitting to search by the police and parole officers.
Examples of special conditions include periodic
drug testing for substance abusers and registration for sex offenders and arsonists.
The principal responsibility of the parole
officer is to monitor this court-imposed contract
and conditions. If offenders fail to live up to their
conditions, they can be revoked and returned to
jail or prison to serve out the remainder of the
original sentence or to serve a new sentence.
Parole can be revoked for two reasons: (1) the
commission of a new crime or (2) the violation of
the conditions of parole (a technical violation).
Technical violations pertain to behavior that is
not criminal, such as the failure to refrain from
alcohol use or remain employed. In either event,
the violation process is rather straightforward.
Given that parolees are technically still in the
legal custody of the court or prison authorities,
their constitutional rights are severely limited.
When parole officers become aware of violations
of the parole contract, they notify their supervisors, who make a recommendation to the parole
authorities and can easily return a parolee to
prison. Two US Supreme Court cases, Morrissey
v. Brewer, 408 U.S. 471 (1972) and Gagnon v.
Scarpelli, 411 U.S. 778 (1973), established minimum requirements for the revocation of parole,
giving parolees some rights (such as written
notice of the violation and the opportunity to
confront the accuser) in revocation proceedings.
The goals of probation and parole supervision
are identical: to protect the community and help
rehabilitate offenders. These dual functions are
referred to as the law enforcement function,
which emphasizes surveillance of the offender
and close control of behavior, and the social
work function, which attempts to provide
supportive services to meet offenders’ needs.
Both have always been a part of community corrections, and debating which should be of higher
priority has always caused strain. Currently, the
social work function has given way to the law
enforcement function, and probation and parole
officers are often less interested today in treating
Parole and Prisoner Reentry in the United States
clients than in controlling their behavior. Higher
risk clients, combined with shrinking resources,
has caused many agents to prioritize offender
monitoring. Parole agents are equipped with
legal authority to carry and use firearms; to
search places, persons, and property without the
requirements imposed by the Fourth Amendment
(i.e., the right to privacy); to order arrests without
probable cause; and to confine without bail.
Monitoring parolee behavior and delivering
services is managed through caseloads (the
number of parolees assigned to a single parole
agent). Higher-risk parolees are placed on
smaller caseloads, which facilitate more intensive services and surveillance. Caseload assignment is usually based on a structured assessment
of parolee risk and an assessment of the needs or
problem areas that have contributed to the
parolee’s criminality. By scoring personal
information relative to the risk of recidivism
and the particular needs of the offender (i.e.,
a risk/need instrument), a total score is derived,
which determines the particular level of parole
supervision (e.g., intensive, medium, regular,
administrative). Each jurisdiction has established
policies that dictate the contact levels (times the
officer will meet with the parolee). Officers may
also contact family members or employers to
inquire about the parolee’s progress. The purpose
of the contacts is to make sure that parolees are
complying with parole conditions.
Parole and Reentry
The term prisoner reentry began affecting
community corrections in early 2000. Unlike
probation and parole, prisoner reentry is not
a legal status or program, but a new conceptual
framework for thinking about the processing of
criminal offenders, from sentencing to parole
discharge. Prisoner reentry is defined as the
process of leaving incarceration and returning to
society and includes all activities and programming conducted to prepare ex-convicts to return
safely to the community and to live as
law-abiding citizens (Petersilia 2003). Reentry
is an inevitable consequence of incarceration.
As Jeremy Travis (Travis 2005, p. xxi) reminds
us, “reentry is not a form of supervision, like
Parole and Prisoner Reentry in the United States
parole. Reentry is not a goal, like rehabilitation or
reintegration. Reentry is not an option. Reentry
reflects the iron law of imprisonment: they all
come back.”
Refocusing the justice system around a reentry
perspective represents a fundamental paradigm
shift that impacts decisions about the timing of
release, the procedures for making the release
decision, the preparation of the prisoner for
release, supervision after release, and the linkages between in-prison and post-release
activities.
Except for those who die naturally or are executed in prison, 95 % of all state prisoners will
eventually leave prison (Petersilia 2003). Estimates from the Department of Justice show that
nearly one-half of all state inmates will be
released within 1 year and three-quarters will be
released within 5 years (U.S. Government
Accountability Office 2001). Parolees represent
11 % of the 7.2 million people “under correctional control” (incarcerated or on community
supervision) (Glaze and Bonczar 2010). Regardless of whether ex-prisoners are on formal parole
supervision, they all return home forever
changed, often facing isolation, stigma, and
a narrowed array of life chances.
Today’s parole population is mostly male
(88 %), although the number of females has
risen steadily over the past decade. The average
parolee is in his or her mid-thirties, with a median
age of 34. For both males and females, most
parolees are members of racial or ethnic minorities (37 % are black and 19 % are Hispanic or
Latino) (Glaze and Bonczar 2008). The Bureau of
Justice Statistics reported that 37 % of all adults
on parole in 2007 were convicted of drug crimes
(Glaze and Bonczar 2008). Many states are
reconsidering the harsh drug laws passed in the
1980s and 1990s, rolling back mandatory
sentences and accelerating releases for some
drug offenders. The anticipated accelerated pace
of drug-involved prison releases will strain parole
and community resources even further.
Most of those released from prison today have
serious social and medical problems. More than
three-fourths of all prisoners have a history of
substance abuse (one-fourth have histories of
3443
P
injection drug use), and one in six suffers from
mental illness (Ditton 1999). Yet less than a third
of exiting prisoners have received substance
abuse or mental health treatment while in prison.
And while some states have recently provided
more funding for prison drug treatment, the
dramatic increase in the prison population has
resulted in a decline in the percentage of state
prisoners participating in such programs that
have been declining, from 25 % in the 1990s to
about 10 % in 2001 (Lynch and Sabol 2001).
A significant share of the prison population
also lives with an infectious disease. Two to
three percent of state prisoners are HIV positive
or have AIDS, a rate five times higher than that of
the US population (Hammett et al. 2002).
According to the Centers for Disease Control,
about 25 % of all individuals living with HIV or
AIDS in the United States had been released from
a prison or jail that year (National Commission
on Correctional Health Care 2002). This proportion may continue to increase as more drug
offenders, many of whom engage in intravenous
drug use, share needles, or trade sex for drugs, are
released from incarceration. Prisoners with
health problems may have a more difficult
reentry process than others, as they are additionally confronted with the tasks of managing their
health problems, obtaining health care, and
keeping up with medications or appointments.
Few inmates have marketable skills or
sufficient literacy to become gainfully employed
at release. A third of all US prisoners were
unemployed at their most recent arrest, and just
60 % of inmates have a GED or high school
diploma (compared to 85 % of the US adult
population). The National Adult Literacy Survey
established that 11 % of inmates, compared with
3 % of the general population, have a learning
disability and 3 % are mentally retarded
(National Center for Education Statistics 1994).
Again, despite evidence that inmates’ literacy
and job readiness has declined, fewer inmates
are participating in prison education or
vocational programs. Just over 25 % of all those
released from prison in 2001 had participated in
vocational training programs, and about a third of
exiting prisoners will have participated in
P
P
3444
education programs – both figures down from the
previous decade (Lynch and Sabol 2001).
In 2007, the nation spent about $47 billion in
state general funds on corrections (which was 7 %
of all state general fund spending), yet spending
on treatment equaled just 6 % of the annual cost
of housing a prisoner (Pew Center on the States
2009). The need for services for substance-abuse
treatment and educational programming in prison
has never been greater, but the percentage of
prisoners receiving these services has declined.
More punitive attitudes, combined with
diminishing rehabilitation programs, mean that
more inmates spend their prison time “idle.” Ironically, as inmate needs have increased and inprison programs decreased, parole supervision
and community services have also decreased for
most returning prisoners.
Current Issues and Controversies
The Decline of Rehabilitation and Discretionary
Parole Release
The pillars of the American corrections systems –
indeterminate sentencing coupled with parole
release, for the purposes of offender rehabilitation – collapsed during the late 1970s and early
1980s. Attacks on indeterminate sentencing and
parole release centered on three major criticisms.
First, there was little scientific evidence that
parole release and supervision reduced subsequent recidivism. Robert Martinson and his colleagues published the now-famous review of the
effectiveness of correctional treatment and concluded, “With few and isolated exceptions, the
rehabilitative efforts that have been reported so
far have had no appreciable effect on recidivism”
(1974, p. 25). Of the 289 studies they reviewed,
just 25 (9 %) pertained to parole, and yet
Martinson’s summary was interpreted to mean
that parole supervision (and all rehabilitation programs) did not work. Once rehabilitation could
not be legitimated by science, there was nothing
to support the “readiness for release” idea and,
therefore, no role for parole boards or indeterminate sentencing.
Second, parole and indeterminate sentencing
were challenged on moral grounds as unjust and
inhumane, especially when imposed on unwilling
Parole and Prisoner Reentry in the United States
participants. Research at the time showed there
was little relationship between in-prison behaviors, participation in rehabilitation programs, and
post-release recidivism. If that were true, then
why base release dates on in-prison performance?
Prisoners too argued that not knowing their
release dates held them in “suspended animation”
and contributed one more pain of imprisonment.
Third, indeterminate sentencing permitted
parole authorities to use a great deal of
uncontrolled discretion in release decisions, and
these decisions often were inconsistent and discriminatory. Since parole boards had a great deal
of autonomy and their decisions were not subject
to outside scrutiny, critics argued that it was
a hidden system of discretionary decision making
that led to race and class bias in release decisions.
It seemed as if no one liked indeterminate
sentencing and parole in the early 1980s, and
the time was ripe for change. Crime control advocates denounced parole supervision as being
largely nominal and ineffective; social welfare
advocates decried the lack of meaningful and
useful rehabilitation programs. A political
coalition resulted, and soon incapacitation and
“just deserts” replaced rehabilitation as the
primary goal of American prisons.
With that changed focus, indeterminate
sentencing and parole release came under serious
attack, and calls for “abolishing parole” were
heard in state after state. In 1975, Maine became
the first state to eliminate parole. The following
year, California and Indiana established determinate sentencing and abolished discretionary
parole release. By 2002, 16 states had abolished
discretionary parole release for nearly all
offenders. In 19 other states, parole authorities
had discretion over a small and decreasing number of parole-eligible inmates. Likewise, at the
federal level, the Comprehensive Crime Control
Act of 1984 created the US Sentencing Commission and phased out discretionary parole for most
federal prisoners in 1997 (Petersilia 2003).
Proponents hoped that determinate sentencing
with mandatory parole would make sentencing
more consistent across offenders and offenses –
and it has (Marvell and Moody 1996). It was also
thought that abolishing parole would lengthen the
Parole and Prisoner Reentry in the United States
time inmates spent behind bars. After all, parole
release was widely regarded as letting them out
early. Solomon et al. (2005) did find that prisoners
released to post-prison supervision served nearly
a year less than those released without supervision
(mandatorily), but it is not clear whether the longer
prison terms served were the result of discretionary parole systems or the nature of the inmate’s
crime or institutional misconduct.
Opponents of parole also assert that there is no
evidence that placing offenders on parole supervision helps reduce their recidivism rates.
A controversial study by the Urban Institute
“Does Parole Supervision Work?” found no difference in the rearrest rates of offenders released from
prison with and without parole supervision. After
statistically controlling for the offenders’ demographic characteristics and criminal histories, the
researchers found that 61 % of mandatory parolees
(those without supervision) were rearrested, compared with 57 % of discretionary (with supervision)
parole releasees. Solomon et al. (2005, p. 37) concluded that “Parole has not contributed substantially to reduced recidivism and increased public
safety.” This is not to say that parole supervision
could not reduce recidivism, only that at the aggregate level in which it was studied, there was no
evidence that it reduced rearrests.
Rehabilitation Programs, the Cost of Parole
Supervision, and Civil Disabilities
Although 70 % of all persons under correctional
control are on probation or parole, nearly nine out
of ten correctional dollars goes to funding prisons.
Nationally, the 2008 average annual cost of a year
in prison was $29,000. In contrast, the average
annual cost of parole supervision was $2,750. In
California, with the nation’s largest prison and
parole population, the figures are dramatic: for
every dollar California spends on prisons, it spent
15 cents on parole (Pew Center on the States
2009). Parole officers complain of growing paperwork and diminishing resources devoted to treatment of parolees. Some parolees do not participate
in any parole programming at all. It is no wonder
that recidivism rates are so high. In a sense, we get
what we pay for, and we have never chosen to
invest sufficiently in parole or reentry programs.
3445
P
In addition to few treatment resources,
parolees are also subject to a number of statutory
restrictions or “civil disabilities” when they
return home. Many restrictions are statutory,
stemming from a common-law tradition that
people who are incarcerated are “civilly dead”
and have lost all civil rights (Travis 2005).
Their criminal record may preclude them from
voting or retaining their parental rights and be
grounds for divorce, and they may be barred
from serving on a jury, holding public office,
and owning firearms. In 11 states in 2004,
ex-prisoners were permanently denied the right
to vote (Manza and Uggen 2006). Employers are
also increasingly forbidden from hiring parolees
for certain jobs and are mandated to perform
background checks for many others. The most
common types of jobs with legal prohibitions
against parolees are in the fields of childcare,
education, security, nursing, and home health
care – exactly the types of jobs that are
expanding. Since the mid-1980s, the number of
barred occupations has increased dramatically.
Even if a parolee is not legally barred from
a particular job, research shows that ex-offenders
face bleak prospects in the labor market, with the
mark of a criminal record representing an important barrier to finding work (Pager 2003). More
than 60 % of employers claim that they would not
knowingly hire an applicant with a criminal background (Holzer et al. 2002). Overcoming the barriers that ex-offenders face in finding a job is
critical to successful reintegration, since employment helps ex-prisoners be productive, take care
of their families, develop valuable life skills, and
strengthen their self-esteem and social connectedness. Research has also empirically established
a positive link between job stability and reduced
criminal offending. Lipsey’s (1995) metaanalysis of nearly 400 studies found that the single most effective factor in reducing re-offending
rates was employment.
Revolving Door Justice: Inmates Release,
Recidivism, and Prison Return
Staying out of prison is a lot harder than getting
out. The landmark BJS study of prisoners
released in 15 states in 1994 found that fully
P
P
3446
two-thirds (67 %) were rearrested and just over
one-half (52 %) were back in prison, serving time
for a new prison sentence or for a technical violation of their release, within 3 years (Langan and
Levin 2002). The “two-thirds rearrest rate” after
3 years has been documented in the United States
for about 35 years, ever since Daniel Glaser
conducted his classic follow-up study of prisoners in The Effectiveness of a Prison and Parole
System (1969). The Langan and Levin (2002)
study also found that certain characteristics
were associated with parole failure: black
parolees had higher rearrest rates, as did males
in general, those with fewer prior arrests, those
incarcerated for property offenses, and those who
were younger at release (under 25). The failure
rate for prisoners who had been previously
released on parole and were now being
re-released was much higher (64 %) than for
prisoners being released to parole for the first
time (21 %) (Travis and Lawrence 2002).
Rosenfeld et al. (2005), using data from the
BJS recidivism study, calculated arrest probabilities by month for each of the 36 months after
release. They adjusted the probability of arrest by
subtracting out persons who were in jail, in
prison, or dead during the month and therefore
not eligible for arrest. Their results show that the
probability of arrest declines with months out of
prison and that the probability of arrest during the
first month out of prison is roughly double than
that during the fifteenth month. Arrest probabilities also differ by type of crime: prison releasees
arrested on property or drug offenses are more
likely to be arrested early in the post-release
period than those arrested for violent offenses.
Clearly, the first few months after prison
release represent high crime risks, but
Binswanger et al. (2007) also found that death
rates for new prison releases – within the first
days and weeks – are much higher than for
matched demographic groups in the general population. The higher prisoner-release death rates
(12 times the average for the general population)
were caused by high rates of homicide and
drug overdoses. These results have led to calls
to front-load parole services and surveillance so
as to reduce these and other negative outcomes
Parole and Prisoner Reentry in the United States
during the first 6 months after release (National
Research Council 2008).
While individual recidivism studies have
always shown high failure rates, until recently
there was little attention paid to the overall
amount of crime returning prisoners were responsible for. Rosenfeld et al. (2005) reanalyzed the
BJS data and estimated that parolees accounted
for 10–15 % of all violent, property, and drug
arrests between 1994 and 1997 and the share of
total arrests attributable to released prisoners
grew as general crime rates declined during the
1990s. In 1994 the arrests of prisoners released in
the previous 3 years accounted for 13 % of all
arrests. By 2001 that figure had increased to more
than 20 %. If former prisoners are accounting for
nearly one-fifth of all the nation’s arrests, then
investing in prisoner reentry is unquestionably
a matter of public safety. Such evidence has
encouraged law enforcement organizations such
as the International Association of Chiefs of
Police (IACP) and the National District
Attorneys Association to develop policies, training, and other tools that support effective prisoner
reentry (La Vigne et al. 2006).
Parole Violators and Their Impact on Prison
Populations
The BJS reports that about 50 % of released
prisoners will be returned to prison in the 3
years following release – and most of them will
eventually be re-released, and the revolving door
process will continually be repeated. Prisoners
refer to it as “doing a life term on the installment
plan.” The constant churning of parolees in and
out of incarceration is a major contributor to the
growing US population. As a percentage of all
admissions to state prisons, parole violators more
than doubled from 17 % in 1980 to nearly 35 % in
2006 (Sabol and Couture 2008).
In these budget-strapped times, many question
the public safety benefits of violating parolees for
technical violations, such as missing a meeting
with a parole agent, being out after a curfew time,
and, increasingly, failing a drug test. Processing
admissions of parole violators takes as much time
and costs as much money as processing admissions of new convictions and for offenders who
Parole and Prisoner Reentry in the United States
often will be in prison for only a few months.
Many advocates argue that technical violations
should not result in a return to prison and can be
handled in the community with less-drastic intermediate sanctions. The contrary view is that some
violations, even of a procedural nature, are
significant signs that a parolee is not respecting
the terms of the parole contract and should suffer
consequences. This can also be a sign that even if
the current violation is minor, if left unattended, it
might lead to more serious ones. Several states
have begun to use parole violation decisionmaking instruments to respond systematically to
violations. The goal of such instruments is to
differentiate technical violations and the response
to them by the severity and risk posed by the
offender, ultimately decreasing prison admissions for technical violations and reducing disparities in parole decision making.
Responding appropriately to parolee recidivism is one way to manage correctional resources
better, but preventing parolees from failing in the
first place is a more proactive strategy. But that
begs the all-important question of whether we can
implement reentry programs to increase the odds
of success.
Future Directions
No one believes that the current parole and
reentry system is working, but the $64,000
question is, can we do better? Can we improve
the outcomes for people returning from prison so
that they are less likely to be rearrested?
Fortunately, the last decade has seen an explosion
of interest in parole and prisoner reentry, and the
“what works?” literature has improved. After
extensively reviewing all of the literature, the
National Research Council (2008, p. 82)
concluded, “‘Nothing works’ is no longer
a defensible conclusion from assessment of
program effects on re-entry outcomes.”
What Works in Reentry Programming?
Using a variety of techniques, researchers have
developed a list of principles of effective intervention and found that programs adhering to
these principles significantly reduced recidivism,
sometimes by as much as 30 % (Andrews and
3447
P
Bonta 2006; Lowenkamp et al. 2006). Petersilia
(2004) summarized these principles and their
applicability to reentry programs:
• Treatment services should be behavioral in
nature; interventions should employ the cognitive behavioral and social learning techniques of modeling.
• Reinforcements in the program should be
largely positive, not negative.
• Services should be intensive, lasting 3–12
months (depending on need), and occupy
40–70 % of the offender’s time during the
course of the program.
• Treatment interventions should be used
primarily with higher-risk offenders, targeting
their criminogenic needs (dynamic risk factors
for change).
• The best strategy for discerning offender risk
level is to rely on actuarial-based assessment
instruments.
• Conducting interventions in the community as
opposed to an institutional setting will
increase treatment effectiveness.
• In terms of staffing, there is a need to match
styles and modes of treatment service to the
learning styles of the offender (specific
responsivity).
It is impossible to know the extent to which
these evidence-based principles are being used
in current parole and reentry programs. Interest
in prisoner reentry over the last decade has
fueled the development of hundreds of programs
across the United States. Some reentry programs
are small and administered by community- and
faith-based organizations, whereas other
programs are large statewide initiatives, often
administered by the state’s corrections department. Some programs start inside the prison and
continue into the community, whereas others
begin when the prisoner has returned home.
Some programs are residential, others involve
day reporting, and still others involve a meeting
every month or two. They rely on existing and
volunteer community treatment services, primarily Alcoholics Anonymous and Narcotics
Anonymous. The academic “what works” literature barely touches these programs, but these
programs are delivering vital services and
P
P
3448
should not be dismissed, as they are the foot
soldiers of the reentry movement.
Many other reentry programs are part of federally funded reentry initiatives implemented in
recent years. The “what works” literature is
influencing their design and evaluations. The
largest effort to implement evidence-based reentry principles systematically is the federal government’s Serious and Violent Offender Reentry
Initiative (SVORI). In 2003, 69 agencies
representing all 50 states received more than
$110 million in federal funds to develop programs to improve the outcomes of serious and
violent prisoners coming home (Lattimore et al.
2005). The federal government’s SVORI goals
were to improve a variety of outcomes, including
family relationships, work, health, community
integration, housing, and reduced crime. Programs were also encouraged to focus specifically
on serious and violent offenders, or both. Most
programs used a formal risk assessment tool to
identify risks and needs, as recommended in the
“what works” literature. Early outcome results
appear promising. SVORI participants were
more likely to receive services and participate in
needed programs, and SVORI participants are
doing better across a wide range of outcomes.
Recidivism outcomes are not yet reported, but
other outcomes, including housing, employment,
mental health, and substance abuse, have
improved. As Lattimore and Visher (2009)
reported to Congress, “In most cases, the difference in outcomes between those participating in
SVORI programs and the comparison subjects
indicates that SVORI program participation
resulted in an improvement in outcomes.” It is
important to note that most SVORI participants
were repeat serious offenders (the males had an
average of 13 prior arrests), so that even
relatively small reductions in recidivism rates
may be quite cost-beneficial.
In sum, it is possible to reduce offender
recidivism in a cost-effective manner. The answer
lies in investing in reentry programs that incorporate “what works” principles, targeting those programs on specific offenders who can most benefit,
and continually evaluating and revising program
models as the science accumulates.
Parole and Prisoner Reentry in the United States
Conclusions and Future Research
Parole and prisoner reentry issues have captured
the nation’s attention, and the US parole system is
entering another chapter in its history. Attacks on
parole and community-based programs have virtually disappeared and have been replaced with
calls for investing more, not less, in parole and
prisoner reentry. Where there was little scholarly
attention paid to parole just 10 years ago, the
volume and visibility of work around parole and
prisoner reentry issues has grown to such an
extent that it is now commonly referred to as
a full-fledged movement (Petersilia 2009). And
the reentry momentum is likely to continue.
Voters are tiring of our dependence on mass
incarceration as a response to crime and more
willing to embrace a more balanced system of
punishment.
For all that has been learned about parole and
reentry in recent years, a number of important
research questions remain. We now know
a good deal about the needs of returning prisoners
and have some evidence about the services they
receive. However, much less research exists on
the effectiveness of particular approaches to
delivering services, and only a small number of
studies have used rigorous scientific methods to
test promising practices. The highest priority for
future research is more credible program evaluations. To date, no studies have analyzed the
differences among low-, medium-, and high-risk
offenders using an experimental design.
A recent panel of the National Research Council
(2008, p. 82) concluded that while there is a great
deal of experiential and practitioner knowledge
with regard to the apparent efficacy of reentry
programs, “the challenge now is to subject these
promising practices to rigorously designed
evaluations.” Rigorous program evaluations
should accompany every significant reentry initiative, and outcomes for these studies should
focus not solely on recidivism but also on other
behavioral outcomes, such as sobriety, stability in
housing and employment, and attachment to
families and communities. Our studies must also
disaggregate the characteristics of the offender
population so that we can design better programs
for specialized populations, such as women,
Parole and Prisoner Reentry in the United States
the elderly, sex offenders, or the mentally ill.
A higher proportion of parolees in the future are
likely to be composed of one of these distinct
population groups, and we know very little
about how to deliver services to meet their
specific needs.
We also need a better understanding of how
neighborhood characteristics affect the reintegration of offenders. Hipp et al. (2010) recently
reported that the presence of more social service
providers nearby (within two miles) led to lower
recidivism rates and that this protective effect
was particularly strong for African American
and Latino parolees. They also found that
parolees living in neighborhoods with higher
levels of concentrated disadvantage experienced
greater rates of recidivism, even after taking into
account the individual characteristics of these
parolees. This study highlights the importance
of social context for successful reintegration.
Research must move beyond simple statistical
models that attempt to explain parolees’ return
to prison solely as a function of the parolees’
background and behavior, since the characteristics of their agent, supervising agency, and community may be significant predictors as well.
Finally, researchers must study successful
parolees in order to uncover the factors that
encourage offenders to shift from formal social
controls to informal social controls. Ultimately,
parolees who make it shift from being accountable to programs and criminal justice agencies
(e.g., police, parole) to being accountable to
more informal social controls (e.g., families,
neighbors). Ideally, formal criminal justice
sanctions should act as “presses” to increase
social bonds to law-abiding family members
and conventional institutions. Ethnographic
studies can identify how to promote positive
social bonds between ex-convicts and community members.
We are now witnessing the start of a new
ideological pendulum shift in US punishment
policy. A declining economy that is pressuring
states to reduce incarceration, combined with
a growing body of evidence identifying effective
reentry programs, has created a window of opportunity for change. Parole and prisoner reentry
3449
P
may well serve as a major conceptual framework
for reorganizing criminal justice policy in the
twenty-first century.
Related Entries
▶ Aging Correctional Populations
▶ Aging Prison Population: Factors to Consider
▶ Crime and the Racial Composition of
Communities
▶ Criminal Careers and Public Policy Responses
▶ Deterrent Effect of Imprisonment
▶ Drug Abuse and Alcohol Dependence Among
Inmates
▶ Drug Courts
▶ Health Issues in Prison Reentry Models
▶ History of Corrections
▶ History of Probation and Parole in the United
States
▶ Intensive Probation and Parole
▶ New Penology
▶ Offender Change in Treatment
▶ Probation and Parole Practices
▶ Probation Officer Decision-Making
▶ Punishment as Rehabilitation
▶ Re-entry
▶ Reentry Courts
▶ Sex Offenders and Criminal Policy
Recommended Reading and References
Andrews D, Bonta J (2006) The psychology of criminal
conduct, 4th edn. Anderson, Cincinnati
Binswanger I, Stern M, Deyo R, Heagerty P, Cheadle A,
Elmore J, Koepsell T (2007) Release from rison –
a high risk of death for former inmates. N Engl
J Med 356:157–165
Ditton P (1999) Mental health and treatment of inmates
and probationers. Bureau of Justice Statistics, Washington DC
Glaze LE, Bonczar T (2010) Probation and Parole in the
United States, 2009. Bureau of Justice Statistics,
Washington, DC
Glaze LE, Bonczar TP (2008) Probation and Parole in
the United States, 2007 – statistical tables. Bureau
of Justice Statistics. U.S. Department of Justice,
Washington, DC
Hammett T, Harmon P, Rhodes W (2002) The burden of
infectious disease among inmates and releasees from
P
P
3450
correctional facilities. In: National Commission on
Correctional Healthcare (ed) The health status of
soon-to-be-released inmates. National Institute of
Justice, Washington DC
Hipp J, Petersilia J, Turner S (2010) Parolee recidivism in
California: the effect of neighborhood context and
social service agency characteristics. Criminology
48(4):947–979
Holzer H, Raphael S, Stoll M (2002) Can employers play
a more positive role in prisoner reentry? Urban Institute, Washington, DC
Langan P, Levin D (2002) Recidivism of prisoners
released in 1994. Bureau of Justice Statistics,
Washington, DC
Lattimore PK, Visher CA, Winterfield L, Lindquist C,
Brumbaugh S (2005) Implementation of prisoner reentry programs: findings from the serious and violent
offender reentry initiative multi-site evaluation. Justice Res Policy 7(2):87–109
Lattimore PK, Visher CA (2009) The multi-site evaluation
of the serious and violent offender reentry initiative.
The Urban Institute. Available at http://www.urban.
org/publications/412075.html
Lipsey MW (1995) What do we learn from 400 research
studies on the effectiveness of treatment with Juvenile
Delinquency? In: McQuire J (ed) What works:
reducing reoffending. Wiley, West Sussex
Lowenkamp CT, Latessa EJ, Smith P (2006) Does correctional program quality really matter? The impact of
adhering to the principles of effective intervention.
Criminol Public Policy 5(3):575–594
Lynch JP, Sabol WJ (2001) Prisoner reentry in perspective. Urban Institute, Washington, DC
Manza J, Uggen C (2006) Locked out: felon disenfranchisement and American democracy. Oxford University Press, New York
Martinson R (1974) What works? Questions and
answers about prison reform. Public Interest
35(Spring):22–35
Marvell TB, Moody CE (1996) Determinate sentencing
and abolishing parole: the long-term impacts on
prisons and crime. Criminology 34(1):107–128
National Center for Education Statistics (1994) Literacy
behind prison walls. U.S. Department of Education,
Washington, DC
National Commission on Correctional Health Care
(2002) The health status of soon-to-be-released
inmates: a report to congress. National Institute of
Justice, Washington, DC
National Research Council (2008) Parole, desistance from
crime, and community integration. National Academy
of Sciences, Washington, DC
Pager D (2003) The mark of a criminal record. Am
J Sociol 108(5):937–75
Petersilia J (2003) When prisoners come home: parole and
prisoner reentry. Oxford University Press, New York
Petersilia J (2004) What works in prisoner reentry?
Reviewing and questioning the evidence. Fed Probat
68(2):4–8
Pathways to Delinquency
Petersilia J (2009) Transformation in prisoner reentry:
what a difference a decade makes. Afterword in
When Prisoners Come Home by Petersilia, Oxford
University Press
Pew Center on the States (2009) One in 31: the long reach
of American corrections. The Pew Charitable Trusts,
Washington, DC
Rosenfeld R, Wallman J, Fornango R (2005) The contribution of ex-prisoners to crime rates. In: Travis J,
Visher CA (eds) Prisoner reentry and crime in America. Cambridge University Press, New York
Sabol WJ, Couture H (2008) Prison inmates at midyear
2007. Bureau of Justice Statistics, Washington
Solomon AL, Kachnowski V, Bhati A (2005) Does
parole work? Analyzing the impact of postprison
supervision on rearrest outcomes. Urban Institute,
Washington, DC
Travis J (2005) But they all come back: facing the
challenges of prisoner reentry. The Urban Institute,
Washington, DC
Travis J, Lawrence S (2002) Beyond the prison gates: the
state of parole in America. Urban Institute,
Washington, DC
U.S. Government Accounting Office (2001) Prisoner
Releases: trends and information on reintegration
programs, Washington, DC
Vigne L, Nancy G, Solomon A, Beckman K,
Dedel K (2006) Prisoner reentry and community policing: strategies for enhancing public safety. Urban Institute, Washington, DC
Pathways to Delinquency
Rolf Loeber
Department of Psychiatry, University of
Pittsburgh, Pittsburgh, PA, USA
Overview
Modern criminological theories aim to explain
crime, but they do so with a different emphasis
on pathways to crime and ranges of explanatory
factors (e.g., Farrington 2005; Thornberry and
Krohn 2003). Although theories of antisocial
and delinquent behavior often have several factors in common (e.g., juveniles’ relationships
with parents and peers), they differ in their relative emphasis on domains, settings, and details of
explanatory factors, and the ways that these factors are developmentally interrelated (see above
Pathways to Delinquency
sources and chapters in Farrington 2005, and in
Lahey et al. 2003). The models also much differ
in the extent to which they are two-dimensional
(i.e., on one surface) on paper or can be visualized
in a better ways. This entry concerns advances in
the conceptualization and visualization of development of behavioral stages, leading to serious
forms of delinquency and explanatory factors that
account for individual differences in how far
individuals progress in their offending.
Issues in the Construction of Models to
Explain Delinquency and Crime
Criminological theories (Thornberry and Krohn
2003) almost always share three themes with the
goals of explaining: (a) antisocial and delinquent
behavior over the life-course, particularly in
terms of prevalence, frequency, and severity of
delinquent acts; (b) individual differences in antisocial/delinquent behavior and developmental
changes in these differences; and (c)
nonoffending or low-level offending. The key is
for researchers, practitioners, and policy makers
to understand how the combination of (a), (b),
and (c) explains why some individuals and not
others become serious property offenders or
violent offenders and understand this both
on the population level (e.g., all youth in a
particular city) and on the individual level
(i.e., a particular juvenile minor offender at risk
of becoming a serious offender). Along these
lines, there is a need for better life-course models
that incorporate the development and accumulation of risk and protective factors.
To date, the integration of these components
in a model has taken place almost exclusively
through the following methods. Firstly, model
specification is usually visually illustrated in
a two-dimensional plane of variables in boxes
(representing usually a single outcome of delinquency, such as violence or property crime) and
their interrelationships are expressed by arrows
between the boxes. Secondly, virtually, all model
specification rests on multivariate statistical
tests of the hypothesized relationships between
independent measures also called risk factors
3451
P
(the boxes) and an outcome, such as serious
delinquency. Risk factors are defined here as
events or conditions that are associated with an
increased probability of serious forms of delinquency, distinguishing risk factors from protective factors, which are factors associated with
lowered risk of delinquency.
However, there are at least five limitations to
the models tested to date (Loeber et al. 2006).
Firstly, the empirical results of the multivariate
models have been enormously diverse. The second disadvantage is that most models have not
delineated developmental pathways to serious
forms of antisocial behavior (such as violence)
by the specification of developmental antecedents to such serious outcomes (e.g., nonviolent
forms of aggression). Third, current models do
not specify differences between individuals in
their exposure to an accumulation of specific
risk (or protective) factors with development.
A fourth disadvantage has to do with the fact
that the majority of studies focus only on risk
factors as a way of explaining antisocial behavior
and either neglect to consider protective factors
or the combined effect of risk and protective
factors on antisocial outcomes.
We argue that criminological theories can be
enhanced in at least two other ways, firstly by
incorporating past development, and secondly,
by incorporating possible future development.
In the first category are differences in individuals’
past history of offending (and history of behavior
problems that are precursors to delinquency), and
differences in the individuals’ exposure to earlier
risk factors (e.g., prenatal exposure to toxins or
child abuse during the preschool period) and protective factors (e.g., a good relationship with an
adult). The second understudied area has to do
with two aspects of individuals’ futures: firstly,
the possible course of subsequent delinquent
development (or pathways) as examined by longitudinal studies of youth who have been
followed up into adulthood; and secondly,
expected exposure to risk and protective factors
based on the observation of such factors in older
populations of youth. Examples of “new” risk
factors that individuals may be exposed to during
adolescence include: gang membership, violent
P
P
3452
victimization, and alcohol intoxication. In summary, we propose that developmental models
should benefit from knowledge of pathways and
exposure to risk and protective factors based on
past studies which can then provide a framework
in which to place the development of specific
individuals. In that sense, population data can
describe both the stages in pathways to serious
delinquency (the latter being the outer, more
severe boundary) and the gradual accumulation
of risk and protective factors that can maximally
take place in populations (another instance of an
outer boundary). Knowledge of these two types
of outer boundaries will make it possible to
describe the past history of an individual’s antisocial and delinquent behavior, his/her past exposure to risk and protective factors, and possible
future exposure to ongoing or new risk and protective factors.
One of the key advantages of such a model is
that, in contrast to current statistical models, it is
likely to provide valuable information about
choices and targets of intervention. This is very
much in line with contemporary principles of
prevention and intervention, with its focus on
behavioral development and an emphasis on
reducing exposure to risk factors while at the
same time increasing exposure to protective
factors (Pollard et al. 1999).
We argue that the complex of factors
described above is easier to understand by policy
makers, informed lay-persons, and practitioners
when visualized in a model of a three-, rather than
a two-dimensional space or model specified as
a multivariate formulae. To construct such
a visual model, it is necessary to review five key
topics: (1) pathways representing individuals’
escalation toward serious property offenses and
violence; (2) differences between individuals in
their exposure to risk and protective factors at
birth (called start-up factors) and subsequently
during childhood, adolescence, and early adulthood; (3) developmental accumulation of risk
and protective factors over time; (4) the existence
of a dose–response relationship between the
number of risk/protective factors and later serious
delinquency and violence; and (5) additive and
compensatory effects of protective and risk
Pathways to Delinquency
factors. Finally, we integrate these five aspects
into a single, three-dimensional cumulative,
developmental model.
Escalation and the Formulation of
Developmental Pathways
Youth who commit acts of violence or serious
theft rarely do so de novo. Instead, many criminologists, child psychologists, and psychiatrists
agree that the majority of youth who commit acts
of violence or serious theft practiced less serious
forms of delinquency earlier in life (e.g., Le
Blanc 2002; Warr 2002). Thus, the onset of violence is usually preceded by a history of escalation in the severity of aggression, which often, but
not always, starts in childhood (Moffitt 1993).
The findings of several studies focus on
a single escalation pathway. However Loeber
et al. (1993) challenged this and tested a triple
pathway model (Fig. 1), which better fitted the
data. The three pathways are as follows: (a) An
Authority Conflict Pathway prior to the age of 12,
that starts with stubborn behavior, and has defiance as a second stage, and authority avoidance
(e.g., truancy) as a third stage; (b) A Covert Pathway prior to age 15 which starts with minor
covert acts, has property damage as a second
stage, and moderate to serious delinquency as
a third stage; and (c) An Overt Pathway that starts
with minor aggression, has physical fighting as
a second stage, and more severe violence as
a third stage. It is possible that within this last
stage, homicide constitutes a separate, most serious component (Loeber et al. 1993). Recent longitudinal research on the Pittsburgh Youth Study
(Loeber et al. 2011) indicates that 94 % of later
homicide offenders have displayed violence earlier in life. The pathways are hierarchical in that
those who have advanced to the most serious
behavior in each of the pathways have usually
displayed persistent problem behavior characteristics at the earlier stages in each pathway. By
a process of selection, increasingly smaller
groups of youth become at risk for the more
serious behaviors. The pathways are also related
to neighborhoods. For example, a higher
Pathways to Delinquency
3453
AGE OF ONSET:
LATE
%BOYS:
FEW
VIOLENCE
(rape, attack,
strong-arm)
SERIOUS
DELINQUENCY
(auto theft,burglary)
PHYSICAL FIGHTING
(physical fighting,
gang fighting)
MINOR AGGRESSION
(bullying, annoying others)
OVERT PATHWAY
EARLY
P
MODERATELY SERIOUS
DELINQUENCY
(fraud, pick-pocketing)
AUTHORITY
AVOIDANCE
(truancy,
running away,
staying out late)
PROPERTY DAMAGE
(vandalism, fire-setting)
MINOR COVERT BEHAVIOR
(shoplifting, frequent lying)
Defiance/Disobedience
COVERT PATHWAY
(before age 15)
Stubborn Behavior
MANY
AUTHORITY CONFLICT PATHWAY
(before age 12)
Pathways to Delinquency, Fig. 1 Developmental pathways to serious delinquency and violence (Loeber et al. 1993)
percentage of youth in the most disadvantaged
neighborhoods escalates from minor aggression
to violence than youth living in more advantaged
neighborhoods (Loeber and Wikström 1993).
The three-pathway model has been subsequently
replicated across two other samples in the Pittsburgh Youth Study. In addition, several other
studies have replicated the pathways model
(Loeber et al. 1999; Tolan et al. 2000).
Risk Factors
The majority of theories about the causes of
crime and antisocial behavior are generally
based on risk factors alone, although different
terms for such factors have been used (e.g.,
Gottfredson and Hirshi 1990; Moffitt 1993).
Risk factors can be distinguished according to
the different domains and contexts in which
they operate. Overall, researchers and theoreticians agree (with some variation) to distinguish
risk factors in the following domains: individual,
family, and peer group. In addition, some risk
factors are associated with children’s exposure
to risk factors in specific contexts, including characteristics of the school attended and the neighborhood in which the children reside or spend
their time (Howell 2003; Loeber and Farrington
1998; Wikström and Loeber 2000). Data from
prediction studies indicate that risk factors from
each of the domains (individual, family, peers,
schools, and neighborhoods) contributes to the
explanation of why some individuals and not
others progress from minor problem behavior
such as bullying, to physical fighting, to violence
(e.g., Farrington 1997; Loeber et al. 2011). Individuals’ exposure to an accumulation of risk
factors in multiple domains rather than in
a single domain heightens the probability of
P
P
3454
later adverse outcomes. For example, DeaterDeckard et al. (1998) find that each of four
domains (child characteristics, sociocultural,
parent, and peer experiences) contributes to the
prediction of externalizing problems.
Many researchers have been relatively silent
about the developmental saliency and priority of
different domains of risk factors (individual,
family, peer, etc.), other than noting that children’s exposure to family factors usually precedes their exposure to peer factors. As children
develop, the configuration of problem behaviors
changes, with children becoming more mobile and
becoming also more active in selecting settings
with their own risk and protective factors, when
children become exposed to risk and protective
factors of peers, school, and, eventually, work.
Thus, with development, there is a reconfiguration
of risk domains with new domains and new risk
factors being introduced over time.
Developmentally graded Risk Factors. Among
the risk factors, some are thought to be present at
birth or become manifest shortly afterward, while
other risk factors emerge later in individuals’ lives.
Before discussing each category of risk factors,
our overall strategy is to view risk factors from
a developmental angle in that children’s exposure
to different risk factors is gradual, that with development, children’s exposure to different risk
domains increases, and that many risk factors persist over time and, consequently, are “stacked”
over time. It would be clearly untenable to insist
that all the risk and protective factors mentioned in
the preceding section are present in juveniles’
lives from a very young age onward. Instead, it is
much more likely that selection processes operate
in which certain categories of children and youth
are incrementally exposed to certain risk and protective factors as they grow into adults (see
below).
Loeber et al. (2006) reviewed the emergence
of risk factors during childhood and adolescence,
starting with risk factors present at birth, and
those emerging during the following periods:
preschool, elementary school, middle/high
school, and early adulthood. Emergence is
defined as the probable earliest exposure of children to risk factors. Systematic knowledge in this
Pathways to Delinquency
respect is still wanting, and as a result, we have
made estimates (knowing that some of the onset
of risk factors, such as in the case of poor executive functioning, partly depends on the state of
measurement in that area). Note that emergence
should not be taken too literally and may apply
differently from one child to another. However,
we maintain that an accumulation of risk factors
may take place over many years, and that risk
factors at birth (e.g., low IQ, negative emotional
temperament) may be followed by exposure to
“new” risk factors emerging during early childhood, such as language problems, which in turn
may be followed by other novel risk factors
emerging during middle to late childhood, such
as poor social skills, or poor academic achievement. During adolescence, novel risk factors may
consist of heavy substance use or delinquent victimization. These examples are risk factors in the
individual domain, but a similar accumulation of
new risk factors may occur in other domains as
well, such as the family, peers, schools, and
neighborhood (including work). It should be
noted that, to some extent, risk factors emerging
in one life period may persist across another life
period and that even when they are time-limited
(e.g., child abuse), they may influence the probability of serious delinquency in a cumulative
manner.
Loeber et al. (2006) concluded that the total
number of risk factors associated with antisocial
behavior and delinquency to which children can
be exposed is about equally divided at or close to
birth (15), emerging in the preschool years (12),
and emerging in the elementary school age years
(18). Much lower numbers of risk factors probably first appear during the middle and high school
years (5). Thus, the most salient risk window of
children’s exposure to risk factors is prior to
adolescence. This must be true for early-onset
cases of antisocial behavior and delinquency.
Even within the category of early-onset cases,
the model is flexible in that we assume that
none of the early risk factors are necessary (i.e.,
absolutely required) for the emergence of antisocial behavior and delinquency. In that sense,
there can be some substitution of risk factors
that apply to some and not to other early-onset
Pathways to Delinquency
3455
cases. In summary, the evidence suggests that
individual risk factors continue to increase in
numbers and in different configurations after
childhood and constitute a substantial proportion
of all known risk factors after early childhood.
Dose–Response Relationship Between
Risk Factors and Delinquency
Pathways to
Delinquency,
Fig. 2 Proportion of boys
committing violent
offenses for different levels
of risk (Based on Loeber
et al. 2005)
Percentage Commiting Violence
Studies agree that the higher the number of risk
factors, the greater the likelihood that individuals
will be affected by a negative outcome or other
deviant behavior such as serious delinquency.
This association is usually called a dose–response
relationship and has been demonstrated for the
full range of indicators of antisocial behavior:
sociopathy, externalizing problems (DeaterDeckard et al. 1998), conduct problems
(Fergusson and Woodward 2000), serious delinquency (Smith et al. 1995), and violence (Loeber
et al. 2011).
As an example (Fig. 2), a prediction index in
the Pittsburgh Youth Study was constructed on
the basis of the 11 strongest predictors of violence
(Loeber et al. 1993): truancy, low school motivation, onset of delinquency before age ten, cruelty
to people, depressed mood, physical aggression,
and callous/unemotional behavior, low family
SES, family on welfare, high parental stress,
and bad (i.e., disadvantaged) neighborhood
(parent report). Figure 2 also shows that the
dose–response relationship applies to the number
of predictors and the probability of later violence
(Odds Ratio (OR) ¼ 6.0 for four or more risk
P
factors). Remarkably, the range of probabilities
in the Pittsburgh data is from 3 % at zero risk
factors to 100 % at 11 or more risk factors. In
summary, these results (and those of several other
studies) indicate a robust association between the
number of risk factors and the probability of later
violence. Remarkably, the association holds even
when slightly different risk factors are measured
from study to study.
Stability. Implied is that some risk factors may
persist over time. Investigations of risk factors
often ignore the fact that such factors, similar to
the outcomes that they purportedly predict, may
vary in their temporal stability over time. Yet few
studies have addressed the issue of stability of
risk factors. An exception is Sameroff et al.
(1998), who found that environmental risk factors
correlate .77 over a 5-year period (ages 13–18),
which is about the same level of stability as
intelligence. Loeber et al. (2000) found that the
year-to-year stability coefficients for the interactions between parents and their boys between
ages 6 and 18 average.66 for bad parent–child
relationship,.70 for poor parent–child communication, and lower for poor supervision (.56) and
physical punishment (.46). Examination of absolute stability reveals that this is higher for bad
relationships and poor communication, but
decreases with age for poor supervision and physical punishment. It should be noted, however, that
despite level changes in risk factors with development, many of the known risk factors continue
to predict later negative outcomes, including
delinquency, at different ages of children (e.g.,
Loeber and Farrington 1998; Loeber et al. 2008).
100%
75%
50%
25%
0%
Base
Rate
0
1
2
3
4
5
6
Number of Violence Risk Factors
7
8
9+
P
P
3456
It is plausible that the presence of some risk
factors set in motion a cascade of other risk factors. For example, the toxic teratogenic effect of
maternal smoking may increase the probability of
several other risk factors in the offspring, including poor executive functioning, poor academic
achievement in school, low motivation to attend
school, truancy, and, subsequently, delinquency.
As another example, the presence of several risk
factors (e.g., parents’ exposure to enduring stress
and parents’ substance abuse) may increase the
chance of antisocial behavior in the offspring at
home, which in turn, may set the scene for heightened risk of problem behavior outside of the
home. Yet, as Howell (2003) points out, it
remains to be seen when and how risk factors
operate in a “sequential causal chain” (p. 110).
It is clearer that the prevalence of different risk
factors is higher in disadvantaged compared to
advantaged neighborhoods (Stouthamer-Loeber
et al. 2002; Wikström and Loeber 2000).
Protective Factors
Most criminology and psychopathology studies
with a public health slant have routinely
neglected the study of protective factors (Lösel
and Bender 2003). Protective factors are associated with the likelihood of reduced antisocial
behavior/delinquency and/or increased positive
outcomes, including positive adjustment and positive mental health.
Developmentally graded Protective Factors.
There is an increasing body of research on protective factors (see review by Lösel and Bender
2003), but that body is still miniscule in comparison to the number of publications on risk factors
pertaining to antisocial behavior and delinquency.
In particular, we know very little about the developmental aspects of the accumulation of protective factors through the life-course. However, the
framework proposed by Hawkins and colleagues
(Catalano and Hawkins 1996) and ourselves
(Stouthamer-Loeber et al. 2002) is unusual in
criminology because each postulates changes
within protective factors to explain individual differences in the development of offending. At the
Pathways to Delinquency
risk of being speculative, we postulate that some
protective factors, like risk factors, may be present
at birth, but that other factors emerge during the
first decades of life. Loeber et al. (2006) shows
protective factors mentioned in the research literature (or factors that are probably protective
according to our view) as being relevant to antisocial behavior and delinquency (based on Loeber
and Farrington in press; Lösel and Bender 2003;
Pollard et al. 1999; Sampson and Laub 1990;
Stouthamer-Loeber et al. 2002, 2004). Most of
the protective factors are the inverse of risk factors, in that many, but not all, of the protective
factors represent the other pole of risk factors.
However, we do not advocate that protective factors are the complete inverse of risk factors (see
Stouthamer-Loeber et al. 2004 for details).
Research shows that there are several protective
factors that do not have a risk factor equivalent
(Loeber et al. 2008). Examples of these unique
protective factors are low Attention-Deficit
Hyperactivity Disorder symptoms, and low anxiety. As with the development of risk factors, we
assume that there is a developmentally graded
emergence of protective factors (Loeber et al.
2006). Some of the protective factors are thought
to be present at birth (e.g., moderate to high intelligence, absence of prenatal exposure to toxins),
while other factors become manifest later. It is also
thought that there is a high degree of temporal
stability of protective factors, although the data
on this have not yet been reported. In the Pittsburgh Youth Study (Stouthamer-Loeber et al.
2004), accountability, perceived likelihood of getting caught for delinquent acts, and low physical
punishment are recurring factors that are associated with desistance in persistent serious delinquency at different ages. Relatively little is
known about increases in protective factors with
development. Stouthamer-Loeber (2002) found
that the proportion of significant associations
between protective factors and delinquency
increases with age. This conclusion, however, is
limited by the reason that the age comparisons are
based on different cohorts, and, therefore, require
replication to be proven.
It is likely that protective factors, such as risk
factors, are correlated, in that one protective
Pathways to Delinquency
factor may set the scene for another protective
factor to emerge. For instance, association with
prosocial peers may generate new prosocial
behavior, which, once adopted by an individual
child, may further decrease the probability that
that child will engage in delinquent acts. Hawkins
and his colleagues (2003) are among the few who
have specified prosocial pathways in the delinquency research that can be conceptualized as
a string of protective factors. They postulate that
opportunities for prosocial behavior set the scene
for interpersonal involvement, which in turn produces rewards, improving bonding to others and
promoting belief in a moral order. Research
shows that the prevalence of different protective
factors is highest in the most advantaged neighborhoods (Stouthamer-Loeber et al. 2002;
Wikström and Loeber 2000).
Inverse Dose–Response Relationship
Between Protective Factors and
Delinquency
Is there an inverse dose–response relationship
between the number of protective factors and
later deviance (the higher the number of protective
factors, the lower the probability of deviance)?
There are very few research studies addressing
this question. Outside the area of delinquency,
Sameroff et al. (1998) found that the higher the
number of protective factors, the lower is the
probability of behavior problems, with the results
mirroring the results of the risk factors: “The more
risk factors, the worse the outcomes; the more
protective factors, the better the outcomes”
(p. 172). Similar results have been reported by
Wikström and Loeber (2000), thus lending support
to the notion that the accumulation of protective
factors in juveniles’ lives counters the likelihood
of later delinquent involvement.
Do Protective Factors Offset the Impact
of Risk Factors?
What is the evidence that protective factors
buffer the risk for later antisocial behavior and
3457
P
delinquency? Stouthamer-Loeber et al. (2002)
showed that even at the level of the number of
risk or protective domains (child behaviors, child
attitudes, school, peers, family, demographic
characteristics), the sum of risk and protective
domains (where the number of protective
domains is deducted from the risk domains) linearly predicts persistent serious delinquency
(Fig. 3). The higher the number of risk domains
to which a youth is exposed and the lower his/her
exposure to protective domains, the greater the
likelihood of later persistent delinquency. Conversely, the higher the number of protective
domains and the lower the number of risk
domains, the smaller the likelihood of later persistent serious delinquency. This was replicated
for the youngest and oldest samples; for different
degrees of neighborhood advantage and disadvantage, for specific risk and protective factors,
and for early and late onset forms of delinquency
(Loeber et al. 2008; Stouthamer-Loeber et al.
2002; Wikström and Loeber 2000). Not all
researchers, however, have found that the relationship between risk and protective factors is
linear. For example, Pollard et al. (1999), studying such diverse outcomes as substance use,
arrest, and attack to hurt, found a nonlinear effect
with the effect of risk factors greater as the level
of risk increased. However, it is plausible that
once neighborhood context is taken into account,
nonlinear associations occur. For example,
Wikström and Loeber (2000) report that, “the
overwhelming majority of boys with a high risk
score were involved in serious offending regardless of the socioeconomic context of their
neighbourhood” (p. 1130). However, neighborhood matters greatly for those with a balanced
score of risk and protective factors.
Toward a Three-Dimensional Model of
Developmental Pathways and
Developmentally Graded Risk and
Protective Factors
To summarize firstly, we have reviewed developmental pathways from minor to serious delinquent behavior. Secondly, we have discussed
P
3458
Pathways to
Delinquency, Fig. 3 The
higher the number of risk
domains (and the lower the
number of protective
domains), the higher the
risk of later persistent
serious delinquency
(Stouthamer-Loeber et al.
2002)
Pathways to Delinquency
* *
100
% Persistent Serious Delinquents
P
80
60
40
20
0
*
*
–6
–5
–4
–3
–2
–1
0
1
2
Protective
4
5
6
Risk
High/Medium SES
developmentally graded, cumulative onset of risk
factors and developmentally graded, cumulative
onset of protective factors (Loeber et al. 2006).
Thirdly, we have seen that there are
dose–response relationships between the number
of risk factors and the probability of later
delinquency and violence, and an inverse
dose–response relationship between the number
of protective factors and the reduced probability
of later delinquency and violence. Finally, we
have seen that knowledge of the proportion of
risk and protective factors is more important
than knowledge of either. The question then is
how to bring these different strands of evidence
together? At this point, we want to take a rather
unconventional route. In science, the verification
of models is usually accomplished by mathematical formulae. We are not aware of formulae that
describe the different aspects that we have
stressed. An alternative or rather a step toward
such a composite model is to display the above
interrelationships visually in a three-dimensional
model.
Before discussing this model, we will present
each component. Figure 4 shows a version of two
of the pathways depicted in Fig. 1, which has
been reoriented into the horizontal plane. These
two pathways are overt and covert behavior
(leaving aside for the moment the authority conflict pathway), with the overt pathway triangle
superimposed on the covert pathway triangle.
Figure 5 represents the developmentally graded,
3
Low SES
*N<5
Top: most
serious forms of
delinquency
Covert
Pathway
Overt
Pathway
Pathways to Delinquency, Fig. 4 Pathway model
flipped horizontally (overt pathway on the top and the
cover pathway underneath)
cumulative onset of risk factors into a number of
vertical panels, each of the same shape, but with
ribs of different sizes, representing different
degrees of individuals’ exposure to risk factors,
so that they would fit on the horizontal pathway
picture (Fig. 4). In the central and highest risk rib,
the number of risk factors present at birth is
thought to be higher than in the adjoining ribs,
the accumulation of risk factors is highest, and
the length of the rib is shown to be longest to
represent the higher degree of severity of delinquency. In contrast, the adjoining smaller ribs
represent fewer risk factors at birth and
a smaller accumulation of later risk factors.
Figure 6 represents the distribution of “start-up”
risk factors at birth in the form of a long triangle.
Pathways to Delinquency
3459
P
Risk factor
planes
Promotive factor
planes
Pathways to Delinquency, Fig. 5 Accumulation of
risk factors at different levels of accumulation
Pathways to Delinquency, Fig. 7 Accumulation of
promotive factors at different levels of accumulation
Start-up
Protective
factors
Start-up Risk factors
Pathways to Delinquency, Fig. 6 Distribution of risk
factors at birth
Pathways to Delinquency, Fig. 8 Distribution of promotive factors at birth (start-up promotive factors)
The central top of that triangle indicates
a category of children with the highest number
of risk factors at birth and at the extremes are
those with the fewest number of risk factors at
that time. Figure 6 represents developmental
rather than chronological time.
Figure 7 represents ribs of the accumulation of
protective factors in an inverse way to the risk
factors. However, the central rib is the least tall
because those who advance to the most serious
acts are usually exposed to the lowest number of
protective factors. In contrast, individuals who do
not advance to serious delinquency outcomes
tend to start out in life with and accumulate
more protective factors than those in the central
rib. Figure 8, in the form of half a butterfly, may
clarify the “start-up” protective factors which are
thought to be lowest in the center and highest at
the extremes.
Finally, Fig. 9 shows how each of the above
elements is put together. Individuals who
advance to the most serious delinquent acts on
the horizontal pathways (the flat, double triangle)
are thought to be exposed to the highest number
of risk factors over time and the lowest number of
protective factors over time. At birth, they also
tend to have the highest number of risk factors
(see the start-up risk panel) and the lowest number of protective factors (see the start-up protective factors). To clarify this cumulative,
developmental model, Fig. 10 shows how it
looks from the front, while Fig. 11 shows how it
looks from underneath.
Relevance of the Model for Assessments
and Interventions
What is the relevance of the cumulative, developmental model for assessment and the evaluation of preventive and remedial interventions?
Assessments. Advances have been made in the
past decades in the area of more precise assessment of juveniles’ problem behavior in terms of
P
P
3460
Pathways to
Delinquency, Fig. 9 The
full model of
developmental pathways as
a function of the
accumulation of risk and
promotive factors (shown
from above)
Pathways to Delinquency
Risk factor
planes
Start-up Risk factors
Start-up
Protective
factors
Pathways to
Serious
Delinquency
Promotive factor
planes
Pathways to
Delinquency,
Fig. 10 The full model of
developmental pathways as
a function of the
accumulation of risk and
promotive factors (shown
from the front)
the types of behavior, their severity, and prognostic validity. In addition, screening instruments are
now available to go beyond the range of problem
behaviors of juveniles that also include the past
history of risk factors to which juveniles have
been exposed. Based on the information provided
in this entry, we argue however that assessments
can probably benefit from two other components:
(a) an appraisal of current and past protective
factors to which juveniles are exposed, because
it is the mixture of risk and protective effects that
appears most crucial in determining the future
risk of serious offending as well as the probability
of full desistance or lower-level offending. We
also argue that the next generation of risk
assessment devices could potentially also benefit
from (b) an appraisal of juveniles’ expected
future exposure to risk and protective factors
based on knowledge from longitudinal survey
studies. Specifically, assessments of developmentally graded exposure to risk factors, based
on longitudinal survey studies, could be
employed to calculate the probability that juveniles will be exposed in the future to risk and
protective factors typically emerging for that
age-group.
One might argue that cross-sectional studies
contain such information. There are several reasons why this is not the case. Firstly, in contrast to
cross-sectional studies, longitudinal studies have
Pathways to Delinquency
Pathways to Delinquency, Fig. 11 The full model of
developmental pathways as a function of the accumulation
of risk and promotive factors (shown from the bottom)
the power to specify selection processes determining which groups of youth are most likely to
be exposed to future risk and protective factors.
The second reason has to do with protective factors associated with stable nonoffending or a deescalation from serious to minor or offendingnonoffending. Since juveniles’ engagement in
delinquency varies from year to year, it is important to establish which protective factors are associated with persistent nonoffending or,
alternatively, with a stable de-escalation in the
severity of offending. Longitudinal studies,
rather than cross-sectional studies, are the best
methods by which to identify protective factors.
Preventive and Remedial Interventions. We
agree with Howell (2003) that a developmental
approach to offending needs a reconceptualization of when interventions, preventive or
remedial, can best take place. He has been
a primary proponent of the influential Comprehensive Strategy that addresses the full continuum of antisocial development with two foci for
implementing change: prevention and early intervention, and graduated sanctions. We believe that
this orientation, which is based on knowledge of
risk and protective factors (called protective factors in this case), can also greatly benefit from
the assessment of future risk through developmentally graded expectations of exposure to risk
and protective factors. We argue that this is especially important because of the fact that even
3461
P
the most effective interventions still do not
reduce the risk of recidivism by more than 40 %
(Lipsey and Wilson 1998) and that the
reoffending rate of high-risk populations of
youth still remains extraordinary high (Loeber
and Farrington 1998).
It should be understood that knowledge of risk
and protective factors (Loeber et al. 2006) does
not mean that we have available the tools to
change each of them, and of course some are
not malleable (e.g., single parenthood). However,
we hope that our list of the risk and protective
factors and our model will stimulate the development of other forms of interventions that can
augment the overall efficacy of interventions currently available. It is our impression that most
treatment programs lack components that focus
on increasing or strengthening protective factors.
We foresee that interventions that mobilize or
enhance protective factors (together with the
reduction of risk factors) eventually are likely to
improve treatment efficacy and, possibly, the
maintenance of treatment effects. We agree,
however, with Pollard et al. (1999) that a sole
focus on interventions enhancing protective factors only, because of the compensatory function
of risk and protective factors, is likely to be
inadequate. Another possible beneficial result of
our model specification is the further individualization of interventions tailored to an individual’s
past, present, and possible future exposure to risk
and protective factors.
Finally, we should stress that our design of
a three-dimensional model of antisocial behavior
has certain limitations. Firstly, we are mostly
focusing on the cumulative processes of risk and
protective factors rather than the specifics of
mediation and moderation within and between
risk and protective factors. Also, we do not
focus on possible reciprocal processes between
risk and protective factors and negative or positive child behaviors, or the fact that as children’s
mobility in the community increases with age,
they become more able to select settings and
situations in which novel risk and protective factors may present themselves. We are also aware
that our model may have different components
depending on the subject population of study
P
P
3462
(girls vs. boys, different ethnic groups), but we
think it is generic enough to serve as a model to
explain a range of forms of maladjustment other
than serious delinquency. Moreover, the model
depicts escalation in the severity of antisocial
behavior better than de-escalation to lower severity levels, but de-escalation could perhaps be
incorporated in the next iteration of the model.
Also, there is a need to transfer the basic mathematical properties of our model into future model
testing. It is clear that many tasks await, and we
hope that this text will inspire others to improve
on where we are at this point.
Related Entries
▶ Age-Crime Curve
▶ Career Criminals and Criminological Theory
▶ Desistance from Crime
▶ Identification Issues in Life Course
Criminology
▶ Longitudinal Studies in Criminology
▶ Optimizing Longitudinal Studies in Offending
▶ Risk Assessment, Classification, and
Prediction
Recommended Reading and References
Catalano RF, Hawkins JD (1996) The social development
model: a theory of antisocial behavior. In: Hawkins JD
(ed) Delinquency and crime: current theories. Cambridge University Press, New York, pp 149–197
Deater-Deckard K, Dodge KA, Bates JE, Pettit GS
(1998) Multiple risk factors in the development of
externalizing behavior problems: group and individual
differences. Dev Psychopathol 10:469–493
Farrington DP (1997) Early prediction of violent and nonviolent youthful offending. Eur J Crim Policy Res
5:51–66
Farrington DP (2005) Integrated developmental and lifecourse theories of offending, vol 14, Advances in
criminological theory. Transaction, New Brunswick
Fergusson DM, Woodward LJ (2000) Educational, psychosocial, and sexual outcomes of girls with conduct
problems in early adolescence. J Child Psychol Psychiatry 41:779–792
Gottfredson MR, Hirshi T (1990) A general theory of
crime. Stanford University Press, Stanford
Hawkins JD, Smith BH, Hill KG, Kosterman R, Catalano
RF, Abbott RD (2003) Understanding and preventing
Pathways to Delinquency
crime and violence: findings from the Seattle social
development project. In: Thornberry TP, Krohn MD
(eds) Taking stock of delinquency: an overview of
findings from contemporary longitudinal studies.
Kluwer Academic/Plenum Publishers, New York,
pp 255–312
Howell JC (2003) Preventing and reducing juvenile delinquency, a comprehensive framework. Sage Publications, Inc., Thousand Oaks
Lahey BB, Moffitt TE, Caspi A (2003) Causes of conduct
disorder and juvenile delinquency. Guilford, New York
Le Blanc M (2002) The offending cycle, escalation and
de-escalation in delinquent behavior: a challenge for
criminology. Intern J Comp Appl Crim Justice
26:53–83
Lipsey MW, Wilson DB (1998) Effective intervention for
serious juvenile offenders: a synthesis of research. In:
Loeber R, Farrington DP (eds) Serious and violent
juvenile offenders: risk factors and successful interventions. Sage Publications, Thousand Oaks, pp 313–346
Loeber R, Drinkwater M, Yin Y, Anderson SJ, Schmidt
LC, Crawford A (2000) Stability of family interactions
from ages 6 to 18. J Abnormal Child Psychol 28:353–
369
Loeber R, Farrington DP (1998) Serious and violent
juvenile offenders: risk factors and successful
interventions. Sage, Thousand Oaks
Loeber R, Farrington DP, Stouthamer-Loeber M, White
HR (2008) Violence and serious theft: development
and prediction from childhood to adulthood. New
York: Routledge
Loeber R, Farrington DP (in press) Protective and risk
factors pertaining to violence in four longitudinal
studies. Am J Prev Med 43:S24–S27.
Loeber R, Wikström PO (1993) Individual pathways to
crime in different types of neighborhood. In:
Farrington DP, Sampson RJ, Wikström PO (eds) Integrating individual and ecological aspects of crime.
Liber Forlag, Stockholm, pp 169–204
Loeber R, Keenan K, Lahey BB, Green SM, Thomas C
(1993) Evidence for developmentally based diagnoses
of oppositional defiant disorder and conduct disorder.
J Abnorm Child Psychol 21:377–410
Loeber R, Keenan K, Zhang Q (1997) Boys’ experimentation and persistence in developmental pathways
toward serious delinquency. J Child Fam Stud
6:321–357
Loeber R, Wei E, Stouthamer-Loeber M, Huizinga D,
Thornberry T (1999) Behavioral antecedents to serious
and violent juvenile offending: joint analyses from the
Denver youth survey, Pittsburgh youth study, and the
Rochester development study. Stud Crime Crime Prev
8:245–263
Loeber R, Slot NW, Stouthamer-Loeber M (2006)
A three-dimensional, cumulative developmental
model of serious delinquency. In: Wikström POH,
Sampson R (eds) The explanation of crime: contexts
and mechanisms. Cambridge University Press, Cambridge, pp 153–194
Penal Abolitionism
Loeber R, Farrington DP, Stallings R (2011) The Pittsburgh youth study. In: Loeber R, Farrington DP (eds)
Young homicide offenders and victims. SpringerVerlag, New York, pp 19–36
Lösel F, Bender D (2003) Protective factors and resilence.
In: Farrington DP, Coid J (eds) Early prevention of
adult anti-social behaviour. Cambridge University
Press, Cambridge
Loeber R, Wung P, Keenan K, Giroux B, StouthamerLoeber M, Van Kammen WB, Maughan B (1993)
Developmental pathways in disruptive child behavior.
Dev Psychopathol 5:101–132
Moffitt TE (1993) Adolescence-limited and life-cyclepersistent antisocial behavior: a developmental taxonomy. Psychol Rev 100:674–701
Pollard JA, Hawkins JD, Arthur MW (1999) Risk and
protection: are both necessary to understand diverse
behavioral outcomes in adolescence? Soc Work Res
23:145–158
Sameroff AJ, Bartko WT, Baldwin A, Baldwin C, Seifer
R (1998) Family and social influences on the development of child competence. In: Lewis M, Feiring
C (eds) Families, risk, and competence. Lawrence
Erlbaum Associates, Mahwah, pp 161–185
Sampson RJ, Laub JH (1990) Crime and deviance over the
life course: the salience of adult social bonds. Am
Sociol Rev 55:609–627
Smith C, Lizotte AJ, Thornberry TP, Krohn MD
(1995) Resilient youth: identifying factors that prevent
high risk youth from engaging in delinquency and drug
use. In: Hagan J (ed) Current perspectives on aging and
the life cycle, vol 4, Delinquency and disrepute in the
life course. JAI, Greenwich, pp 217–247
Stouthamer-Loeber M, Loeber R, Wei E, Farrington DP,
Wikström PO (2002) Risk and protective effects in the
explanation of persistent serious delinquency in boys.
J Consult Clin Psychol 70:111–123
Stouthamer-Loeber M, Wei E, Loeber R, Masten AF
(2004) Desistance from persistent serious delinquency
in the transition to adulthood. Dev Psychopathol
16:897–918
Thornberry TP, Krohn MD (2003) Taking stock of delinquency: an overview of findings from contemporary
longitudinal studies. Kluwer Academic/Plenum Publishers, New York
Tolan PH, Gorman-Smith D, Loeber R (2000) Developmental timing of onsets of disruptive behaviors and
later delinquency of inner-city youth. J Child Fam Stud
9:203–230
Warr M (2002) Companions in crime: the social aspects of
criminal conduct. Cambridge University Press, Cambridge, UK
West DJ (1982) Delinquency: its roots, careers and prospects. Heinemann, London
Wikström PO, Loeber R (2000) Do disadvantaged neighborhoods cause well-adjusted children to become adolescent delinquents? A study of male juvenile serious
offending, risk and protective factors, and neighborhood context. Criminology 38:1109–1141
3463
P
Patrol
▶ Preventive Patrol
Patterns in Youthful Offending
▶ Specialization in Juvenile Offending
Peer Effects
▶ Network Analysis in Criminology
Penal Abolitionism
Vincenzo Ruggiero
School of Law, Middlesex University,
London, UK
Overview
Abolitionism is not only a strategy or a set of
demands aimed at the reduction (or suppression)
of custody, it is also a perspective, a philosophy,
an approach. Penal abolitionism challenges conventional definitions of crime and the law, while
defying official views of the meanings and effects
of punishment. This entry examines the origin,
philosophy, and achievements of penal abolitionism, collecting and discussing the key arguments
and views expressed by its leading proponents. It
sets off with reference to Tocqueville’s view that
prisoners lose their right to humanity and
Foucault’s remarks, during his visit to Attica,
about the “consumption and elimination” of the
excluded enacted by prison institutions. It then
discusses the degree to which the abolitionist
stance can be deemed utopian and presents the
classic arguments which constitute the components of that stance. Three areas are addressed:
crime, the law, and punishment, namely, the key
P
P
3464
areas in which abolitionism elaborates its own
philosophy while distancing itself from conventional sensibilities and, at the same time,
responding to its mainstream critics.
Introduction
Abolitionism fights against the “curious mechanism of circular elimination” identified by
Foucault during his visit to Attica, where he
noted that society excludes certain specific types
of its members, and by sending them to prison, it
breaks them up, crushes them, physically eliminates them. Then
The prison eliminates them by “freeing” them and
sending them back to society; the state in which
they come out insures that society will eliminate
them once again, sending them to prison. Attica is
a machine for elimination, a form of prodigious
stomach, a kidney that consumes, destroys,
breaks up and then rejects, and that consumes in
order to eliminate what it has already eliminated.
(Simon 1991, p. 27)
Similarly, observing the American prison system, de Tocqueville (1956) noted that offenders
released from custody remained among humans,
but they lost their rights to humanity; people fled
them as impure, and even those who believed in
their innocence abandoned them. Once released,
they could go in peace, with their live generously
left to them, but a life worse than death.
Against this strategy of elimination, the primary challenge for penal abolitionists is “to construct a political language and theoretical
discourse that disarticulates crime from punishment” (Davis 2008, p. 3). However, such strategy
is accompanied by alternative conceptualizations
of crime, critical analyses of law, and radical
thinking around the very nature, function and
philosophy of punishment.
According to one of the most representative
figures of this multifaceted school of thought,
abolitionism is a manifestation of
the general human urge to do away with and to
struggle against those phenomena or institutions of
a social, political or religious nature that at a given
time are considered to be unjust, wrong or unfair.
(Bianchi 1991, p. 9)
Penal Abolitionism
The “general human urge” in the definition
provided by Bianchi is the same urge that gave
strength to the ancestors of contemporary abolitionists, namely, the women and men who fought
against slavery, and later the campaigners who
battled, and continue to do so, for the abolition of
the death penalty. But the battle also continues on
other fronts. Abolitionism posits that the criminal
justice system as a whole constitutes a social
problem in itself. Some forms of abolitionism,
however, are already in place. Surely, there is
an abolitionist stance in the proposition that the
state-centralized administration of penal justice
should be replaced by decentralized forms of
autonomous conflict regulation. Echoes of abolitionism are also heard when critics warn that, if
we want to reform the penal system, we have to
start a process of collective conversion away
from the traditional and conventional grammar
that characterizes it. It should be reiterated, therefore, that abolitionism does not advocate the
immediate suppression of all custodial institutions, a program which, on the other hand,
would enjoy the support of many a critic who
find the conditions of imprisonment all over the
world shameful. Abolitionism t is not
a “program,” but an approach, a perspective,
a methodology, and most of all a way of seeing.
By observing the concepts and the cultural matrix
which give shape to abolitionism, one may discover that this school of thought sits very comfortably within Western popular culture, which
guides the conduct of many and which many
could mobilize to justify their conduct.
Utopia?
If the activist ancestors of abolitionism are the
men and women who fought against slavery and
the death penalty, their philosophical ancestors
are harder to identify with precision. But let us
start with a general, preliminary characterization.
Abolitionism, I would submit, is grounded in
a variety of social philosophies which are primarily concerned with discussing processes of social
development that can be viewed as pathological
or as “misdevelopments” (Honneth 2007).
Penal Abolitionism
Among these philosophies are those expressing
the view that societies should support a rich plurality of activities, each valuable in its own right,
and that each person should be treated as an end,
not as a mere means to the ends of others. This
Aristotelian view would deny “that a society can
be flourishing as a whole when some members are
doing extremely badly” (Nussbaum 2000, p. 106).
There is nothing utopian in attempts to redress
“remediable injustices”: abolitionists do not pursue perfect justice, rather, they aim at enhancing
justice. Their focus on social interactions rather
than institutions, on precise settings in which
people live rather than official norms and extraneous professionals, locates them in a specific
political and philosophical tradition. According
to a distinction suggested by Amartya Sen
(2009), there are contractarian approaches and
comparative approaches to the idea of justice.
The former establish general, universal, principles of justice and are concerned with setting up
“just institutions.” For such institutions to function, total compliance of people’s behavior is
required. The latter assess the different ways in
which people lead their lives, actually behave and
interact. A contractarian approach is described by
Sen as “transcendental institutionalism,” in that it
searches for the ideal institutions capable of forging a perfectly just society. By contrast,
a comparative approach is led by the search for
social arrangements that satisfy people in their
concrete collective life.
When people across the world agitate for more
global justice – and I emphasise here the comparative word “more” – they are not clamouring for
some kind of “minimal humanitarianism”. Nor are
they agitating for a “perfectly just” world society,
but merely for the elimination of some outrageously unjust arrangements to enhance global justice. (Sen 2009, p. 26)
Comparison entails information, which in turn
presupposes proximity to the actors involved in
the process of forging ideas of justice. Abolitionists propend for this type of approach.
Some forms of human suffering may be
unavoidable, and perhaps cannot be remedied in
some particular place or at some particular time.
Correspondingly, however, there are forms of
3465
P
suffering that are needless at particular times
and places. According to Gouldner (1975), it is
the sociologist’s job to give special attention to
the latter. Penal suffering is avoidable, particularly if its ineffectiveness can be proven, and the
prisoners’ standpoint deserves to be heard not
because they have any special virtue and not
because they alone live in a world of suffering.
A sociology of the underdog is justified because,
and to the extent that, his suffering is less likely to
be known and because – by the very reason of his
being underdog – the extent and character of his
suffering are likely to contain much that is avoidable. (Gouldner 1975, p. 37)
Similarly, Nils Christie often reminds us that
all he wants to do is to reduce the amount of
suffering in the world: “I have never known
someone who wants to increase it” (personal
communication). Christie may have been particularly fortunate in his encounters, but his views
and those of other abolitionists sit comfortably
with a variety of insights found in sociology,
politics, and philosophy.
Problematic Acts
Are crimes exceptional events? Louk Hulsman’s
reply to this rhetorical question takes the form of
an excursus into the criminal justice system as
a “special” set of procedures addressed to a “not
so special” category of individuals. We are
accustomed, he argues, to regarding the criminal
justice system as a rational apparatus specifically devised to respond to crime and to control
or neutralize its effects. We are also accustomed
to interpreting crimes as exceptional events,
“events which differ to an important extent
from other events which are not defined as criminal” (Hulsman 1986, p. 63). In conventional
perceptions, criminal conduct is at odds with
the average conduct in that it deviates from the
type of interactions supposedly established by
the majority. Such deviation, statistically exceptional, is thus deemed to trigger the special
responses offered by official institutions: special
events justify the special nature of the reaction
against them.
P
P
3466
According to abolitionist analysis, conducts
classified as criminal are only a small proportion
of a variety of similar conducts which escape
criminalization. On the one hand, therefore,
only certain types of events are selected as specific, undeniable, self-evident, representatives of
what we understand as crime. On the other hand,
the events selected are brought together in
a discreet category, as if they constituted
a homogenous group of conducts. “Within the
concept of criminality a broad range of situations
are linked together. Most of these, however, have
separate properties and no common denominator” (Hulsman 1986, p. 65). What do violence
within the family and street violence have in
common? Do shoplifting, drug selling, and
armed robbery belong to the same rubric of
behavior? What makes dangerous driving similar
to fencing? And political violence to pollution?
Neither the motivation of those involved nor the
techniques required, let alone the consequences
of those specific conducts, displays a precise set
of common components. If we are led to analyze
them by referring such conducts to a common
logical and moral framework, it is because the
framework we adopt is inspired by the criminal
justice system: “All these events have in common
is that the criminal justice system is authorised to
take action against them” (Hulsman 1986, p. 65).
This is not denying that human interactions
may cause considerable degrees of suffering
to those involved: abolitionism would only
question why some types of suffering mobilize
the intervention of institutional agencies while
others do not. Moreover, in an imaginary scale
of personal hardship, the distress caused by some
interactions of a criminal nature would not score
particularly high when compared to those of
a quotidian, apparently prosaic, nature. “Matrimonial difficulties, difficulties between parents
and children, serious difficulties at work and
housing problems will, as a rule, be experienced
as more serious.” In brief, there is nothing that
intrinsically distinguishes, for those directly
involved, criminal events from a variety of other
unpleasant events. Most episodes causing hardship are normally dealt with informally by the
individuals affected, at times with the mediatory
Penal Abolitionism
participation of people or groups belonging to the
community in which such episodes occur. Many
conducts that generate serious harm do not elicit
responses from the criminal justice system, while
those involved in problematic situations mostly
attempt to find solutions within the context in
which they arise: the family, the group of friends
or neighbors, and the work place. All this means,
Hulsman concludes, “that there is no ‘ontological
reality’ of crime” (Hulsman 1986, p. 66). Some
examples may provide support to this conclusion.
A murder occurs in a town inhabited by some
300,000 people, and the alleged murderer is
a university lecturer. A few days earlier, the lecturer had delivered a public speech and no one in
the audience noticed anything unusual in the
speech and in the speaker. The event is incomprehensible and the investigators feel the necessity to
entrust a team of psychiatrists with the explanation
of the mystery. Let us now imagine another murder occurring 200 years before in a small town.
If we had lived in that small town at that time, we
would probably have found it ridiculous to call in
an expert of the mind to explain why the killer had
killed. Ridiculous because we all knew why he
killed. After all, we would not have been surprised,
and we would have agreed among friends that this
was exactly what we all might have expected all
along, (Nils Christie, personal communication)
The difference between the experience and
perception of the two murders resides in the
nature and amount of information the residents
share about one another. So many people live in
a middle-sized modern town that it is impossible
to know them all. In addition, life is organized in
ways that only allow us to hold a superficial or
segmented knowledge of other beings. We have
a narrow basis for predicting behavior outside the
specific group to which we belong.
Another example comes from the Norwegian
valleys, where the traditional institutional figure
of the lensmann is still operative. A local sheriff
of sort, the lensmann performs a variety of civil
tasks and his/her ability to function is dependent
upon the support received by fellow citizens. The
lensmann may direct auctions, help mothers get
money from absent fathers, and deal with crime.
When Christie interviewed one such institutional
Penal Abolitionism
figure, he was told that there was no crime in the
district. But during the course of the interview,
several things happened that testified to the
contrary.
The telephone rang, a lady had lost her purse: the
lensmann asked his assistant to drive down to the
close-by café; the purse was found and brought
back to the lady. So was the young man who was
using the purse. He happened to be the lady’s son.
A report came in on breaking and entering into
a store of weapons. The lensmann jumped into his
car, drove far up into the mountains in the direction
of the store, met a car high up there, stopped the
car, found Ole drunk as usual, with a carload of
guns he had stolen just to irritate his farther. He
brought Ole home and took the guns to a safer
depot. What a cliff-hanger story lost for the massmedia! Helicopters and anti-terrorist police might
have been engaged in the crime-hunt of the century. Now, it was Ole. An old story of misery and
family quarrels. (Christie 1982, pp. 73–74)
These cases and the one following show that
definitions of crime are based on “ignorance” or
lack of familiarity with the events at hand.
A man administers some drugs to his wife and
causes her death by suffocation. He then writes to
the police, saying that he has committed murder
and that he is going to drown himself. The letter
reaches the police 2 days after the man’s suicide.
The flat where the coupled lived is visited by the
officers, who find the dead woman. She had
Alzheimer’s and, before killing her, the man had
looked after her with deep love. “To some this is
a story of Romeo and Juliet. To others, it is one of
plain murder” (Christie 2004, p. 1). From numerous stories and situations such as this, abolitionists
derive the argument that crime is a shallow concept, too imprecise to allow the subtle distinctions
and understandings we need. The alternative concept of “trouble” may be a better starting point,
trouble being something which is widely experienced and simultaneously caused by people
interacting. Moreover, trouble requires responses,
solutions, lest it causes harm and suffering. “The
danger is too hastily to define troubles as crime.”
The next step is, therefore, to examine in detail
which acts are beneficial and which harmful.
Then follows an analysis of these acts perceived as
bad – a classificatory scheme with categories as
irritation, unpleasantness, disgust, sin - and then,
3467
P
but only as one among so many alternatives –
crime. When crime is the last concept in the line,
it is easier to raise the analytical question: What are
the social conditions for acts to be designated as
crimes? Crime does not exist. Only acts exist,
acts often given different meanings within various
social frameworks. (Christie 2004, pp. 2–3)
Let us take a final example of how social
frameworks determine the meaning of acts.
A man carrying a bag-load of beer cans arrives
in a park. He starts drinking one can after the
other, and because he acts “funny,” he attracts
the attention of the children playing nearby who
surround him and enjoy the show. Suddenly, the
man goes behind a bush with the intention of
relieving himself and the children, who are carried away by the entertaining situation, follow
him. It is summer, and many people are enjoying
the sun from their windows or balconies in two
adjacent blocks of flats. Christie names the two
buildings as, respectively, The House of Perfection and The House of Turbulence. The former
was built by an efficient and reliable company,
the tenants moved in smoothly, on the agreed
day, and to their satisfaction they found everything they expected perfectly in place. The latter
was built by a company that went bankrupt, causing delays and distress among buyers, who organized themselves with the aim of claiming some
form of compensation. The inhabitants of The
House of Turbulence, in other words, knew each
other fairly well, had experienced some form of
collective action, and identified some common
understanding of their needs. Now, a situation in
which a man relieving himself while surrounded
by children lends itself to highly different interpretations. For the inhabitants of the House of
Turbulence, the case was clear.
The man is Peter, son of Anna. He had an
accident when he was little, behaves generally
a bit strangely, but is as kind as the midsummer
night is long. When he drinks too much, it is just to
phone his family and someone comes to take him
home. (Christie 2004, p. 5)
On the contrary, in The House of Perfection
nobody knows Peter: he is just a man urinating in
front of children. The police must be called,
because this is a case of indecent exposure. In
Christie’s analysis, the inhabitants of this house
P
P
3468
never had a chance to get to know each other, nor
were they ever forced to establish forms of cooperation with their neighbors. In conclusion,
a “limited amount of knowledge inside a social
system opens the possibility of giving an act the
meaning of crime.” (Christie 2004, p. 5).
Substantial Law
Moving on to the abolitionist analysis of the
law, it is necessary to start from a general,
conventional premise. Juridical equality may be
described as everybody’s right to mobilize state
institutions for the protection and safeguard of
their wellbeing. In this perspective, it is, therefore, the right to mutual coercion. Disrespect for
the liberty of others amounts to the denial of their
freedom. The state intervenes to deny that denial
and restore the initial situation. Coercion is therefore legitimate in that it denies an act which has
denied the freedom of others. Abolitionists would
retort that such arguments might only be suitable
for societies in which equal access to the law is
complemented with equal access to resources.
Their critique of the law and the criminal justice
system is addressed to the iniquitous societies in
which we live, while critical responses are also
provided to other key assumptions, namely, that
the law addresses individual rather than collective actors, and that liability and non-liability can
be scientifically assessed.
Institutional intervention into problematic
situations aims to obliterate an ethics of shared
responsibility for conduct while affirming an
ethics of individual responsibility. In this way,
institutions can establish their monopoly over
the power to punish or pardon, namely, their
right to destroy or “repair” the individual
responsible. The weakening of any networks
contesting this right, be they neighborhood,
religious, political, or cultural aggregations of
people claiming their own right to deal with
problematic situations and responsibility, is crucial for that monopoly to be accomplished.
When the official power to punish or pardon
solidifies, penal sanctions as chief characteristics of the criminal law emerge.
Penal Abolitionism
It is only then that crime – as defined by the state
itself – becomes possible. It is only then that the
continuity of crime in some form becomes fundamental to the very existence of the state. And this
form is determined by what the state chooses to
sanction penally. (Kennedy 1976, p. 63)
The abolitionist stance, here, echoes some
aspects of conflict theory. Conflict theorists propose that law in the books and law in action favor
the interests of special groups, and that, therefore,
rather than being concerned with the explanation
of crime, theory should engage in explaining the
function of criminal law (Vold 1958; Chambliss
1969, 1975; Turk 1969; Quinney 1974; Akers
1997). Similarly, abolitionists do not ask why
some people commit crime, but rather why some
acts are defined as criminal? As succinctly stated
by Quinney (1970, p. 18): “criminal definitions
describe behaviour that conflicts with the interests
of the segments of society that have the power to
shape public policy.” From another, though adjacent, perspective, the criminal law itself is
described as devoid of rationality and replete
with conflicting principles, as its central features
derive from struggles taking place in previous
epochs. Such features took shape amid social
and political conflicts and, inevitably, bear the
mark of previous historical events: “The central
principles of the law are the site of struggle and
contradiction. . . The fate of law as a rationalising
enterprise is tied up with the nature of law as
a social, historical force” (Norrie 1993, p. 9).
Elements of Marxist analysis are detectable in
these formulations, and in general in all conceptions of law as expression of group conflicts. The
law, according to Marx, allows states to deal with
“social imperfections” as if they were “evil dispositions of the poor” rather than products of
specific economic and political arrangements.
Marx, however, defines the law as the form in
which the individuals of a certain ruling class
assert their common interests.
Abolitionist analysis, however, develops some
central tenets of conflict theory in an original
manner. Where conflict theorists seem to limit
themselves to the critique of the criminal justice
system as an expression of antagonistic values
and interests operating from above, abolitionists
Penal Abolitionism
reappropriate the very notion of “conflict” and
turn it into a critical tool to be utilized from
below. Nils Christie (1977, p. 1), for instance,
remarks that conflicts are hijacked by the criminal justice system and that criminology lends
a helpful hand in the process: “conflicts have
been taken away from the parties directly
involved and thereby have either disappeared or
become other people’s property.” Law specialists
are not the only ones to “steal” conflicts; conflict
theorists in criminology, in effect, do the same,
though in a different fashion: they turn interpersonal conflicts into class conflicts, thus again
taking them away from those directly involved.
Comparing two opposite scenes, an adjudication case in a small village in Tanzania and a case
dealt with in a juvenile court in England, Christie
provides the following description. In the former,
the contending parties, a man and a woman,
occupy the center of the judicial scene, they
make their claims, while their respective friends
and relatives take part without “taking over,” and
the audience intervenes with questions, information, and jokes. In the latter, the predominance of
professionals is visible, while the young defendant appears to be confused by those surrounding
him, whose role he does not understand and who
do not take any notice of him. “The truth is that,
for the most part, the business of the criminal
courts is dull, commonplace, ordinary and after
a while downright tedious” (Christie 1977, p. 2).
Christie (1998, p. 19) provides two pictures of
how rules are created. In the first, Moses descends
from the mountains carrying the rules engraved in
granite. “Moses was only a messenger, the
people – the populus – were the receivers, controlled from far above.” The other picture sees
a group of women gathering at the fountain,
around the well, or along a river, who while fetching water exchange information and evaluations.
The point of departure of their conversation will
often be concrete acts and situations. These are
described, compared to similar occurrences in the
past or somewhere else, and evaluated – right or
wrong, beautiful or ugly, strong or weak. Slowly,
some common understanding of the occurrences
might emerge. This is a process whereby norms
are created. It is a classical case of horizontal or
equalitarian justice. (Christie 1998, p. 119)
3469
P
The act whereby conflicts are hijacked by
professionals entails a specific construction of
reality hinging on an incident, narrowly defined
in time and space, whereby the individual is separated from the context in which the action takes
place. According to the abolitionist critique of
legal professionalism, those who act, in this
way, are isolated from the very incident involving
them, from the environment in which it occurs,
from their “friends, family, and the material substratum of their life world” (Hulsman 1991,
p. 23). Also, they become separated from those
people who, in that specific situation, feel victimized by certain acts. Hulsman focuses on the
culture and the organizational make up that
extrude people from their social context and artificially sets them against one another: “In this
sense the cultural organisation of criminal justice
creates ‘fictitious individuals’, and a ‘fictitious’
interaction between them” (Hulsman 1991).
The core function of such an organization is the
apportioning of blame, whereby events are
assembled and sanctions catalogued on the basis
of seriousness and gravity, respectively.
A hierarchy is, therefore, postulated and patterns
of acts and institutional responses to them are
artificially established. Events, in their turn, are
chosen among a limited range of conducts and
interactions, namely, those which the criminal
justice organization feels more comfortable in
explaining and evaluating. “No comparison is
made with events and behaviours outside that
range.” Patterns and hierarchies are drawn in
a universe which is far removed from the context
in which events took place, so that what appears
to be consistent within the criminal justice profession is experienced as inconsistent by those
who constitute the objects of that profession.
Hulsman remarks, in this respect, that “values
and perceptions in society are not uniform.”
And in this way he echoes conflict theorists
such as Sellin (1939), who emphasizes “culture
conflicts,” that is the differences between social
groups, their aspirations and behavioral models,
leading to acts defined as crime. But he adds that
the “program” for blame allocation typifying the
criminal justice system is a true copy of the doctrines of “the last judgment” and “purgatory”
P
P
3470
developed in certain varieties of Western Christian theology. The criminal justice system, in his
view, maintains the features of “centrality and
totalitarianism which belong to those theological
doctrines. Naturally, those origins – this ‘old’
rationality – are hidden behind new words:
‘God’ is replaced by the ‘law’ and the consensus
of the people” (Hulsman 1991, p. 24).
The notions of separateness and isolation,
applied by abolitionism to the offender, the victim, and the problematic situation characterizing
their interaction, are also mobilized by conflict
theorists, when their analyses focus on the law as
the realm of individualist philosophies and as the
reflection of egoistic interest. Abolitionism
enriches this approach with detailed descriptions
of contexts and events, suggesting that these are
subjected to a process of abstraction or outright
obliteration. It is the very cultural organization of
the criminal law that fosters this process,
a process whose “frame of reference,” as
Hulsman terms it, gives no room “for active participation and guidance for the parties involved.”
When the police operate within a criminal justice
frame they tend not to be directed by the wishes and
desires of the complainant, but by the requirements
of the legal procedure which they are preparing.
The complainant becomes a guide for their activities as “witness”. A witness is mainly a tool to bring
legal proceedings to a successful end. In
a comparable way the frame of court proceedings
precludes – or makes it anyway specially difficult –
for the victims to express freely their view on the
situation or to enter in an interaction with the person who is standing as a supposed offender before
the court. (Hulsman 1991, p. 24)
If classical conflict theory would mainly
examine structures and agencies while omitting
to look in any detail into their specific social
composition, abolitionism devotes particular
attention to such composition, particularly to the
increasing separate specialization characterizing
it. In brief, it is argued that each institutional
agency develops its own criteria for action, its
own ideology and culture which may lead it to
collide with other agencies. Agencies are bureaucracies pursuing internal goals, such as expand,
attenuate internal problems, monitor the
wellbeing of their members, and ultimately
Penal Abolitionism
ensure their own survival. “The process of
bureaucratisation and professionalisation within
the criminal justice system makes it a soulless
machinery” (Hulsman 1982, pp. 56–57).
Limits to Pain
Among the most critical traits characterizing
abolitionist theories and practices are those
concerned with the nature, function, and
philosophy of punishment. The infliction of
punishment might be justified if it contributed to
the maximization of happiness rather than the
maximization of pain. Hence the need for any
theorizing on punishment to consider its social
consequences. Those who defend institutional
coercion in the form of punishment may advocate
rehabilitative treatment, may value its general or
individual deterrent effect or its function of incapacitation aimed at prolific offenders. Abolitionists argue against such philosophies and their
practical outcomes. In Mathiesen’s (1990) view,
by contrast, prison has no defense whatsoever,
particularly when inspired by penal retribution,
whereby offenses are said to possess
a “punishment value” that can be translated into
a specific amount of “time.” Mathiesen argues
that time is only measurable subjectively, and
that the perception of its entity depends on one’s
proximity to those serving a prison sentence.
After crossing almost four centuries,
Mathiesen (1990, p. 33) identifies four major
components in the rehabilitative ideology which
have been remarkably constant through time.
“Work, school, morality and discipline have run
through the centuries as main pillars of thinking.”
While the emphasis may constantly shift from
one to the other component, rehabilitation itself,
understood as “return to competence” is rarely
achieved. Evidence of this failure is found by
Mathiesen in a variety of studies grouped in
three main areas. First, he refers to prison treatment studies, which yield largely uniform findings: the effects of treatment, in terms of
preventing recidivism, and regardless of strategy
or intensity, are generally poor. He then considers
the notion of treatment in relation to the actual
Penal Abolitionism
environment in which it is expected to be carried
out. The organization of prison life, the authoritarian principles governing it, along with the
bureaucratic arrangement stifling creative projects, provide a setting which is conducive to
everything but treatment. Moreover, prisons are
usually “overcrowded, run-down, and more or
less dangerous places to those who inhabit
them,” and if anything, they exacerbate the conducts of those they purport to treat (Mathiesen
1990, p. 41). Finally, he looks at a range of
studies devoted to prison as a social organization
or focused on prisoners’ communities, arguing
that inmates become soon “prisonized,” that is
they internalize the values and rules of a violent
and coercive environment and adopt a culture
which is impervious to change (Clemmer 1940).
Such culture, which protects the inmates from the
very setting that they inhabit, makes prison perform the function of a crime school, and prisoners
more or less immune to treatment or readjustment
programs.
Moving on to the discussion of deterrence as
general prevention, Mathiesen (1990) suggests
that the failure of this alleged function of imprisonment should prompt a dramatic reduction of
the use of imprisonment as a whole. Research
into the actual effectiveness of general prevention is so inconclusive that, when seen against the
harmful outcomes of custodial punishment, it
would be wise to limit the use of custody altogether, until its effectiveness is soundly proven.
General deterrence consists of a message
addressed to the general population, but it
would appear that something goes horrendously
wrong in the communicating process: “general
prevention functions in relation to those who do
not ‘need’ it. In relation to those who ‘need’ it, it
does not function” (Mathiesen 1990, p. 69). It is
for moral reasons, therefore, and due to lacking
evidence of the actual working of general deterrence, that the use of imprisonment should be
reduced to a minimum. There is, however,
a supplementary argument that would strengthen
Mathiesen’s reductionist stance, namely, that
deterrence makes no difference between guilty
and innocent individuals, in the sense that if
non-offenders were punished its goal would still
3471
P
be achieved. By punishing people at random,
irrespective of whether or not they have committed crimes, one may well assume that the general
population is discouraged from violating the law.
But, although acceptable from a utilitarian perspective, punishing the innocent would run counter the Kantian moral principle whereby
individuals must not be treated as mere instruments for the achievement of a social good: all
individuals should be treated as ends in themselves, in accord with their dignity and worth as
persons.
Considering how ideas of general deterrence
have increasingly replaced notions of treatment,
Christie (1982, p. 29) notes that at least the latter
“had a formally clear target: those receiving treatment.” The former, instead, chooses as a target
a vague entity described as the general population, which is composed of people who, if
deterred from one type of offense may resort to
another type; people who, in the face of a punitive
message launched in one country, may decide to
move to another country; or finally, people who
might find it hard to receive the punitive message
altogether. “Conceptually, as well as empirically,
ideas of general prevention or deterrence are thus
more cumbersome to handle than ideas of treatment” (Christie 1982, p. 29).
As for individual deterrence, the belief that
imprisonment amounts to a form of training for
criminal activities to come is persuasive enough,
in the abolitionist view, to discard this specific
official function of prison as inappropriate.
Schools of crime teach resentment and hostility
rather than respect for norms and others. In addition, it could be stressed that deterrent effects are
notoriously difficult to establish, “since they
involve counterfactuals: would these individuals
have acted differently had the threat of criminal
penalties been other than it was?” (Duff and
Garland 1994, p. 25).
The history of punishment may be interpreted
as an evolutionary process leading to distinct
stages characterized by increasing leniency.
Some commentators may describe the development occurring as one bringing a gradual
decrease in pain, a progressive shift from brutality to sweetness. Of course, there is difference
P
P
3472
between gruesome public executions and incarceration, between suffocating reclusion and custody in a single room, with cold and hot water and
a TV set. But does this indicate an objective
reduction of the pain caused by imprisonment?
I just do not know. Each form would have to be
evaluated according to its own time, by those
receiving the pain, in the framework of their usual
life and other people’s life, and in the light of what
they saw as their sins. I do not see how a scale could
be established. (Christie 1982, p. 9)
Law texts establish when punishments have to
be inflicted, while judges decide how long they
should last; neither examines their effects on
bodies and minds, the suffering produced, how
it feels. The belief in slow progressive humanization accounts for the lack of this type of information (Hulsman 1986).
Criticizing the humanization thesis, Hulsman
(1986) notes that the overall amount of punishment does not tend to decline, as the number of
convicted people per 100,000 of population
shows a strong cyclic movement. The trend of
convictions, in many countries in the industrialized world, has in fact for some time moved
upward. Hulsman finds the assessment of the
qualitative aspects of punishment more difficult.
It is true that the application of the death penalty
has been greatly reduced in recent centuries, and
in normal times has been abolished in many
countries. The same can be said, he adds, of
many forms of corporal punishment. It may also
be true that some progress has been made in
improving the regime in prison systems. He, nevertheless, suggests caution when judging qualitative amelioration and humanization. The amount
of suffering incorporated in legal penalties cannot
be measured on a scale of absolute values. “To
a large extent it consists of the difference between
the normal living situation of people and that
which is created by the intervention of the criminal justice system.” Hulsman is alluding to the
concept of less-eligibility whereby the conditions
in prison must be worse than the worst social
condition of people in liberty. Because the prison
system “has always drawn its clientele mainly
from the most disadvantaged sections of the population,” conditions in prison will reflect the
Penal Abolitionism
lowest standard of living experienced by this
social sector. Now, as “the living standards of
those same sections have in Europe improved
considerably in recent years,” one may presume
that prison conditions have improved accordingly, but the reality is that “improvement inside
prisons during the last 30 years do not appear to
have kept pace.” The conclusion is that “if this
supposition is correct, then the degree of suffering from the penal sanction has in a sense
increased” (Hulsman 1986, pp. 64–65).
The “evolution” of pain, moreover, is alleged
to have rendered physical punishment obsolete,
which is far from reality. Prison still entails forms
of corporal affliction: it degrades the body, it
deprives of air and light, it imposes humiliating
sanitary conditions, it causes diseases, and it produces sterile suffering.
Not all suffering is bad; some is beneficial, in that it
makes our consciousness develop while opening
up new existential paths, making us better humans
and getting us closer to the others. Imprisonment is
a type of suffering that does not create anything,
that does not generate any meaning. (Hulsman
1982, p. 59)
There are, finally, problems caused by
institutional intervention per se: custodial and
non-custodial measures alike erect.
obstacles to the development of social solidarity
and respect for differences that manifest themselves in “lived life”. . . Rather than alternatives to
custody we should devise alternative approaches
by which information is collected about the
needs of those involved in problematic situations.
(Hulsman 1982)
Intellectual Biographies
As remarked above, abolitionism does not possess one single theoretical or political source of
inspiration, but a composite backdrop from
which, wittingly or otherwise, it draws its arguments and proposals for action (Ruggiero 2010,
2011). The intellectual biography of Louk
Hulsman, for example, should be related to
some crucial passages in the Bible where mercy
is advocated while judgment and retribution are
rejected. The Gospels of Mark, Luke, and Paul
Penal Abolitionism
seem to provide an apposite theological
underpinning for Hulsman’s abolitionism,
which can also be assimilated to Saint Francis’
ecumenism and his view that thieves are not those
who steal, but those who do not give enough to
the needy. Radical theology or the theology of
liberation also provide significant sources of
inspiration. With Bakunin’s anarchism, Hulsman
shares the belief that the realization of freedom
requires that political action be conducted religiously. In some pages of Marx, Engels, Tolstoy,
and Hugo an echo is felt of Hulsman’s concepts
of redemption of punishment, self-government,
mercy, and pietas. Hulsman’s system of thought,
in brief, is shown to display a high degree of
syncretism.
An equal if not a higher degree of syncretism
shows the intellectual trajectory of Thomas
Mathiesen, whose materialist approach is punctuated by arguments for a pluralistic, interdisciplinary, sociology of law. The writings of Marx
and Engels constitute an ideal background for an
understanding of Mathiesen’s work, which on the
other hand draws on a number of other theoretical
sources. The focus on offenders and prisons, and
on social movements traditionally excluded from
orthodox notions of class struggle, makes
Mathiesen’s stance an implicit critique of classical Marxism. What constitutes the originality of
Mathiesen’s work is the coalescence of research,
action, and theorizing that characterizes his entire
career, as an academic as well as an activist.
Action is inherent in the research method
adopted, and those researched are the prime subjects involved in research as well as in action. His
radical analysis, therefore, is a tool which constantly translates knowledge about conflict into
collective praxis for those producing it.
In his work, Nils Christie stands in favor of
simplicity and intelligibility. He says that when
we write we should keep our favorite aunt in
mind, like Kropotkin said that anarchist literature
had to keep in mind the workers to whom it was
addressed. Christie’s system of thought echoes
anarchist theories of law and authority. His critique of legal professionalism is akin to libertarian arguments against the proliferation of laws,
while his appreciation of conflict as a resource
3473
P
brings to mind the anarchist idea that problems
within communities can only been resolved if
those involved possess sufficient autonomous
resources to do so. One of Christie’s arguments
is that communities and groups, irrespective of
their dimension, may find abolitionist experimentation possible only if the interactions within
them are highly frequent and intense. In this
sense, he expresses the purest of anarchist
notions, namely, that a better social life is experienced when communities develop social feelings and, particularly, a collective sense of justice
that grows until it becomes a habit.
Conclusion
On the nature and concept of crime abolitionism
engages in a debate that our ancestors Plato, Aristotle, and Spinoza saw as the distinction between
good and evil. Far from exhausted, such debate
continues in an array of disciplines ranging from
theology to psychology, from political philosophy
to social theory. The abolitionist critique of the law
contains some Kantian moral principles, in that
existing law is seen as using people as an instrument, a chance to signal the strength of transcendental and worldly authority. Abolitionists hold
the view that the law performs a transferral of
vendetta from the sacred into the judicial system
of the state, and that it is mainly addressed to
scapegoats, chosen violators among many, who
are at the same time “evil” and “sacred,” in that
they reproduce social life through sacrifice.
In abolitionist analysis, issues such as the designation of what constitutes crime, the intervention of the law, and the infliction of penalties that
take place in institutional settings, which are distant from the situations, are addressed. Within
such settings, punishment is the outcome of
a cognitive process whereby the institutions
make sense of events and claim that their response
is based on the knowledge of such events. Abolitionists, instead, claim that knowledge is mainly
achieved within precise contexts in which problematic situations arise, and it is for those involved
to “repair” such situations through the development of restorative collective capabilities.
P
P
3474
Related Entries
▶ Juvenile Diversion
▶ Monitoring and Evaluation of Restorative
Justice
▶ Restorative Justice
▶ Restorative Justice and Practice
Recommended Reading and References
Akers RL (1997) Criminological theories: introduction
and evaluation. Roxbury Publishing Company, Los
Angeles
Bianchi H (1991) Abolitionism in the past, present and
future. In: Lasocik Z, Platek M, Rzeplinska I (eds)
Abolitionism in history: on another way of thinking.
Institute of Social Prevention and Resocialisation,
University of Warsaw, Warsaw
Chambliss W (1969) Crime and the legal process.
McGraw-Hill, New York
Chambliss W (1975) Criminal law in action. Hamilton,
Santa Barbara
Christie N (1977) Conflicts as property. Br J Criminol
17:1–15
Christie N (1982) Limits to pain. Martin Robertson,
Oxford
Christie N (1998) Between civility and state. In:
Ruggiero V, South N, Taylor I (eds) The new
European criminology: crime and social order in
Europe. Routledge, London
Christie N (2004) A suitable amount of crime. Routledge,
London/New York
Clemmer D (1940) The prison community. Holt, Rinehart
and Winston, New York
Davis AY (2008) History is a weapon. The challenge of
prison abolition: a conversation between Angela
Y. Davis and Dylan Rodriguez. http://www.
historyasweapon.com/defcon/davisinterview.html
de Tocqueville A (1956) Democracy in America. Mentor
Books, New York
Duff RA, Garland D (eds) (1994) A reader on punishment.
Oxford University Press, Oxford
Gouldner AW (1975) For sociology: renewal and critique
in sociology today. Penguin, Harmondsworth
Honneth A (2007) Disrespect: the normative foundations
of critical theory. Polity, Cambridge
Hulsman L (1982) Peines perdues: le système pénal en
question. Le Centurion, Paris (interview with J. Bernat
de Célis)
Hulsman L (1986) Critical criminology and the concept of
crime. Contemp Crises 10(1):63–80
Hulsman L (1991) Alternatives to criminal justice:
decriminalization and depenalization. In: Lasocik Z,
Platek M, Rzeplinska I (eds) Abolitionism in history:
on another way of thinking. Institute of Social Prevention and Resocialization, Warsaw
Penal Justice and Social Injustice
Kennedy MC (1976) Beyond incrimination: some neglected
facets of the theory of punishment. In: Chambliss W,
Mankoff M (eds) Whose law? What order? A conflict
approach to criminology. Wiley, New York
Mathiesen T (1990) Prison on trial. Sage, London
Norrie A (1993) Crime, reason and history: a critical
introduction to criminal law. Weidenfeld and
Nicolson, London
Nussbaum M (2000) Women and human development: the
capabilities approach. Cambridge University Press,
Cambridge
Quinney R (1970) The social reality of crime. Little,
Brown, Boston
Quinney R (1974) Critique of the legal order. Little,
Brown, Boston
Ruggiero V (2010) Penal abolitionism. Oxford University
Press, Oxford
Ruggiero V (2011) An abolitionist view of restorative
Justice. Int J Law Crime Justice 39(2):100–110
Sellin T (1939) Culture, conflict and crime. Social Science
Research Council, New York
Sen A (2009) The idea of justice. Allen Lane, London
Simon JK (1991) Michel Foucault on Attica: an interview.
Soc Justice 18(3):23–37
Turk A (1969) Criminality and the legal order.
Rand-McNally, Chicago
Vold GB (1958) Theoretical criminology. Oxford
University Press, New York
Penal Justice and Social Injustice
Matt Matravers
Morrell Centre for Toleration,
University of York, York, UK
Overview
The connections between retributive justice,
broadly conceived, and distributive or social (in)
justice have long troubled criminologists, moral
and legal philosophers, and criminal justice practitioners. One reason for this is that the economically deprived make up a disproportionate
number of criminal offenders (and victims),
a fact made even more striking in many places
when overlaid with statistics about race. Another
is that inequalities within liberal democratic societies have risen since the 1970s; dramatically in
some places such the USA and UK. This rise has
accompanied a dramatic rise in punishment and
Penal Justice and Social Injustice
incarceration. Assuming inequality is a proxy for
distributive injustice, then, the world is becoming
less just and, in some places, more punitive.
These facts compel our attention in (at least)
two ways. First, there is a practical public policy
question. If crime rates are correlated to poverty
and/or inequality, then one way to address crime
is through social justice and the redistribution of
income, wealth, and opportunities. This is no
doubt sensible, but it is not the subject of discussion here. Rather, second, juxtaposing criminal
and distributive justice brings into focus
a normative question of whether social justice is
a precondition of legitimate criminal justice, or
more minimally, whether – and if so in what
ways – social injustice renders the imposition of
criminal liability and sanctions morally
problematic.
It is this normative issue that is discussed
below. Certainly, some philosophers concerned
with the justification of punishment have held
that conditions of social injustice – such as
those prevalent in much of the developed
world – render punishment at best problematic
and at worst morally impermissible. Thus,
Antony Duff (1986, p. 294) argues that punishment “will not be justifiable unless and until we
have brought about deep and far-reaching social,
political, legal, and moral changes in ourselves
and our society”. Similarly, Jeffrie Murphy
(1973) claims that existing social injustice fatally
undermines the only successful theory through
which punishment can be made permissible.
And David Bazelon (1976, p. 385) that “there
can be no truly just criminal law in the absence
of social justice.”
This entry will consider the arguments for the
proposition that social injustice renders the imposition of criminal liability and sanctions morally
problematic. It does not deal with the question of
whether lesser punishments are merited by those
who are disadvantaged by injustice, but rather
with the moral legitimacy of punishment as
a whole in circumstances of distributive injustice.
The first set of arguments focuses on theories of
punishment; the second on the responsibility of
the offender; the third on whether an unjust society has the standing to call offenders to account.
3475
P
In all cases, it is assumed that the background is
one in which the law is not egregiously unjust.
That is, the discussion does not encompass societies in which, for example, the law formally
discriminates against non-white persons with
respect to political and property rights. Such societies clearly raise issues about whether obedience
is owed by those against whom discrimination is
practiced and about the moral and legal liability
of those who break the law. However, the
assumption here is that the system has no such
formal rules, but that the social and economic
distribution of goods is unequal and unjust
(which will have implications for the value of
the formal rights enjoyed by all). In short, the
focus is on societies such as the USA and UK,
not on those such as South Africa in the era of
Apartheid.
Ideal Theories and Actual Practices
There are many theories that purport to justify
or render permissible state punishment. Such
theories tend to be ideal; that is, they develop
philosophical accounts of the justification of punishment assuming certain ideal conditions. The
question then arises as to whether those theories
remain compelling when applied to the actual
world. It is not possible to discuss each and
every such theory here, but consider some general approaches and how they might be affected
by social injustice.
Consequentialist Theories
Consequentialist theories justify punishment by
appeal to its good consequences. A system of
punishment is justified if it brings about better
overall consequences than any alternative system
including the option of doing nothing. In general,
consequentialists claim that punishment is justified because it secures good consequences
through individual and general deterrence, rehabilitation, and incapacitation. These are empirical
claims. If punishment were to fail to realize good
consequences, or fail to realize as good a set of
consequences as could be attained by an alternative policy, then it is not justified on this basis.
P
P
3476
Clearly issues of social justice will as a matter
of fact have an effect on any such empirical
claims. It may be best to divert resources away
from criminal justice and toward social justice if
that will be more effective in reducing crime (or
realize other good consequences that outweigh
any resulting bad consequences). It may be that
rehabilitation is better realized by providing
social and economic opportunities rather than
by punishment, and so on. These matters can
only be settled by empirical enquiry, not by philosophical armchair reflection.
Varieties of Desert
According to one variety of desert-based retributive justice, punishment is justified if and only if
it restores the fair balance of benefits and burdens
that the original crime disturbed. This so-called
fair-play theory was pioneered by Herbert Morris
(1968) and Jeffrie Murphy (1973).
In Murphy’s formulation, the argument relied
on a hypothetical social contract. In return for
security and assurance, self-interested persons
would accept an authority whose purpose was to
ensure those goods by upholding and enforcing
the agreed rules. That is, the citizen benefits from
the rule of law and all that comes with it and in
return undertakes to maintain society’s rules
through his obedience to them. Failure to abide
by the rules is a form of “free-riding”; the
offender gains the advantages of the system without having to “pay” for them through his own
self-restraint. In being punished – in
(hypothetically) agreeing to the system of punishment under which he is now made to suffer –
the offender pays a different price, one designed
to remove the extra benefit he has attained and
thus restore the status quo ante.
However, as Murphy was quick to point out,
the contract here is reciprocal. The authority of
the law – and so the duty of obedience on the part
of the citizens – stems from its being part of
a system of rules to which reasonable selfinterested persons could agree. Thus, if the system of rules is unjust – if, for example, it enforces
fundamentally unjust distributive arrangements
or discrimination between persons on the basis
of their skin color – then citizens no longer have
Penal Justice and Social Injustice
a duty of obedience and the state no longer has the
moral authority to punish breaches of the rules.
This account of punishment has been subject
to criticism and few (even among its early adherents) continue to adhere to it (for a sympathetic
contemporary account, see Dagger 1993; for
a summary of the main criticisms, see Matravers
2000, pp. 52–72), but the contractualist worry is
a genuine one. To the degree that society is conceived of as a cooperative endeavor in which
reciprocally agreed rules are enforced and
obeyed, there is a concern that the state’s failing
to keep its side of the bargain, so to speak, renders
illegitimate its enforcing of those rules.
A different, but related worry, affects the censure account of retributive punishment that
replaced the “fair-play” account as the leading
desert theory in the literature. According to
Antony Duff and others, the justification of punishment lies wholly, or in part, in its calling the
alleged offender to account and, should the
alleged offender prove to have performed the
wrongful act without excuse or justification, in
its expressing censure of the offender on behalf of
the polity. In imposing a penal sanction, the polity provides a vehicle for this censure and
a potential vehicle for the offender’s penitential
response.
In his early work (1986, pp. 291–299), Duff is
clear that social injustice renders this account
inapplicable to the real world. Such injustice is
a symptom, and cause, of the absence of
a genuine moral community on whose behalf,
and in whose name, the offender is called to
account. If there are no shared values, then the
offender cannot be censured in their name and
cannot in turn use punishment to restore herself to
those values and so to the community. Like
Murphy, Duff argues that the gap between the
ideal and the actual is so great as to render the
ideal unusable. However, in later writings, Duff
becomes less pessimistic although he still poses
the question of whether those who have been
systematically excluded from their fair shares of
society’s goods can be thought to be bound by the
laws of that society, a question he thinks likely to
be answered in the negative for many offenders
(2001, pp. 183–184). This motivates a different
Penal Justice and Social Injustice
worry, considered below, which is that social
injustice undermines the standing of the society
to call the offender to account.
Consequentialist, fair-play, and censure-based
theories of punishment, then, are each in their
own way threatened by conditions of social injustice. In all these cases, the fact of injustice
threatens the viability or applicability of the philosophical theory that purports to justify punishment in the real world. However, desert theorists
(in particular) also have a different worry about
social injustice; this is not that it renders the
theory – that punishment is justified because and
to the degree that it is deserved – inapplicable, but
that it undermines the deservingness of individual
offenders; it renders previously unjustly treated
citizens less blameworthy and so less deserving
of punishment.
Blameworthiness and Distributive
Injustice
The arguments of the section above suggest that
social injustice might challenge the legitimacy of
the law, the account of political obligation, or the
justification of the system of punishment as
a whole. A (perhaps) more familiar worry is that
in some way or other, defendants from indigent,
or unjustly disadvantaged, backgrounds are less
to blame than are others. This section considers
this worry in relation to cases of excuse and of
justification.
Rotten Social Background
Perhaps the most famous debate in relation to
doing penal justice in conditions of social injustice arose out of a 1973 case, US v Alexander.
Two black defendants were charged with the
murder of two white victims. The defendants
had shot the victims following an altercation at
a fast-food restaurant during which the victims
had called the defendants “God-damned niggers.” On behalf of one of the defendants, who
was not mentally ill according to any recognized
diagnostic categories, a psychiatrist nevertheless
testified that the defendant suffered from an
“emotional illness” that was rooted in his
3477
P
upbringing in a very poor and deprived neighborhood. This upbringing had left the defendant with
an acute sensitivity to racial discrimination such
that when he had been subject to racist abuse, he
had felt “an irresistible impulse to shoot” (for
useful accounts of the issues here, see Green
2011; Heffernan 2000). The judge allowed the
jury to consider the mental capacity of the defendant, but mindful of the possibility that the jury
might be moved by pity or sympathy for the
defendant given his upbringing, the judge went
on to say to the jury that it ought not to be
concerned “with the question of whether a man
had a rotten social background.” The defendants
were convicted.
The court of appeals upheld the conviction,
but one of its judges, David Bazelon, in part
dissented. He argued that there was a relevant
causal claim linking the defendant’s rotten social
background and the defendant having no “meaningful choice when the racial insult triggered” his
response in the restaurant (United States
v Alexander, 471 F.2d 923: 960). The truth of
such a causal claim would need to be considered
by the jury, but Bazelon wrote confidently of the
social science evidence in its support and of the
need to consider what to do with such defendants
should they prove innocent, but dangerous
(holding out the options of therapy or preventive
detention).
The idea of a defense of rotten social background was taken up by Richard Delgado (1985).
Delgado argued that the law should recognize
a new excuse based on extreme poverty and deprivation. As an excuse, the claim would not be
that the conduct was justified. Rather, the defendant would have to prove that at the time of
acting, her conduct was the result of behaviors
shaped by her background such that she met
(an extended understanding of) normal excusing
conditions: For example, she was unable accurately to understand the nature and consequences
of her conduct; she was suffering from involuntary rage; or was otherwise unable to control her
conduct.
It is important to note that neither Bazelon nor
Delgado argues that the fact of an (unusual)
causal history to the agent’s conduct is enough
P
P
3478
to excuse. All actions have causes, some are more
unusual than others and some evoke sympathy,
but as Stephen Morse insists (2000, p. 130), “causation is neither an excuse per se nor the equivalent of hard choice (so-called compulsion), which
is an excusing condition.” Rather, the argument is
that the particular background of deprivation is
such as to render some defendants excused at the
time of their conduct because it interferes with
those ordinary capacities on which the law
depends when holding persons responsible. In
this, a defense of rotten social background
would come close to the defense of “battered
woman syndrome,” in which it is argued that
a history of violent abuse can render a woman
incapable of (legally) responsible behavior
even at some later point when her conduct
seems otherwise to be chosen.
Justification
Proponents of the excuse of rotten social background accept that the defendant’s conduct was
wrongful; they merely argue that it may be
excused. However, another response to social
injustice is to regard some conduct performed
by the disadvantaged as justified. Consider
a case imagined by Jeremy Waldron (2000,
pp. 104–105). A homeless, unemployed, destitute
man in a society with little or no welfare assistance scavenges in a park. He comes across
a half-eaten hamburger and begins to eat it. However, the person who had bought the hamburger,
having eaten half, was planning to feed the rest to
the birds. He complains to the police and the
destitute man is arrested and charged with theft.
The defendant claims the justification of
necessity. Before considering this, it is worth
noting how this differs from the claim of excuse
in rotten social background. In the case of
excuses, the conduct is wrong, but excused. In
the case of justification, the conduct itself is not
wrongful. Thus, in rotten social background
cases, there is no obvious restriction on the possible conduct to which the excuse might attach or
on the range of victims of the conduct. As is clear
from the Alexander case, the excuse potentially
covers serious wrongs such as homicide and it
would have made no difference to the case
Penal Justice and Social Injustice
whether the victim was poor or rich. In the justification case under discussion, it is clear that the
conduct that is justified is severely restricted. The
defendant might claim justification for picking up
the hamburger, but he could not do so for killing
the owner so as to get the hamburger (no matter
how hungry he was). Similarly, the agent could
not claim justification for taking the last morsel of
food from a similarly situated indigent victim.
These differences, of course, reflect the different
structures of excuses and justifications.
Although Waldron is clear that there is a moral
case for the justifiability of the destitute man’s
actions, he is confident that any legal system
dedicated to the protection of property rights
will reject the necessity defense in this instance.
To establish necessity in the common law tradition, the defendant must show imminent threat;
that the conduct would be effective in ameliorating that threat; that no alternative legal means to
do the same was available; that the conduct was
less harmful than the harm threatened; and that
the circumstances that gave rise to the threat are
not attributable to her.
The threshold for the use of the necessity
defense is very high. To rebut it, the prosecution
would only have to show that the indigent eater of
the hamburger was not in danger of death or
serious injury, or that he could have gone from
the park to the local charitable shelter, or that his
condition was due to his culpable failure to find
work. In short, necessity and long-term poverty
do not map easily onto each other. Moreover, as
Waldron insists, systems of law protect the
orderly and predictable governance of property,
an order that cannot be subject to disruption
whenever someone feels the need to appropriate
something of someone else’s.
The arguments in this section demonstrate the
difficulties in matching concerns about social
justice with issues in criminal justice. In one
way, this is right; not all poor people commit
crimes (and not all crimes are committed by the
poor) and it would be demeaning to large sections
of the population to pretend otherwise. If poverty
or disadvantage is to be relevant to judgments of
the offender’s liability, then it has to be shown
either that the offender’s background relevantly
Penal Justice and Social Injustice
impaired him at the time of acting, or that his
actions were legally justified. It is likely that these
hurdles will not in fact be often cleared.
However, this seems to leave the issue
unresolved. It does matter that society is distributively unjust and it does trouble us that those
who systematically suffer most from injustice
also disproportionately appear in the criminal
justice system as victims and offenders.
The State’s Standing to Hold
the Offender to Account
The focus has so far been on the applicability of
theories of punishment to the real world and on
the offender. Social injustice raises genuine questions for the first, but the focus on the blameworthiness of the offender seems either to lead in
directions covered by the existing criminal law
(with respect to excuses and justifications) or to
risk widening the scope of the enquiry in ways
that potentially demean the poor and disadvantaged. Returning to the original issue, however,
suggests an alternative approach in which the
focus is not on the offender, but on the state.
Assume – reasonably enough in most cases –
that the offender acts culpably. The law
prohibiting the conduct is reasonable, the offense
is wrongful, and the offender is blameworthy for
it. To take an example, a disadvantaged offender
defrauds the state by claiming unemployment benefits while working on the black market for cash.
In response to cases like this, Antony Duff
asks not whether the offender is responsible, but
to whom he is responsible. That is, criminal
responsibility is a relational concept; an alleged
offender is responsible for some conduct to someone or some institution. Thus, a teacher is responsible for marking her students’ essays and that
responsibility is owed to her students. Moreover,
the teacher can call her students to account for
failing to submit their essays, and so on. In short,
with respect to essays (and no doubt other things),
the teacher has standing.
The question Duff poses is whether a state’s
maintaining of chronic and systematic injustice
removes its standing to call offenders to account.
3479
P
As the paper of one of his papers suggests –
“‘I Might Be Guilty, But You Can’t Try Me’:
Estoppel and Other Bars to Trial” (2003), Duff
is skeptical that unjust societies retain the standing to call to account those whom they have
treated unjustly.
This is an attractive response to the issues. It
does not deny the responsibility of indigent
offenders, and it surely captures a plausible intuition (and mirrors many real life situations in
which persons may think, “I may be wrong, but
who are you to criticize?”). There is surely something wrong in being held to account by an institution that is itself in violation of the values to
which it appeals in its calling the alleged
offender to answer for her actions (cf Sykes and
Matza 1957 on “The Condemnation of the
Condemners”).
However, as stated, the account is underdeveloped. Two questions arise: First, does the unjust
state lose the standing to judge all offenders and
offenses (even those unrelated to the injustice)?
Second, does the unjust or immoral state lose all
standing across all dimensions? To see this, first
contrast the unemployed offender mentioned
above who defrauds the state by cheating on his
social security payments and the same person
who beats his wife. It may be thought that the
state lacks the standing to call him to account for
fraud, given a background of systemic injustice,
but why should this affect its standing to call him
to account for assault? Second, consider the
standing of the UK state to call a well-off defendant to account for insurance fraud, given recent
revelations that the same state had deliberately
misled its citizens in the run-up to the second
Iraq war and had, when fighting terrorists in
Northern Ireland in the 1970s, connived in the
extrajudicial execution of (at least one) political
agitator (for a summary of these criticisms, see
Matravers 2006).
In short, the argument from standing seems to
do too much. It is in that sense reminiscent of
Murphy’s argument that legitimate punishment is
impossible in a real world in which distributive
injustice is prevalent. Yet, it undeniably captures
something of the worry that many people have
when confronted by the facts of distributive and
P
P
3480
retributive justice and, if nothing else, it should
evoke in all citizens what Duff (2003, p. 259)
calls “a properly humble and cautious spirit
which recognizes how far from clean our collective hands and consciences are” with respect to
penal justice.
Conclusion
Not all the poor suffer from the same injustices,
not all of them commit crimes, and not all of
those who commit crimes do so in similar circumstances (for a nuanced discussion of a “caseby-case” approach to social injustice in relation
to the criminal law, see Green 2011,
pp. 364–373). Nevertheless, the poor and disadvantaged are more likely to be victimized by
crime, they are more likely to engage in criminal
activity, and they appear disproportionately
among those arrested and convicted (Green
2011, pp. 354–355; Western 2006). If one were
to visit any prison throughout much of the world,
one would be confronted with populations predominantly made up of young men from poor and
disadvantaged backgrounds (in the USA, young
black men from the same) (Wacquant 2009). At
the same time across much of the developed
world, the inequalities between the richest and
the poorest in societies are growing. These facts
compel attention not just in themselves, but
because in punishing the poor, “we run the risk
of compounding the sins of socio-economic
injustice with those of retributive injustice”
(Green 2011, p. 376).
The purpose of the arguments above is to show
that confronting this possible double injustice
from within the existing structures of the criminal
law is deeply problematic. Some disadvantaged
offenders will suffer from “abnormality of mind”
(or similar) consequent upon their disadvantage,
but this will be a small minority and, even among
those, difficult to establish in court. Few, if any,
will be able to claim a justification such as necessity. Ideal theories of punishment may not apply
to the world we confront, but absent some other
form of social organization, the necessity to
punish will persist.
Penal Justice and Social Injustice
Perhaps the route out of these difficulties is to
look beyond the resources of the criminal law and
of philosophies of punishment and to political
philosophy more generally. Much contemporary
discussion of justice revolves around the question
of how we might distinguish between those
things – actions, omissions, character features,
and so on – for which the agent is responsible
and on the basis of which he might legitimately
claim the benefits or be made to bear the burdens
and those things that are merely chance or
circumstance – things that, so to speak, simply
happen to her – on the basis of which she may
deserve assistance or compensation. Of course,
while in many cases, poverty and injustice simply
happen to people, the decision to commit an
offense may fall into the category of choice rather
than chance. Nevertheless, asking “when is it fair
to make people bear the burdens or enjoy the
benefits of their conduct?” may open up avenues
of enquiry that might inform the criminal law
and, if not that, at least our attitudes to those
who fall liable to it (for an initial development
of such a strategy, see Matravers 2007: Chap. 3;
Scanlon 1998: Chap. 6). That said, given the
extent of disagreement among political philosophers as to the answers to these questions, criminal law issues such as those discussed above
remain both alive and worthy of attention.
Recommended Reading and References
Bazelon D (1976) The morality of the criminal law. South
Calif Law Rev 49:385–405
Dagger R (1993) Playing fair with punishment. Ethics
103:473–488
Delgado R (1985) “Rotten Social Background”: should
the criminal law recognize a defense of severe
environmental deprivation?. Law Inequal 39:12–23
Duff RA (1986) Trials & punishments. Cambridge
University Press, Cambridge
Duff RA (2001) Punishment, communication, and
community. Oxford University Press, Oxford
Duff RA (2003) “I Might Be Guilty, But You Can’t Try
Me”: Estoppel and other bars to trial Ohio State. J Crim
Law 1:245–259
Green S (2011) Just deserts in unjust societies. In: Duff
RA, Green S (eds) Philosophical foundations of the
criminal law. Oxford University Press, New York,
pp 352–376
Penal Paradigms: Past and Present
Heffernan WC (2000) Social justice/criminal justice. In:
Heffernan WC, Kleinig J (eds) From social justice to
criminal justice: poverty and the administration of
criminal law. Oxford University Press, New York/
Oxford, pp 47–83
Matravers M (2000) Justice and punishment: the rationale
of coercion. Oxford University Press, Oxford
Matravers M (2006) ‘Who’s Still Standing?’ A comment
on Antony Duff’s preconditions of criminal liability.
J Moral Philos 3(3):320–330
Matravers M (2007) Responsibility and justice. Polity
Press, Cambridge
Morris H (1968) Persons and punishment. Monist
52:475–501
Morse SJ (2000) Deprivation and desert. In: Heffernan
WC, Kleinig J (eds) From social justice to criminal
justice: poverty and the administration of criminal law,
Practical and professional ethics series. Oxford
University Press, New York/Oxford, pp 114–160
Murphy J (1973) Marxism and retribution. Philos Public
Aff 22:17–43
Scanlon TM (1998) What we owe to each other. Harvard
University Press, Cambridge, MA
Sykes GM, Matza D (1957) Techniques of neutralization:
a theory of delinquency. Am Sociol Rev
22(6):664–670
Wacquant LJD (2009) Punishing the poor: the neoliberal
government of social insecurity. Duke University
Press, Durham
Waldron J (2000) Why indigence is not a justification. In:
Heffernan WC, Kleinig J (eds) From social justice to
criminal justice: poverty and the administration of
criminal law. Oxford University Press, New York/
Oxford, pp 98–113
Western B (2006) Punishment and inequality in America.
R. Sage Foundation, New York
Penal Paradigms: Past and Present
Leslie Sebba
Institute of Criminology, Faculty of Law,
The Hebrew University, Mt. Scopus,
Jerusalem, Israel
Overview
Accounts, critiques, and theoretical conceptualizations of penal systems reveal that the characteristics of these systems may vary in very significant
ways. Penological histories show how the “classical school” of the late eighteenth and early nineteenth centuries (epitomized in the famous essay
of Cesare Beccaria’s Of Crimes and Punishments)
3481
P
believed that the primary rationale for the penal
system was general deterrence and advocated that
this should be reflected in sentences proportionate
to the seriousness of the offence. They then
describe how this ideology was in the latter half
of the nineteenth century replaced by individualized sentencing – which in turn gave way a century
later to a “just deserts” philosophy (again proportionate). These successive approaches are sometimes referred to as paradigms.
A focus on the discourses prevalent in the
literature on contemporary criminal justice, however, indicates that in fact a variety of competing
models are identified or advocated – perhaps
most famously illustrated by Herbert Packer’s
seminal article outlining two apparently alternative models – “due process” and “crime control.”
This entry will attempt to provide an introductory
overview of these developments – both the historical and the contemporary. The usefulness of
these approaches will be considered, as well as
confusions arising from the terminology
employed, given that neither the terms “paradigm” nor “model” are always used consistently.
Introduction
While penal systems, whether formal or informal,
appear to constitute an integral part of every society (and thus presumably, as argued by Durkheim,
fulfill some significant social function), they differ from each other in almost every conceivable
aspect: the forms of conduct punished, the procedures for determining guilt, the aims of punishment, its modalities, and its severity. Moreover,
individual penal systems fluctuate and indeed
undergo radical changes over time. In elaborating
on these characteristics, traditional penological
texts (e.g., Barnes and Teeters 1959) presented
largely descriptive accounts of the historical
development of penal systems, mainly in Europe
and the Americas – while there has also been
a fascination with ancient and “primitive” systems (Diamond 1971). Social theorists since
Durkheim have offered generalized theoretical
explanations for the form and severity of
punitive practices – a development analyzed by
P
P
3482
Garland in Punishment and Modern Society and
applied to contemporary society in his Culture of
Control (Garland 1990, 2001). More recently,
there has been a wave of research endeavoring
to explain the wide differences among nations
in the severity of their sanctioning policies
(Tonry 2007).
What many of both the descriptive accounts of
penal systems and the more critical analyses share
in common is that they tend to view the penal
system as – indeed – a “system” with a coherent
purpose and philosophy, thereby implying
a degree of interconnectedness and compatibility
among its different component elements and
shared values among the various actors. It is also
common to identify somewhat dramatic ideologically driven policy transformations believed to
have taken place at certain periods of time. Thus,
historical accounts will invariably refer to the
eighteenth century as the era of classical criminology (and in particular penology), the late nineteenth and most of the twentieth century as the
era of positivism, with the 1970s onward as being
dominated by the “just deserts” model. The penal
systems identified under these different headings
are shown to diverge in fundamental ways, in
particular having different sentencing rationales,
thereby giving rise to their designation as alternative “penal paradigms.” One purpose of this entry
will be to elaborate on and evaluate such historical
presentations.
It is suggested here that such analyses may be
viewed as being complemented by contemporary
discourses relating to the characteristics and
ideologies of the present-day criminal justice
system, which sometimes echo the typification
of past ideologies but with a tendency to describe
a multiplicity of trends or ideologies operating
simultaneously or in competition.
Such a connection between past and present is
indeed made by Jock Young in his much-cited
essays “Thinking Seriously about Crime” and
“Incessant Chatter: Recent Paradigms in Criminology” (Young 1981, 1994) which incorporate
critiques of the traditional presentation of the
historical paradigms while using this concept to
analyze a number of more recent theoretical
approaches. However, Young also uses the term
Penal Paradigms: Past and Present
“model” in this context, and this is the term more
frequently adopted in current penological discourse when referring to alternative penal
options. Much of the discourse relating to criminal justice models, however, seems to have been
inspired by the dichotomy developed by Herbert
Packer (1964) in his well-known law review
article “Two Models of the Criminal Process,”
and indeed the relevant literature frequently
acknowledges such a debt.
This entry will provide a brief survey both of
the historic “paradigms” and of the contemporary
“models,” followed by some reflections on the
nature of their contribution and related conceptual
issues. The survey will be preceded, however, by
a discussion of the terminology employed here and
elsewhere in the literature in this field.
Paradigms and Models in a
Criminological-Penological Context
The term paradigm was developed in the context
of the hard sciences to denote the prevailing
understanding or interpretation of natural phenomena. It was popularized by Thomas Kuhn in
his Structure of Scientific Revolutions (Kuhn
2012) when analyzing how changes come about
in such understandings on the part of the scientific
community and came to be applied also to the
social sciences. While Kuhn’s analysis is essentially restricted in its application to phenomena
which may be studied empirically and is thus not
applicable to many concepts and ideas pertaining
to punishment, it has now a more widespread
usage, connoting “a cluster of concepts such as
assumptions, values, practices and methodologies shared by a community of researchers in
a given discipline” (New World Encyclopedia,
online). There may, however, be competing paradigms at any given time, where the evidence is
equivocal and/or the scientific community
divided in their views and interpretations.
In this context it should be noted that while
paradigms relate to “scientific findings” and are
thus closely linked to positivist conceptions, each
paradigm is rather a social construct, describing the
collective understandings of a given community as
Penal Paradigms: Past and Present
perceived by an observer (or participant). In the
social sciences where “evidence” tends to be less
hard than in the natural sciences (and the role of
values more salient), it is probable that there will be
less consensuality, thereby increasing the possibility of “competing paradigms.”
The term “model” has similarly acquired a
wider usage than its original meaning of a (usually
small-scale) replication of a product for the purposes of demonstrating or testing its operation, or
“providing insight into the consequences of
a decision” (Chaiken et al. 1975: v). It is simply
used to describe the main features and mode of
operation of a system – whether existing or proposed. Both the expressions “paradigm” and
“model” implicitly refer to a system or framework
which comprises a number of components operating as a cohesive whole – and, where relevant,
sharing the same values and ideology. Only the
paradigm, however, has a connotation of exclusivity – although also that such exclusivity may be
ephemeral.
Other related terms used in the penological
context include penal strategies (Garland 1985)
and styles of social control (Black 1976). Young
(1981) not only used both the terms “models” and
“paradigms” but also Weber’s concept of ideal
types, which suggests an idealized account of the
institution in question which is unlikely to be
found in that form in real life. Recent analyses
of Packer’s models have considered their relationship with Weber’s ideal types (see below).
It should also be noted that the first two paradigms to be discussed here, the “classical” and
the “positivist” paradigms, have generally been
described in the literature as schools. Finally,
certain distinct perspectives on criminological
issues today may be referred to as criminologies,
such as “feminist criminology” and “green
(environmentalist) criminology.” The relative
significance and usefulness of these terms has
yet to be fully explored.
Historic Paradigms
The three historical paradigms to be considered
here are set in somewhat different theoretical
3483
P
frameworks. While the classical school, at least
as reflected by its most notable representative,
Cesare Beccaria, focuses almost exclusively on
the formal institutions of the criminal justice system (in this respect resembling the contemporary
models to be considered later in this entry), the
positivist school is much wider in its concerns,
being equally – or indeed primarily – concerned
with the etiology of crime. By contrast, the third
of our historical paradigms, just deserts, has the
narrowest focus, being concerned exclusively
with the purposes and procedures of sentencing.
The Classical Paradigm
This paradigm is identified with the European
enlightenment movement and the emergence of
the ideas which were to become identified with
liberal theory and English constitutionalism
(Young 1981) as an antidote to the monarchical
despotism prevailing in much of continental
Europe. Building on Montesquieu’s doctrine of
the separation of powers and Rousseau’s social
contract theory, Beccaria (1764/1995) in his seminal essay advocated a strict separation of powers
and equality before the law in order to counter the
arbitrary practices prevalent in certain contemporary European regimes which occurred as a result
of executive intervention in trials and sentencing.
Beccaria advocated uniform penalties, proportional to the seriousness of the offence, which
would strengthen their deterrent power: rational
persons would refrain from committing crimes
attracting high penalties, especially if law
enforcement was systematic, certain, and immediate. The attribution of free will and rationality
to potential offenders and correspondingly the
importance of deterrence as the primary aim of
punishment are thus perceived as identifying
characteristics of the classical paradigm.
This paradigm may also be seen to be linked to
the development of the modern prison. While
Beccaria was advocating the prison as a more
effective deterrent than the death penalty, his
contemporary John Howard was documenting
the practices of the then local prisons, which led
the way to the enactment in 1779 of the Penitentiary Houses Act in order to establish a new and
rationalized form of prison – the penitentiary.
P
P
3484
As is well known, this system, based upon uniform
treatment in identical cells which it was assumed
would lead to penitence and reform, was ultimately pioneered in the United States a few
decades later. Another key figure identified with
this paradigm was Jeremy Bentham, also famous
for developing his own “rational” model of prison
development – the Panopticon (Semple 1993).
The Positivist Paradigm
Concepts of equality and uniformity in penal
policy were challenged during the second half
of the nineteenth century by the rise of the social
sciences and the claim of Auguste Comte that in
the course of human development “metaphysical” or abstract thinking was to be replaced by
the “positivist” stage – and the application of
purportedly scientific methodology and empiricism. One outcome of this was Lombroso’s theory of delinquent man. The focus on human
differences resulted in a penchant among positivist criminologists and their disciples in the field
for the classification of offenders for treatment
purposes. Such techniques derived encouragement from the mid-century experiments with the
“progressive stage system” (pioneered in Ireland)
whereby prisoners advanced through the prison
system and into the community in the light of
their conduct.
Under the influence of positivism, the reformation of the offender became the declared purpose of punishment rather than deterrence, and
the ideal of standardized penalties proportional to
the seriousness of the offence as posited by
Beccaria came to be replaced by the individualization of punishment, as expounded in the treatise bearing this name written by Raymond
Saleilles (1911). The new approach to sentencing
was epitomized by the principle of the indeterminate sentence widely adopted in the USA
whereby terms of imprisonment imposed by the
courts were open-ended, the actual duration of
the sentence to be determined by a parole board
in the light of the prisoner’s progress.
This “individual treatment model” of sentencing, as it developed towards the end of the nineteenth century and throughout much of the
twentieth century, was characterized by the
Penal Paradigms: Past and Present
creation of specialized penal and judicial institutions ostensibly adapted to the needs of the
offenders in question, such as reformatories and
juvenile courts for younger offenders and institutions for dangerous psychopaths, the involvement
of behavioral experts such as psychiatrists and
probation officers in decision-making, and the
increasing recognition of the helping professions
as agents of offender rehabilitation in both closed
institutions and in the community. Moreover,
while the determinist conceptualization of
human behavior and the focus on the individual
led to the application of more flexible criteria of
criminal responsibility, some versions of the
Social Defense ideology (a branch of the positivist movement which focused on penological
issues) proposed also to dispense with the legal
requirement of a “guilty mind” (cf. Ancel 1965).
The threshold requirement for treatment would
be the personality of the alleged offender – the
main criterion being his or her dangerousness.
The Just Deserts Paradigm
The positivist ideology with its emphasis on rehabilitation meshed well with the welfarist ethos
that was widely supported during the period following the Second World War. However, if the
individualized treatment paradigm was most
widely accepted (at least in the prevailing rhetoric) in the USA, it was there too that it was the
most forcefully rejected during the 1970s. Critiques of this model focused on the inequality and
arbitrariness of the system, in particular in the
context of parole decision-making (American
Friends Service Committee 1971). Critics
included not only radicals and liberals but also
conservatives (cf. Blomberg and Lucken 2000).
Adding fuel to the fire was the critique of
sentencing disparities deriving from the wide
discretionary powers held by the judges
(Frankel 1973), and – perhaps most influential
of all – the empirical claim that the effectiveness
of the treatment methods on which the system
was based was unproven (Martinson 1974).
The structure which was to take account of
these critiques, as formulated most articulately
in Doing Justice, the report published by Andrew
von Hirsch (1976), was a return to fixed and
Penal Paradigms: Past and Present
proportional sentencing. While outwardly this
bore some resemblance to the system advocated
by the Classical School (see above), its underlying philosophy was not deterrence, or indeed any
other consequentialist purpose – as none had been
shown to be effective by the empirical evaluations, but desert, a contemporary version of the
historic aim of retribution. According to this doctrine, the penalty would match the seriousness of
the offence as measured by the harm inflicted and
the culpability of the offender – and thus become
standardized. The mechanism whereby this was
to be achieved was the establishment of
a sentencing commission which would issue
detailed guidelines to the sentencing judges,
whose discretion would be severely curtailed.
Since sentences were to be for fixed terms, the
discretionary power of the parole boards would
be abolished.
Support for such reforms in the USA was such
that they were adopted in principle by the federal
system and around half of the states – as well as
some overseas countries, as documented in other
entries. The concept of “structured sentencing,”
whether by guidelines or otherwise, attracted
even wider support. The “individual treatment
model” as the foundation of the penal system
had in general been rejected.
Historical Paradigms: A Critique
The preceding overview suggests that in the
context of this entry, the term “paradigm” is
most appropriately (or least inappropriately)
used in the context of the positivist school, since
the policies advocated in accordance with this
approach were the most rooted in the empirical
understanding on the part of its adherents of the
world about them. It was indeed this characteristic that ultimately contributed to the demise of
this paradigm when research findings suggested
that the ability of decision-makers to make
accurate prognoses and identify successful treatments was very limited.
There are, however, other problems with what
Young has described as the typical textbook
presentation of a linear development of
a succession of paradigms. Young is correct in
pointing out that despite the enormous impact of
3485
P
positivist thinking on the twentieth century
penological
discourse,
criminal
justice
structures – particularly in European countries –
remained throughout the twentieth century
strongly influenced by the tenets of eighteenth
century classicism. The nature and seriousness
of the offence remained the primary consideration in determining the sentence, with the
characteristics of the offender (other than past
convictions) playing a secondary role. Contemporaries therefore referred to the model adhered
to in Europe during the course of the twentieth
century Europe as “neoclassical.”
Today too it is hard to identify a dominant
paradigm. As noted above, the parameters of the
third of the paradigms referred to, the desert
paradigm, were somewhat narrowly drawn
(being confined to sentencing practices), and in
any case, most contemporary western systems
have been strongly influenced by consequentialist ideologies – incapacitation, deterrence and
even rehabilitation. It should also be recalled
that the concept of the paradigm refers in principle not directly to the practices of the actors in the
field (in this case the criminal justice agencies)
but to the perception of these practices – or
perhaps of the ideologies which are thought to
motivate them – on the part of the relevant disciplinary group. Thus, radical criminologists such
as Young, who perceive the criminal justice system primarily in terms of control, if not exploitation, of the powerless by the powerful, have since
the 1960s been offering alternative paradigms to
explain its operation.
It may always have been the case that it was
more problematic to identify monolithic trends
(such as are implicit in the concept of the paradigm) in one’s own era, when different views or
nuances will be more apparent than when looking
back in history. However, this problem has surely
been aggravated by the dichotomization of social
scientists into positivists and social constructivists, further compounded by the fragmentation
caused by postmodern developments – leading
Garland and Sparks (2000) to opine that further
paradigms in the (consensual) Kuhnian sense can
no longer be anticipated. These authors point to
the salience of crime in the contemporary politics
P
P
3486
of neoliberal societies. This has had the effect of
reducing the relevance of the prevailing views
of academia and the professions, which have
perhaps themselves become more fragmented
and politicized.
Criminal Justice Models
The Legacy of Herbert Packer
Unlike the paradigm, the concept of a model does
not imply consensus and exclusivity but rather
variation and experimentation. While criminologists have sometimes used them in the context of
empirically oriented work (e.g., Feeley 1973;
Chaiken et al. 1975), much of the modern discourse on criminal justice models has been
devoted to normative and/or ideological issues,
having been inspired, as noted above, by Herbert
Packer’s seminal article “Two Models of the
Criminal Process” (Packer 1964 – subsequently
expanded into a book – Packer 1968). The article
followed a period of activism in the area of civil
rights and criminal justice on the part of the US
Supreme Court headed by Chief Justice Warren,
but Packer was doubtless concerned by the
minority dissenting views which were to become
the majority in the era of Chief Justice Burger. It
seemed to Packer that the Court and indeed other
criminal justice agencies were being required to
choose (or at least prioritize) between two alternative types of process which he identified as due
process and crime control, respectively,
representing alternative sets of values and
priorities.
A system based upon due process he perceived
as being analogous to an obstacle course, in that it
sought to ensure protection of the rights of the
innocent (and to risk the acquittal of the guilty) by
imposing restraints on the prosecution. Under the
crime control model, on the other hand, the system was comparable to an assembly line in
a factory – designed to maximize the numbers
who would be processed through the system to
conviction and punishment in minimum time.
This was to be achieved by focusing on the
investigation stage and offering plea bargains to
the “factually guilty.” Although Packer did not
Penal Paradigms: Past and Present
perceive these models as mutually exclusive, his
dichotomization has penetrated deeply into
(at least) the academic criminal justice culture.
Critiques of Packer’s Models
Packer’s Models have been the brunt of a number
of critiques, some trenchant, with Griffiths (1970)
being a notable early example, and most recently
and comprehensively by Macdonald (2008).
These critiques have focused on three main
aspects of Packer’s work: (a) his conceptualization of the model, (b) the validity of the due
process and the crime control models and the
relationship between them, and (c) the need for
additional models having (in the view of their
proponents) greater validity or irrelevance.
(a) Conceptualization
The lack of precision in the concept of the
model in the contemporary literature of the
social sciences – on one view its “indiscriminate use” (Macdonald 2008) – has already
been referred to. Zedner (2004: 116) has
observed that “It is often difficult to determine whether models are intended as analytical devices or as normative statements of
ideals to which the process should aspire” –
an ambivalence conveyed by Packer too. In
the wake of Damaska’s discussion as to
whether Packer’s models are consistent with
Weber’s ideal types, Macdonald notes that
Weber distinguished between ideal types
and ideals, the former constituting an analytical construct based upon accentuated tendencies which was in his view how Packer
conceptualized his models, without fully
developing them. (Whether in fact the models
were quite unconnected with value
judgments, as Weber described his ideal
types, is another question).
Macdonald sought, as the title of his article
indicates, to “learn from Packer’s mistakes”
and to “succeed. . . where Packer failed”
(Macdonald 2008: 269), for which purpose
he found necessary to differentiate between
the investigative and the trial stages of the
process, while arguing the need for
a multidimensional model. On similar grounds
Ashworth and Redmayne (2005: 38ff.)
Penal Paradigms: Past and Present
reject the “models” approach as an unsatisfactory basis for the evaluation of the criminal
process. These writers find models based on
simple dichotomies to be too crude – tending
“to reduce the world to black and white”
(Zedner 2004: 119–120).
Additional conceptual issues arise in the
context of the evaluation of the two models as
outlined by Packer, and the relationship
between them.
(b) The Two Models and the Relationship
Between Them
A number of criticisms have been directed
at the specific models described by Packer, as
well as suggesting that the dichotomy may be
a false one (see, e.g., Macdonald 2008:
264ff.). Both Damaska (1973) and Smith
(1997) find fault with due process as an
objective (in Damaska’s words “obstacles
without a course”), rather than a principle
imposing restraints on the pursuit of other
objectives. In this context Macdonald
(2008: 289) refers to the work of the Swedish
jurist Jareborg who differentiated between
a “defensive model” and an “offensive
approach” to criminal policy. Zedner (2004:
118) has pointed out that speed and
efficiency – characteristics attributed by
Packer to the crime control model – are also
important for due process and protection of
the defendant’s liberty. McBarnet (1981),
echoing the literature of radical criminology,
has argued that rather than constituting an
alternative approach, due process serves to
legitimize the goals of (repressive) crime
control, thus giving rise to “due process. . .
for crime control.”
(c) Additional Models
With Packer’s dichotomy as a “given,”
other writers have proffered additional
(often designated “third”) models. Griffiths’
“family model” (which clearly falls into
Zedner’s
normative
category)
was
a proposal for a more therapeutic procedure.
This, too, has been the direction of at least
some of the recent writings seeking to accommodate a role in the system for the victim.
Thus, while Beloof (1999) has proposed
3487
P
a Victim Participation Model within a due
process framework, Roach (1999) has
portrayed the alternatives of a punitive and
a non-punitive model of victims’ rights, and it
is the latter – designated a “Circle Model”
and based upon restorative justice – which he
seeks to promote.
A unique attempt to incorporate a broader
model approach into an empirical analysis of
the criminal justice system was undertaken
some years ago by King (1981). King identified six alternative models of the system for
which accounts could be found – whether by
name or implicit – in the criminological or
criminal justice literature. To Packer’s duo
(due process and crime control), King
added: the medical (or treatment) model,
and three further models deriving from criminological research or sociological critiques –
the bureaucratic model, the status passage
model, and the power model. He then identified the “features of the court” which one
would expect to find under each model. The
validity of these models could then be tested
(in the instant case, in his study of English
magistrates’ courts) by observing to what
extent these features were actually found in
the courts. “As in the physical and social
sciences, the term model is used here to
denote a hypothetical but coherent scheme
for testing the evidence” (p. 12).
There is no doubt that the integration of
a multi-model approach into a field study
served to enrich this discourse. However,
there is a measure of disjuncture here in the
construction of the models in that the first
three, which King calls “participant” models,
reflect the idealized expectations of a certain
category of participant (thus, again, falling
into the category of normative models),
while the other three which he refers to as
“social” models are based on academic critiques of the existing system. On the other
hand, the attribution to each model of
a “social function” bestows upon the models
a unifying ideology or rationale such as generally characterized the paradigms discussed
in the first part of this entry.
P
P
3488
It should be noted that the “penal process” to
which the literature reviewed in this section relates
is that of common law jurisdictions (primarily the
USA). Models relating to other systems will be
very different, but some “Packer scholars” have
dwelt on the relationship between the common law
and developments in continental Europe’s “inquisitorial system”: see, e.g., Damaska (1973).
Special Categories of Offender
Variations in the manner in which criminal justice systems operate tend to be augmented when
dealing with offenders with distinctive traits or
pathologies – or whose offences are perceived as
particularly threatening or distinctive in some
other way. Examples would be sex offenders,
offenders against state security laws (“terrorists”), offences related to drug use, minors, or
the mentally unstable. In such cases different
paradigms or models may be identified. Thus,
the modification of human rights protections
when dealing with security offences (including
punishment without trials) may be such that the
“war” model is arguably the most appropriate
metaphor to attach in this case.
By contrast, purportedly softer alternatives to
the prevailing model(s) may be adopted when
dealing with minors. Identifying the optimal policy for his purpose, however, has been the subject
of intense debate over the past century. It was in
relation to juvenile offenders that the treatment
model of the positivist school assumed its most
comprehensive application, in particular via the
vehicle of the juvenile court, first established
(in Chicago) in 1899. The treatment of juvenile
offenders became virtually indistinguishable from
“welfare” cases such as children in need of care
and supervision. They were thus deprived of the
guarantees provided under the US Constitution’s
Bill of Rights. Juvenile courts in other common
law jurisdictions, however, retained somewhat
more of a rights framework – while the US
Supreme Court case Re Gault in 1967 (387 U.S.
1) reversed the welfarist trend in the American
system. If for some decades the juvenile justice
system has seen these two approaches – the “welfare” model and the “justice” or “legal” model – in
combat (or fusion), recent decades have seen
Penal Paradigms: Past and Present
strong claims, particularly in the antipodes, for
their replacement by a system of restorative justice. Three distinct approaches to juvenile justice
have thus emerged. As observed by Borowski and
O’Connor (1997: 229): “Unlike other aspects of
the criminal justice system, it is possible to identify coherent theoretical positions reflected in legislation, policy and practice.”
O’Connor’s review of these models considers
how each model responds to such underlying
questions as follows: What is the cause of
crime? How should it be viewed? How should it
be defined in law? How should the state respond?
What is the relationship between the individual
and the state? What is the purpose of the sanction? O’Connor’s models arguably present both
an analytical and a normative framework for the
comparison of these three ideologies, while his
approach bears a resemblance to Young’s analysis of criminological paradigms referred to in the
Introduction.
Conclusions
This entry has analyzed the use made of two
concepts – the paradigm and the model – in the
course of the history of the criminal justice system. It seems to the present writer that the paradigm has been used primarily to describe changes
in the main orientation of the system at different
periods, while the model has been more extensively used in contemporary settings. Both terms
have had a usage somewhat different from their
original meaning (in the case of the model,
a multiplicity of usages), and partly for this reason, the discourses in which they have been
incorporated have been challenged.
The departure point for this entry was the
concept of the paradigm – perhaps the more
interesting concept because of the connotations
conveyed by this term in a penological context of
(a) being unequivocally linked to an ideology,
(b) indicating an interpretation uniquely accepted
by the relevant disciplinary group, and (c) having
played a key role in the history of penology and
the penal system. Yet while the brief historical
overview presented at the beginning of this entry
Penal Paradigms: Past and Present
may have been sufficient to indicate the dramatic
contrast between the main historic paradigms (in
particular classicism and positivism), Young is
almost certainly correct in denying the occurrence of a “paradigm shift” as a momentary
historical event, the reality having been considerably more complex. Further, the idea that the
“community of [criminology] scholars” will
identify with a common vision of penal policy
has become increasingly unlikely from generation to generation. Even less likely is the prospect
of shared perceptions between the “criminology
community” and policy-makers, such as Garland
suggests may have occurred in the era of “penal
welfarism” and “government criminology.” As
observed by Garland and Sparks (2000), in an
era of ideological fragmentation and politicization, paradigms – like cultures – are likely to be
multiple.
The relatively fluid concept of the model is free
from such connotations, but its conceptual ambiguities have given rise to much confusion. This
confusion came to the fore in the aftermath of the
publication of Herbert Packer’s famous two
models of the criminal process – the crime control
model and the due process model, and the numerous critiques thereof. Nevertheless, Packer’s
models have become an inseparable part of the
discourse of criminal justice. Indeed, the two concepts considered in this entry are likely to survive
and enhance both critical analyses and normative
debates on the many facets of penality – while
methodological inquiries as to their validity and
applicability may serve to clarify underlying epistemological issues.
Acknowledgement The author would like to thank Prof.
Lucia Zedner for her helpful comments on a previous
draft.
Related Entries
▶ Historical Criminology
▶ History of Corrections
▶ History of Juvenile Justice
▶ Legal Control of the Police
▶ Problem-Solving Courts
▶ Prosecution and Wrongful Convictions
3489
P
Recommended Reading and References
American Friends Service Committee (1971) The struggle
for justice. Hill & Wang, New York
Ancel M (1965) Social defence. Routledge & Kegan Paul,
London
Ashworth A, Redmayne M (2005) The criminal process,
3rd edn. OUP, Oxford
Barnes HE, Teeters NK (1959) New horizons in criminology, 3rd edn. Prentice-Hall, Englewood Cliffs
Beccaria C (1764/1995) On crimes and punishments and
other writings (trans ed: Bellamy R). Cambridge University Press, Cambridge
Beloof DE (1999) The third model of criminal process: the
victim participation model. Utah Law Rev 2:289
Black D (1976) The behavior of law. Academic, New York
Blomberg T, Lucken K (2000) American penology:
a history of control. Aldine de Gruyter, New York
Borowski A, O’Connor I (1997) Models of juvenile justice. In: O’Connor I, Borowski A (eds) Juvenile crime,
justice and corrections. Addison Wesley Longman,
Melbourne (Chap 12)
Chaiken J et al (1975) Criminal justice models: an overview. Rand, Santa Monica
Damaska M (1973) Evidentiary barriers to conviction and
two models of criminal procedure: a comparative
study. Univ Pa Law Rev 121:506
Diamond AS (1971) Primitive law: past & present.
Methuen, London
Feeley M (1973) Two models of the criminal justice
system – an organizational perspective. Law Soc Rev
7:403
Frankel ME (1973) Criminal sentences: law without order.
Hill & Wang, New York
Garland D (1985) Punishment and welfare. Gower,
Aldershot
Garland D (1990) Punishment and modern society.
Clarendon, Oxford
Garland D (2001) The culture of control. OUP, Oxford
Garland D, Sparks R (2000) Criminology, social theory,
and the challenge of our times. In: Garland D, Sparks
R (eds) Criminology and social theory 1. Clarendon,
Oxford
Griffiths J (1970) Ideology in criminal procedure or
a ‘third model’ of the criminal process. Yale Law
J 79:359
http://www.newworldencyclopedia.org/entry/Paradigm.
Accessed 22 Jan 2013
King M (1981) The framework of criminal justice. Croom
Helm, London
Kuhn T (2012) The structure of scientific revolutions.
4th edn, University of Chicago Press, Chicago
MacDonald S (2008) Constructing a framework for criminal justice research: learning from Packer’s mistakes.
New Crim Law Rev 11:257
Martinson R (1974) What works? Questions and answers
about prison reform. Pub Interes 35:22
McBarnet DJ (1981) Conviction: law, the state and the
construction of justice. Macmillan, London
P
P
3490
Packer H (1964) Two models of the criminal process.
Univ Pa Law Rev 113:1
Packer H (1968) The limits of the criminal sanction.
Stanford University Press, Stanford
Roach K (1999) Four models of the criminal process.
J Crim Law Criminol 89:671
Saleilles R (1911) The individualisation of punishment.
Little, Brown, Boston
Semple J (1993) Bentham’s prison: a study of the
panopticon penitentiary. Clarendon, Oxford
Smith DJ (1997) Case construction and the goals of criminal justice. Br J Criminol 37:319
Tonry M (ed) (2007) Crime, punishment and politics in
comparative perspective. U of Chicago Press, Chicago
von Hirsch A (1976) Doing justice: the choice of punishments, Hill & Wang, New York
Young J (1981) Thinking seriously about crime: some
models of criminology. In: Fitzgerald M (ed) Crime
and society: readings in history and theory. Routledge,
London
Young J (1994) Incessant chatter: recent paradigms in
criminology. In: Maguire M et al (eds) The Oxford
handbook of criminology. Clarendon, Oxford, p 69
Zedner L (2004) Criminal justice. OUP, Oxford
Penal Philosophy and Sentencing
Theory
Matt Matravers
Morrell Centre for Toleration
University of York, York, UK
Overview
Many penal philosophers think that offering
a justification of the state’s right to punish is
particularly urgent (and difficult) precisely
because punishment involves doing things to
people that would, outside the practice of punishment, be gross rights violations. That is, punishment typically involves imposing some
deprivation – for example, the removal of property or freedoms – on a person for committing an
offense and this stands in need of justification.
Yet, despite this motivation, many theories of the
justification of punishment say surprisingly little
about the detail of sentencing and penalties. This
entry considers the various rationales for sentencing with a view to showing both how different
justifications of punishment have different
Penal Philosophy and Sentencing Theory
implications for sentencing theory and practice
and how reflecting on sentencing matters can
shed light on the justification of punishment.
The entry considers first consequentialist theories – that is, theories that justify punishment by
reference to the net benefits that accrue through
punishment in comparison with any other form of
social control (or doing nothing) – and then
desert-based, retributive, theories. Finally, it considers some questions of the relationship between
theory and practice.
Punishment and Sentencing
Punishment
Punishment – or at least state punishment –
involves the intentional infliction on an offender
by a proper authority of “pain or other consequences normally considered unpleasant” for an
offense against legal rules (Hart 1968, pp. 4–5).
In addition, it involves the expression of moral
condemnation or censure (Duff 2007; Feinberg
1970). When a court passes sentence, it gives
shape to the particular consequence to be visited
on the offender. In practice, the options available
to the court may be limited by guidelines or
statute and the court may not be explicit in justifying the use of one penalty rather than another.
However, the use of punishment in general, and
of particular penalties, requires justification and
those justifications tend to fall into two broad
categories: consequentialist and retributive.
For consequentialists, punishment is justified
because its consequences – specifically, having
the system of punishment with its associated
penalties – can be expected to yield higher net
benefits than having an alternative system of
social control or having no such system. This
may be because punishment reduces future
instances of criminal behavior through deterrence, rehabilitation, or incapacitation.
For retributivists, punishment is justified
because it is deserved by the offender for his
offense. It may be that good consequences also
follow, but the justification of punishment does
not appeal to these consequences. What, then,
do these broad theories of punishment imply
Penal Philosophy and Sentencing Theory
for sentencing theory (insofar as we can distinguish sentencing theory and penal philosophy)
and practice?
Deterrence
Deterrence is typically split into “special” and
“general.” Special deterrence is aimed at the particular offender and is sometimes discussed in
relation to the giving of more substantial punishments to repeat offenders. However, it is not
advanced as a general theory of sentencing. This
is perhaps a reflection of the lack of empirical
evidence when it comes to special deterrence and
of the fact that a sentencing scheme that was
genuinely tied to special deterrence would have
to allow specific, individualized sentences for
each offender in relation to his particular likelihood of reoffending. Such a sentencing scheme
would lack all consistency and would, in any
case, be unworkable.
General deterrence addresses not the particular offender in court, but potential offenders
throughout the population. If the general justifying aim of punishment is to secure good consequences by reducing criminal behavior, then it is
argued that the mechanism by which it does that
is general deterrence. The origins of the modern
approach are most often associated with the great
utilitarian philosopher Jeremy Bentham. Bentham thought the infliction of pain was, other
things equal, an evil, but the threat – and where
that threat failed, the infliction – of pain could be
justified if it secured better future consequences
than would be achieved by any alternative action
or policy. Assuming that the addressees of the
system (that is, ordinary citizens) were motivated
by the pursuit of pleasure and the avoidance of
pain, Bentham argued that criminal sanctions
should be set at just that level required to change
the outcome of the “hedonic calculus” for each
individual in favor of what would be conducive to
the maximization of the general good. Moreover,
Bentham went on, penalties should be such as to
“induce a man to choose always the least mischievous of two offenses” by ensuring that
“where two offenses come into competition, the
3491
P
punishment for the greater offense must be sufficient to induce a man to prefer the less” (Bentham
1970, pp. 168). That is, he developed an account
of penalty scales (Bentham 1843).
In thinking of people as rational calculators set
on maximizing their own interests and susceptible to changes in the payoffs of different actions,
Bentham’s theory has much in common with
contemporary proponents of the economic theory
of law such as Richard Posner (Posner 1985).
However, critics of the position have pointed to
flaws both in the empirical claims made on behalf
of general deterrence and in its normative
foundations.
Empirical Questions Around General
Deterrence
Empirical questions around general deterrence
revolve around the overall claim that the threat
of legal punishment reduces offending and the
claim that marginal changes in the severity of
sanctions has an effect on rates of offending.
The difficulty with the first of these claims is in
establishing that it is the threat of punishment
rather than anything else that stops people from
breaking the law. Moreover, insofar as the theory
posits an effect that happens because people reason about what to do as rational, self-interested
calculators, the account may explain some people’s behavior and not that of others. It may
simply not be true that certain kinds of offenders
do accurately balance the factors for and against
their offending in advance (even when they
know, for example, what penalty they are likely
to be given and the likelihood of being caught). In
short, general deterrence may be at best
a selective explanation of law-abiding behavior
because many offenders do not think as Bentham
and modern economists believe they think (Doob
and Webster 2003).
A similar problem arises with the empirical
evidence in relation to the effects of marginal
changes in sentencing severity or probability of
conviction. In a review of research, Andrew von
Hirsch and his coauthors identified the factors
that would need to be true for such marginal
changes to have an effect on a potential
offender’s behavior. That offender, they argued,
P
P
3492
would need to know about the change; he would
need to include the increased or decreased risks
of suffering punishment in his decision as to
whether to offend; he would need to believe in
a realistic possibility that he might be caught and
that the penalty would be applied to him; and he
would need to be willing to adapt his behavior in
the light of these new considerations (von Hirsch
1999). The conclusion of the study is that while
there is some evidence of a link between
offending and the certainty of punishment, there
is far less evidence of a link between offending
and the severity of punishment (see also Bottoms
and von Hirsch 2011).
Normative Questions Around General
Deterrence
Insofar as deterrence theories are embedded in,
and result from, consequentialist accounts of
punishment, they are subject to the moral criticisms that are often leveled at those accounts. The
problem is a general one: Consequentialist
accounts value the aggregate (sometimes the
average) good achieved in the future. We are to
do what will, in Bentham’s famous phrase,
achieve “the greatest happiness of the greatest
number.” But, in being concerned for the greatest
happiness (or utility), the theory leaves out questions of the distribution of utility; it fails to pay
attention to what John Rawls (1971, p. 27) calls,
“the distinction between persons.” For this reason, it is alleged, if the calculation dictates it, it
may be not merely morally permissible, but morally obligatory, to frame and punish an innocent
person (McCloskey 1968) or to punish an
offender far in excess of what is usual to make
an example of him.
Consequentialist writers have responded to
this kind of criticism by insisting that the theory
is best thought of as addressing rules and institutions and not individual actions (Goodin 1995;
Rawls 1955/1999). In that case, they argue the
system of rules in relation to criminal justice and
the range and type of penalties that best promote
long-term good will be much the same as is found
in current liberal systems. That is, the rule “only
punish those (believed to be) guilty and not the
innocent” will better promote the good than a rule
Penal Philosophy and Sentencing Theory
that allows the innocent to be sacrificed each and
every time that particular act will have better
long-term consequences than any alternative.
However, few have been convinced by this argument in part because of doubts about the claim
that the usual procedural rights and protections of
(suspected) offenders will ensure long-term utility, and in part because even if it delivers the right
answer (not to hang the innocent man), it does so
for the wrong reasons. The innocent man
deserves to be spared because it would be
wrong to kill him, not because a rule against
hanging innocents secures better consequences
than its opposite (Matravers 2000, pp. 17–23).
The idea that the justification of punishment
and of the penalties it imposed must be grounded
in the requirement “to preserve and increase the
welfare of the state” (Michael and Adler 1933,
p. 340) was prevalent for much of the century
prior to the 1960s (Tonry 2011b, pp. 14–18). It
seemed to some writers to be the only “rational”
basis for sentencing (Walker 1969). However,
from the 1970s onward, the theory came under
sustained criticism. In moral theory from neoKantian writers including, and inspired by, John
Rawls and in penal theory by the “Just Deserts”
movement. Yet, we continue to believe – and to
find evidence for – the claim that the whole criminal justice system must surely have something to
do with reducing future crime rates and that taken
as a whole, it does deter crime (Robinson and
Darley 2003, 2004). Moreover, statutes and
judges continue to refer to deterrence at the
point of sentencing. So, whatever its philosophical difficulties, it is clear that deterrence is not yet
moribund in practice.
Incapacitation and Rehabilitation
At the high point of utilitarian welfarist thinking
about punishment (in roughly the middle of the
Twentieth Century), it seemed obvious to writers
on the criminal law that a rational approach
to punishment would eschew retributivism as
“fallacious” (Michael and Adler 1933, p. 341)
and barbaric. Instead, it seemed clear that if the
welfare of the state was to be increased, then
Penal Philosophy and Sentencing Theory
antisocial behavior, and those who committed it,
needed to be dealt with “scientifically.”
Offenders would be “treated” under the careful
eye of experts and until “cured” would generally
be held away from the general population. For the
influential social and legal thinker, Barbara
Wootton, the criminal law was irrational in its
concern with intentions and mental states prior to
conviction. Rather, mens rea considerations
mattered only “after a breach of law has been
proved” because of “the light which they throw
on the likelihood of [the offender] offending
again, and upon the most hopeful way of dealing
with him” (Wootton 1978, p. 224). Sentences
should be indeterminate and offenders released
when, or if, their individualized treatment plans
proved successful.
There were individual voices of dissent. For
example, CS Lewis’s (1949) “The Humanitarian
Theory of Punishment” offered the argument that
the treatment model reduced persons from choosing agents to manipulable things. By the late
1960s, those voices had increased and by the
1970s become a clamor (Frankel 1972; Kleinig
1973; Morris 1974; von Hirsch and Committee
for the Study of Incarceration 1976). The charge
was normative in that utilitarianism treated people as mere objects to be manipulated; political in
that indeterminate sentences were racially biased
and unfair; and empirical in that increasing
doubts were expressed about the efficacy of
treatment.
Remnants of both incapacitation and rehabilitation nevertheless remain in contemporary penal
philosophy and sentencing theory. With respect
to incapacitation, governments (of all
persuasions) have responded to the risks created
by so-called dangerous persons – often sex
offenders – by allowing incapacitative sentences
or post-sentence restrictions such as sex-offender
registers. However, the policy has largely drifted
free of any theoretical foundation in a theory of
punishment. Rather, it exists as a bolt-on public
policy that philosophers have approached as
a discrete problem. The question is, can – and if
so, how should – a liberal society restrict the
freedoms of some for purely or primarily preventive reasons? The best answer to that is that any
3493
P
such policy is justifiable only if: (1) the potential
harm was sufficiently grave; (2) the prediction
technology was sufficiently accurate; (3) the preventive response was maximally humane and
minimally intrusive under the circumstances;
and (4) if the preventive action was preceded by
adequate due process (Morse 1999, p. 297). For
many critics, condition (2) is seldom met given
the unreliability of current predictive judgments
(Monahan 2004).
Similarly, rehabilitation is not regarded by
most contemporary penal philosophers as providing an overall theory of punishment or criminal
justice, but that is not to say that it is redundant in
sentencing decisions. Two different motivations
have sustained rehabilitative programs: first,
a humanitarian desire to help those in incarceration many of whom have mental health issues,
drug issues, and basic training needs; second,
increased frustration with rates of recidivism
that spawned the “What Works?” movement –
a political movement keen to subject policy to
evidence-based evaluation – which revitalized
interest in techniques for anger control, the use
of cognitive behavioral therapy, and other techniques of behavioral modification. This focus on
the “good done” by legal processes (including
arrest, trial, and punishment) has developed into
a small but significant movement under the title
of “therapeutic jurisprudence”, which (as the
name suggest) aims to assess criminal justice
institutions against a standard of therapeutic usefulness (Wexler 2008; Winick 1997).
In short, incapacitation and rehabilitation
remain important parts of public policy and sentencing practice. But, they do so largely as discrete elements and not as the result of an overall
theory of punishment in which the welfare of
society is the only desideratum.
Retributive Theories
Retributive theories of punishment go back (at
least) to the great German philosophers Immanuel Kant and Georg Hegel (see ▶ Traditional
Retributivism). Following the assault on indeterminate sentencing in the 1970s, and the more
P
P
3494
general philosophical critique of consequentialist
thinking at around the same time, retributivism
came back into fashion for the rest of the century.
Retributivists hold that punishment is justified
when, and because, it is deserved (although the
exact sense of desert is contested (Matravers
2011)). More precisely, the argument is that
those who culpably commit criminal offenses
deserve censure and that censure should take the
form of a penal sanction – so-called hard treatment. Most importantly for the development of
a retributive sentencing scheme, that hard treatment must be proportionate both in the sense of
cardinal proportionality (the penalty should be
fitting or commensurate with the wrong) and of
ordinal proportionality (crimes of comparable
seriousness ought to receive penalties of comparable severity).
The precise way in which retributivists defend
their theories varies. For Michael Moore (Moore
1987, 1997), punishment is good when and
because it is deserved and this judgment and the
associated penalties are a matter of what is dictated by a correct account of morality. For
Antony Duff, it is censure that is deserved and
hard treatment is merely the vehicle through
which that censure is communicated and through
which the offender can show penance (Duff 1986,
2007). For Andrew von Hirsch and Andrew
Ashworth, retributivism is a mixed theory in
which censure addresses us as moral beings and
hard treatment offers a “prudential supplement”
encouraging us to keep to the law (von Hirsch
1993; von Hirsch and Ashworth 2005).
Proportionality
The principle of proportionality is very attractive.
If punishment is an exercise in blaming, then the
penalties must reflect this in the sense both that
the penalty must be at an appropriate level and
that more serious wrongs deserve more blame
and so more severe punishment. This also captures a very important commitment to equality
and justice often glossed as “treat like cases alike
and different cases differently.” That is, if two
offenders commit crimes of equal seriousness
with equal culpability, then they should be
punished equally. Moreover, once an account
Penal Philosophy and Sentencing Theory
can be given of the anchoring points of the scales
(the least and most serious offenses and the least
and most severe penalties), it appears to offer
a workable translation of a theory of punishment
into a sentencing scheme.
However, the proportionality account of punishment is not without its critics. For “pure”
censure-based theorists, the addition of hard
treatment as a prudential supplement treats citizens as things to be coerced rather than agents to
be reasoned with, and threatens to “drown out”
the moral message of the law (see the exchange
between Duff and von Hirsch in Matravers 1999).
However, even those who accept the necessity of
hard treatment as a legitimate part of punishment
have queried the usefulness of proportionality as
a guiding principle. The criticisms in part focus
on the theory’s workability and in part on its
implications for sentencing practice.
A punishment scheme based on proportionality presumes that we can accurately measure
offense seriousness and culpability, and that we
can then match that to a point on a penalty severity scale. These things pose a significant challenge to proportionality theorists. In relation to
offense seriousness, what is needed is a way to
put together on one scale crimes such as assault
by an individual and commercial pollution. In
relation to culpability, in principle, the theory
needs to assess the subjective state of the
offender, but in practice may have to rely on
objective measures such as the absence or presence of previous convictions. Proportionality theorists are aware of these (and other challenges)
and have tried to meet them (Ashworth 2010; von
Hirsch and Ashworth 2005), but not everyone is
convinced that they have been successful (for
a helpful summary of the criticisms, see Tonry
2011b, pp. 224–34).
Perhaps even more significant is the need to
anchor the scale of penalty severity. The bottom
anchoring point – the penalty to be associated
with the least serious offense – may be reasonably
easy to fix in that many jurisdictions may have
something like a police caution. However, what of
the upper anchoring point? Should this be capital
punishment (or worse), life imprisonment without
the possibility of parole, 21 years, or less?
Penal Philosophy and Sentencing Theory
Without an answer to this question, the scheme
is inadequate.
Proportionality theorists have reached for
a variety of principled, and socially contingent,
ways of fixing the upper anchoring point. As
noted above, one might worry, for example, that
too severe punishments will “drown out” the
moral message conveyed by the law. The problem the account faces is that insofar as the penalty
conveys proportionate blame, different societies,
with different histories and practices, will think
different overall levels of suffering needed to do
this accurately.
Parsimony
In contrast to the equality and proportionality
required by the proportionality account, Norval
Morris (1974) and others have argued for
a principle of parsimony. By “parsimony” it is
meant that punishment levels ought in general to
be set such as that offenders receive “the least
severe sentence consistent with the governing
purposes of sentencing” (Tonry 1992, p. 81). In
developing this account, Morris and others have
come to think of desert as a “limiting” principle
(Frase 2004) such that offenders should receive
the least severe “not undeserved” sanction from
a range in which the upper point is consistent with
proportionality, but the lower bounds allow for
individualization of sentences and parsimony.
Such an account is still hostage to a theory of
cardinal desert, but its relative vagueness would,
its proponents suggest, at least allow greater justice to be done to each particular offender; justice
that would be sacrificed by the fetishizing of
proportionality. Exactly that discretion, though,
is what bothers proponents of proportionality
(perhaps mindful of the way it was used to disadvantage minority groups in the past).
Theory and Practice
Following the breakdown of the consequentialist
consensus in the final third of the twentieth century, broadly retributive theories of punishment
took center stage and, by-and-large, remain
there (although see the essays in Tonry 2011a).
3495
P
The associated “just deserts” movement, and its
various political analogues like the campaign for
“truth-in-sentencing,” was dominant, but for
a shorter period and not without challenge.
Throughout the same period, and in particular
after the mid-1980s, many sentencing policies
that explicitly violated proportionality were
passed. Mandatory minimum sentences for
drugs offenses, sexual offenses (particularly
against children), and possession offenses
(particularly of child pornography) have meant
that these crimes attract penalties far above those
given for what might be thought of as roughly
equally serious violent crimes. Another example
is California’s notorious three-strikes law, which
means that relatively trivial third offenses could
attract very severe penalties, and there are many
others.
As noted above, other less retributive policies – grounded in therapeutic jurisprudence or
restorative justice – also flourished in corners of
the criminal justice system often because they
were perceived to “work”; in particular to reduce
recidivism.
Very few of these policies can be said to flow
directly from a fully worked-out philosophical
position on the justification of punishment. Yet,
it would be absurd to say that “theory does not
matter.” Policies do change as a result of theoretical reflections and will no doubt continue to do
so. However, sentencing – perhaps more than any
other part of the criminal justice system – is
political and pluralistic. It is political in that it is
a highly visible and politically controversial. To
be “soft on crime” is, in many political systems,
to court political disaster (Lacey 2008). It is pluralistic in that, at the point of sentencing, judges
in systems where there is discretion will often call
upon many considerations (an intuitive notion of
desert, the need to express the community’s
blame, deterrence, the risk posed by the offender,
the offender’s potential for reform, and so on).
Such pluralism can sit ill with philosophers of
punishment who seek solutions which are logically consistent and morally coherent and
defensible.
That said, precisely because crime is
a politically sensitive issue, and because in
P
P
3496
many jurisdictions, judges are limited in their
discretion by mandatory sentences or restrictive
sentencing guidelines, it is vital that moral and
penal philosophers continue to subject sentencing
rationales to critical analysis. And, in doing so,
they will inevitably call on overall accounts of the
general purpose, and justification, of punishment
even if those accounts do not always translate
easily into particular sentencing outcomes.
Recommended Reading and References
Ashworth A (2010) Sentencing and criminal justice,
5th edn. Cambridge University Press, Cambridge
Bentham J (1843) Principles of penal law. In: Bowring
J (ed) The works of Jeremy Bentham, 1st edn. William
Tait, Edinburgh
Bentham J (1970) In: Fred R, Philip S (eds) An introduction to the principles of morals and legislation. Oxford
University Press, Oxford
Bottoms A, von Hirsch A (2011) The crime preventive
impact of penal sanctions. In: Cane P, Kritzer H (eds)
The Oxford handbook of empirical legal research.
Oxford University Press, Oxford
Doob AN, Webster CM (2003) Sentence severity and
crime: accepting the null hypothesis. Crime Justice
30:143–195 (ArticleType: research-article/Full publication date: 2003/Copyright # 2003 The University of
Chicago Press)
Duff RA (1986) Trials and punishments. Cambridge University Press, Cambridge
Duff RF (2007) Answering for crime: responsibility
and liability in the criminal law. Hart Publishing,
Oxford
Feinberg J (1970) The expressive function of punishment.
In: Doing and deserving: essays in the theory of
responsibility. Princeton University Press, Princeton,
pp 95–118
Frankel ME (1972) Criminal sentences: law without order.
Hill and Wang, New York
Frase R (2004) Limiting retributivism. In: Tonry M (ed)
The future of imprisonment. Oxford University Press,
New York, pp 83–120
Goodin R (1995) Utilitarianism as a public philosophy.
Cambridge University Press, Cambridge
Hart HLA (1968) Punishment and responsibility: essays in
the philosophy of law. Oxford University Press,
Oxford
Kleinig J (1973) Punishment and desert. Springer, New
York
Lacey N (2008) The prisoners’ dilemma: political economy and punishment in contemporary democracies.
Cambridge University Press, Cambridge
Lewis CS (1949) The humanitarian theory of punishment.
20 Century Aust Q Rev 3(3):5–12
Penal Philosophy and Sentencing Theory
Matravers M (1999) Punishment and political theory. Hart
Publishing, Oxford
Matravers M (2000) Justice and punishment: the rationale
of coercion. Oxford University Press, Oxford
Matravers M (2011) Is twenty-first century punishment
post-desert? In: Tonry M (ed) Retributivism has
a past: has it a future? Oxford University Press,
New York
McCloskey H (1968) A non-utilitarian apporach to punishment. In: Bayles M (ed) Contemporary utilitarianism. Doubleday, New York, pp 239–259
Michael J, Adler M (1933) Crime, law and social science.
Harcourt, Brace, New York
Monahan J (2004) The future of violence risk management. In: Tonry M (ed) The future of imprisonment.
Oxford University Press, New York
Moore M (1987) The moral worth of retribution. In:
Schoemann F (ed) Responsibility, character and the
emotions. Cambridge University Press, New York,
pp 179–219
Moore M (1997) Placing blame: a general theory of the
criminal law. Oxford University Press, Oxford
Morris N (1974) The future of imprisonment. University
of Chicago Press, Chicago
Morse SJ (1999) Neither desert nor disease. Legal Theory
5:265–309
Posner RA (1985) An economic theory of criminal law.
Columbia Law Rev 85:1193–1195
Rawls J (1955/1999) Two concepts of rules. In: Rawls J,
Freeman SR (eds) Collected papers. Harvard University Press, Cambridge, pp 20–46
Rawls J (1971) A theory of justice. Harvard University
Press, Cambridge
Robinson P, Darley J (2003) The role of deterrence in the
formulation of criminal law rules: at its worst when
doing its best. Georgetown Law J 91:949
Robinson P, Darley J (2004) Does criminal law deter?
A behavioural science investigation. Oxf J Legal
Stud 24:173–205
Tonry M (1992) Proportionality, parsimony, and interchangeability of punishments. In: Duff A et al (eds)
Penal theory and penal practice. Manchester University Press, Manchester
Tonry M (2011a) Retributivism has a past: has it a future?
Oxford University Press, New York
Tonry M (ed) (2011b) Why punish? How much?: a reader
on punishment. Oxford University Press, New York
von Hirsch A (1993) Censure and sanctions. Clarendon
Press, Oxford
von Hirsch A (1999) Criminal deterrence and sentence
severity : an analysis of recent research. Hart,
Oxford/Portland
von Hirsch A, Ashworth A (2005) Proportionate sentencing : exploring the principles. Oxford University Press,
Oxford/New York
von Hirsch A, Committee for the Study of Incarceration
(1976) Doing justice: the choice of punishments:
report of the Committee for the Study of Incarceration,
1st edn. Hill and Wang, New York
Philosophical Basis of the Forensic Process
Walker N (1969) Sentencing in a rational society. Allen
Lane/The Penguin Press, London
Wexler DB (2008) Rehabilitating lawyers: principles of
therapeutic jurisprudence for criminal law practice.
Carolina Academic Press, Durham
Winick B (1997) The jurisprudence of therapeutic jurisprudence. Psychol Public Policy Law 3(1):184–206
Wootton B (1978) Crime and penal policy: reflections on
fifty years’ experience. George Allen & Unwin,
London
3497
P
Personal Control: Criminal Behavior,
Social Control, Self-Control, Models,
Constraints
▶ Control Theory
Personalized Policing
▶ Community Policing
Penalties
▶ Sentencing Research
Philosophical Basis of the Forensic
Process
Perceptions of Risk
▶ Rational Choice, Deterrence, and Crime:
Sociological Contributions
Camilla Hald
Danish National Police,
Research & Development, Copenhagen,
Denmark
Overview
Perceptual Updating
▶ Bayesian Updating and Crime
Performance Management
▶ Judicial Leadership and Performance
Perpetrator Characteristics
▶ Estimator
Identification
Variables
and
Eyewitness
Persistent Offenders
▶ Career Criminals and Criminological Theory
In recent years controversies have arisen over
the definition and demarcation of science in
relation to forensic science. These controversies
are among other expressed in debates following
a report from the National Academy of Sciences
in the US, questioning the scientific status of
forensics and its interpretations of evidence
before courts of law. The controversies highlight
an apparent incommensurability of the two
fields of knowledge straddled by the forensic
process: science and law. These controversies
are however as old as science and law, respectively, and can be traced in the major philosophical principles’ underpinning, shaping, and
informing the origins and the characteristics of
the forensic process. The philosophical basis
of the forensic process consists of a meshwork
of ideas. It originates from and is influenced by
a range of different disciplines, models of
inquiry, and forms of validating truth claims,
which have been decisive in bringing the
P
P
3498
discipline about. Some major ones runs from
divination, classical rhetoric, theology, law, philology, philosophy, history and the arts, over the
model of Galilean science, medicine, psychotherapy, art connoisseurship, the infant social
science and statistics, to natural sciences such
as biology and chemistry, and, new philosophies
of science born from these disciplines. Following the trajectories of thought illuminates how
there is a philosophical basis to the controversies. The forensic process is informed by ideas
from two different models of science. One is a
model of science based on identifying and
interpreting “signs”, the other is one based on
observation and verification of general facts.
Each model plays a significant part in the workings of the forensic process. The lack of
acknowledgement of the contributions from
each, however, plays a major part in the controversies. Hence the way forward may prove to be
related to attaining a more nuanced understanding of the philosophical roots of the forensic
process as well as in incorporating each model
equally into the theoretical and operational
framework of the discipline.
Key Issues/Controversies
The forensic process is the process of inquiry
belonging to forensic science. Forensic science
is most often defined as an applied science, whose
raison d’être is to provide scientific analysis of
evidence in the service of law. The forensic process is part of the investigative inquiry carried out
within the legal process which is oriented toward
providing answers to questions essential to the
legal deliberation and arbitration of cases. This
however is not a straightforward business, since
the questions asked and the methods of supplying
answers differ markedly between law and
science.
Put crudely, the ultimate goal of legal inquiries is to establish “what happened” (how, when,
where, and why) and “who did it” in a given case
under scrutiny. The goal of inquiry of the sciences is however rather to establish “what
a given phenomenon is” and “how it works.”
Philosophical Basis of the Forensic Process
Also the methods of inquiry, and for validating
“truth” differ between law and science. Legal
inquiry makes use of qualitative methods, testimonial evidence, and a judgment of and deliberation between, probable arguments derived
here from, whereas the sciences use observation
and experiments, favoring the quantifiable,
reproducible, and empirically verifiable.
Differences between law and science are
grounded in a difference between the object and
goal of inquiry of each discipline. The legal process deals with questions of a social, moral, and
cultural kind and in the deliberation of individual
cases in which truth is of an approximate kind as
versed in the dictum “beyond reasonable doubt.”
Natural science (with which forensic science is
most commonly associated and from which the
most influential forensic techniques and forms of
analysis originates: physics, molecular biology,
chemistry, etc.) deals with questions about
nature, its laws, and general characteristics to
the point of establishing the truth about it in
a more absolute sense. The knowledge sought
by science is one which may be reproduced independent of the particular researcher or situation,
and from which it is possible to predict results
based on true premises. Thus for law the goal of
inquiry is the most reasonable explanation for
particular events, while the goal of scientific
inquiry is the universal truth about “what is.”
These two fields of knowledge is straddled by
the forensic process, in which legal questions
must be translated into scientific questions in
order to provide scientific answers which may
inform legal argumentation and deliberation
(see Fig. 1). Forensic science and its role in
legal deliberation is in other words a matter of
using science in the service of law and legal
deliberation, to help establish “matters of fact”
which may aid the process of deliberating
between “probable arguments”.
Evidence produced through the forensic process is thus also submitted to different forms
of validation in different contexts and phases of
the process. Within the laboratory, evidence is
evaluated according to the soundness of analysis
and integrity of the process of analysis (an
uncompromised chain of custody and accuracy
Philosophical Basis of the Forensic Process
3499
P
LAW
RECONSTRUCTIVE (his)’STORY’ of the CASE presented before competent tribunal
According to the 6Ws: “WHO”, “WHAT”, “WHERE”, “HOW”, “WHEN”, “WHY”
Investigative inquiry into ‘matters of fact’
FORENSIC SCIENCE
TESTIMONIAL EVIDENCE REAL EVIDENCE to ‘facts’ of case
‘Story’/testimony of the ‘thing’
Refinement phase
Examination of exhibit
EXHIBIT recovered from scene
SCENE INVESTIGATION
Testing phase
Investigative phase
‘Thing’ at scene
Philosophical Basis of the Forensic Process, Fig. 1 The forensic process – science in service of law (From Tilstone
et al. 2012)
of instruments and methods applied) in order to
assure the scientific “rigor” of the results through
recognized methods of analysis and scientific
procedures for testing. Within the legal process,
the evidence is in addition to the assurance of the
integrity of analysis and evidence, evaluated for
their meaning in relation to the legal argumentation presented before a court of law. The former
demands that what is infered from the evidence is
verifiable and stand up to tests in accordance with
scientifically acknowledged theories and
methods. The latter demands that evidence is
seen in the context of the case and the circumstances surrounding the actions or event in
question and is based on evaluation of possible
or probable causes or sources of particular forms
of evidence.
The evidential value of the information
provided through scientific analysis and the
professional judgment of the analyst is, in other
words, ultimately decided in the context of
the case and the legal process of deliberating
between the arguments presented by the advocates of prosecution and defense.
This fundamental condition of the forensic
process has resulted in intense debates about the
role of the forensic scientist in providing service
to law by analyzing and interpreting physical
evidence. Questions such as: How this is done;
where the boundaries of professional judgments
in interpreting the evidence, are set and how far
the forensic analyst may stretch in order to help
answer the questions of the legal process – and
how the authority and integrity of the
scientist should be assessed – are however subject
to different interpretations. Within forensic science itself, tensions are evident in the way that
different schools have arisen, especially within
criminalistics, each school emphasizing the role
of the forensic analyst differently, as well as the
degree to which criminalistics should be defined
by the needs and methods of law versus those of
the pure sciences (Inman and Rudin 2001).
These inherent tensions have become more
obvious in the United States, in the wake of
a report from the National Academy of Sciences:
Strengthening Forensic Science in the United
States: A Path Forward (2009), evaluating the
P
P
3500
field of forensic science. The report raises many
issues and questions echoing debates, which have
a long history – both within forensic science
disciplines and beyond. These are issues such
as: that the current forensic science system is
not based on science; that the analysis is frequently subjective; and that the same technique
used by different analysts can lead to different
results, just as the same technique used by the
same analyst on the same sample at different
points in time and in different contexts can lead
to a different result, and that anecdotal information, phrased like “In my training, knowledge,
and experience.. . .” makes its way into the courtroom. Reactions to the NAS report have been
swift and all across the board from those that
deny that there is anything wrong, to those who
have adopted the NAS report and endeavored to
suggest new modes of operation to improve on
the lack of scientific rigor and culture of forensics, especially within criminalistics (e.g.,
Mnookin et al. 2011).
The report does however not deal with the
fundamental philosophical basis of the forensic
process, Therefore there is no recognition of the
philosophical roots of this debate, and the fact
that the issues have existed in different forms
since well before the development of the different
disciplines of forensic science. In order to understand the controversies a proper historical and
theoretical context is needed. The purpose of the
entry is therefore to explore the major philosophical principles that have inspired and shaped the
forensic process. Given that the word “forensic”
in contemporary use has two meanings: (a) as
a form of legal evidence and (b) as a short term
for forensic science or criminalistics (forensics),
this entry will inspect the philosophical history
along two trajectories. Firstly, through
a visitation of the history of legal rhetoric – the
predecessor of modern law – secondly, through
the history of the growth of science and the scientific method and the philosophical “goods”
being, respectively, embraced and discarded in
the process. Ultimately, it is in the combined
history of these two trajectories of thought that
the philosophical basis of the forensic process is
to be found.
Philosophical Basis of the Forensic Process
Fundamentals of Legal History
The Art of Rhetoric: The Foundation of
Modern Law
Etymology is a proper place to start when digging
for the roots and the philosophical basis of the
forensic process. The word “forensic” originates
from the Latin adjective forensis, meaning “of or
before the forum.” In Roman times, a criminal
charge meant presenting the case before a group
of public individuals in the forum. Both the person
accused of the crime and the accuser would be
represented by an advocate giving speeches on
their behalf, presenting their side of the matter.
The individual with the best argument and delivery
would succeed. This origin is the source of the
modern usage of the word “forensic” to denote
a form of legal evidence (forensic evidence) and
as a short representation of “forensic science,” that
is, science related to and of relevance to courts
(forensics). Historically, courts of law (forums or
public assemblies) were intimately connected to
rhetoric. In classical Greece, rhetoric was taught
to citizens so that they could defend themselves in
the court and deliberate in the assembly. In Rome,
advocates (called logographs) used oratory to plead
on behalf of their clients before the forum. Law, as
a subject of its own, did not emerge until the
twelfth century (Schoeck 1983), and law students
continued to be educated in rhetoric well into the
Early Modern period (Balkin 1996). A key element
of the art of rhetoric was how to find and present
arguments in a case to be presented before a court
of peers. To guide the search for arguments and
proofs to their fact, which would persuade the court
to find the case presented by the rhetorician the
most believable and worthy of support, the Ancient
Greeks developed a meticulous system or inventory of possible forms and modes of argument
to this end, also known as topoi or in English
“topics.” The following presents a brief history of
the origins of the topics and their incorporation into
modern legal procedure and the forensic process of
which the topics form the backbone.
The Topics of Classical Rhetoric
The first attempts at setting out guidelines for
those who were to investigate and present cases
Philosophical Basis of the Forensic Process
before a tribunal of decision makers seems to be
dated back to one of the deemed founders of the
modern art of rhetoric, Hermagoras of Temnos
(300 BC). Hermagoras divided the materials of
rhetoric into two parts: thesis and hypothesis.
A thesis involves an abstract, general question,
whereas a hypothesis involves a question
concerning concrete particulars. Hermagoras
provided a list of circumstances which needed
to be addressed in order to form hypotheses and
to be able to demonstrate their truth-value, in the
form of a list of questions. He contended that no
hypothetical question or questions involving particular persons and acts can arise without reference to these circumstances, and no
demonstration of such a question can be made
without using them.
The questions defining Hermagoras’ circumstances are known only in Latin because they
arose as part of the theological writings of St.
Augustine recognizing Hermagoras as the original source. St. Augustine quoted Hermagoras’
questions in Latin as follows: quis, quid, quando,
ubi, cur, quem ad modum, quibus adminiculis
(who, what, when, where, why, how, in what
way, and by what means/which aids) (Robertson
1946). Hermagoras’ list of questions for producing hypothesis and proof is a mnemonic device
intended to provide a systematic methodology to
the task of producing rhetorical evidence (proof).
This form of mnemonic belongs to a part of the
methodology of the discipline of rhetoric, derived
from Aristotle, known as topoi – or “the topics” –
of which he was the first to write systematically
about.
The word “topic” originates from the Greek
word topos (pl. topoi), meaning “place,” and
refers to the “places” (on the mind’s horizon)
where one can go to look for and discover the
relevant arguments (i.e., get ideas for argumentation). In Latin, the topics are referred to as loci
(places) or loci communes – meaning the “common places.” The use of the word topos/topoi
may refer to the ancient mnemonic principle of
using “spatial locations” or “activities” by
which to remember even complex or huge
amounts of information. One may for instance
remember a long list of dinner guests by
3501
P
imagining them around the table. Generally,
the word “topics” is commonly used to refer
both to the actual “places” where one can go to
find material for arguments and to the methodology of the process of searching these places.
The topics have since their very conception in
classical rhetorical theory been a very confused
concept, not easily defined and incorporating
a bewildering diversity of meanings. Initially,
“Topics” is the title of a treatise of Aristotle on
the art of dialectic – the invention and discovery
of arguments in which the propositions rest upon
commonly held opinion or what he termed
endoxa (Aristotle 384–322 BC). These commonly held beliefs, the topoi or “topics,” are
not merely popular notions held by any given
individual in a city or culture about any and all
subjects. Rather, they are elements of reason
upon which those who conscientiously dispute
agree in principle, i.e., that which is firmly
embedded within opinion or belief among
those who engage in disputation. The topics
belong according to Aristotle to the art of rhetoric, in which they inform the invention and
discovery of arguments (Aristotle and LawsonTancred 1991).
Aristotle’s theory of the topics emphasized a
particular aspect of evidence related to forming
and proving hypotheses. Hypotheses was thought
to belong to the form of inference pertaining to
“practical knowledge” (techné and phronesis)
which was opposed to “theoretical knowledge”
(episteme). Aristotle contended that when
lacking the certainty and necessity of the true
premises described in the syllogism and logic,
one needed to take point of departure in commonly held beliefs which could be viewed as
“law-like” in nature because of their general
acceptance as “true.” The concept of topics to
Aristotle was thus both (a) a minutely devised
inventory or system of cases, in which one
might profitably take point of departure when
devising the argument and providing for them
and (b) the theory of hypothesis building and
testing within the area of rhetoric, to which they
pertained. Thus when a “belief” is considered of
a high level of generality, i.e., being “common
sense”, it belongs to the topoi.
P
P
3502
The Philosophical History of the Topics
The ideas of rhetoric conceived by the Ancient
Greeks were spread and became known to the rest
of the European continent up through the Middle
Ages and during the Renaissance. They were
distributed through the works of Roman advocates and orators and the treatises they drafted
in the language of church and scholars: Latin.
A key figure in the history of European rhetoric
and philosophy, linking Classical philosophy and
Medieval and Renaissance thinking, was the
Latin rhetorician Cicero (106–43 BC) and his
works Topica, De Oratore and De Inventione.
Cicero divides his work on the circumstances
into two parts: attributes of the person (personae)
and attributes of the act (factum). As attributes of
the person, he lists nomen, naturam, victum,
fortunam, habitum, affectionem, studia, consilia,
facta, casus and rationes as the topics to consider
in building a credible statement of the case, but
more importantly, to develop the arguments
which should make the main body of the text or
speech, and according to which the logical arguments of the case should be elaborated. Attributes
of the action are of three kinds: those concerning
the performance of the action itself, those related
to the action, and those consequent upon the
action; only the first of these reflects the system
of Hermagoras. In relation to the action itself,
Cicero names locus, tempus, modus, occasio
and facultas (Robertson 1946).
The most important discussion of the topics,
considering their relation to legal rhetoric, is
made by Anicius Manlius Severinus Boethius
(ca. 480–524 AD) in the early Middle Ages.
Boethius,
in
his
influential
writings
De differentiis topicus, makes the seven topics
(or circumstances) fundamental to the arts of
prosecution and defense. The loci proper to rhetoric are confined to the attributes of the person
who is called for judgment and those of the act or
statement for which he is to be judged. All arguments, either in the defense of the person or for
his prosecution, must be taken from these materials. Thus, Boethius associated the topics with
the process of determining the extent of guilt of
a person accused of violating legal rules or customs, which may be significant to the adoption of
Philosophical Basis of the Forensic Process
these by papal forums and priests in their work
with ascertaining as accurately as possible the
magnitude of each sin confessed to them.
The topics arise in a new guise in the pastoral
and theoretical theology in the thirteenth century,
as Mnemonic verses consisting of variations of
Hermagoras questions. This is evident in the 21st
canon of the Fourth Lateral Council arranged by
Pope Innocent III, which stipulated that confessors should diligently seek out et peccatoris
circumstantias et peccati (the circumstances of
the sin and of the sinner) – resembling to no small
extent Cicero’s person and act personis et
negotiis – in order that they might justly weigh
the sins confessed to them and administer suitable
remedies. This appropriation of the topics to
theological matters and theological inquisitorial
processes can be attributed the theological writings of St. Augustine (Augustine of Hippo) and
Thomas Aquinas. Verses made of the questions in
different variations, afforded a flexible instrument for interrogation. It allowed the priest to
associate conventional theological divisions of
the sins with the questions so that they afforded
a frame of reference against which to place his
theological knowledge. In this way they provided
a practical device for spontaneous and quick
analysis (Robertson 1946, pp. 6–7).
The topics were then revived by Renaissance
rhetoricians, such as Thomas Wilson, The Art of
Rhetorique, who made an English verse of them:
“Who, what, and where, by what helps, and by
whose: Why how, and when, doe many things
disclose” (cit. in Robertson 1946, p. 14).
A principle of such fundamental importance to
both rhetoric and theology soon spread to other
areas of literary work and studies. Italian writers
adapted them to historiography. One of the
important figures in Renaissance rhetoric
Giambattista Vico (1668–1744), trained in jurisprudence, but widely read in Classics, philology,
and philosophy, all of which informed his highly
original views on history, historiography, and
culture, revisited the topics in his mature work:
Scienza Nuova (or The New Science). In this
work he argued for a central role of the topics in
all disciplines including the sciences – even
mathematics.
Philosophical Basis of the Forensic Process
Vico’s point was that “topics” is not to be
learned from a book but is more akin to
a disposition whereby one has analogies and
arguments at one’s hand. Thus for the doctor the
topics provide ready mental access to the stock of
medical precedents that have historically enabled
successful diagnosis and treatment, and to the
mathematician pressing the boundaries of knowledge further, the topics consist of the accepted
theories preceding the development of new
hypotheses. Vico found that the principle
described in the use of the topics was to be considered a necessary part of making judgments
about everyday life in general, and that their
cultivation is required for creative discovery
and quick action in all disciplines, from
physics to law.
In his own time, Vico was relatively unknown,
but from the nineteenth century onward where the
methodology and philosophical content of rhetorics and linguistics converged with other developments in the history of ideas giving rise to new
critical theories in the philosophy of science, his
views found a wider audience. Today his influence is widespread in the humanities as well as in
the social sciences. Before delineating this turn,
tying the knot between the two histories of ideas
by describing how they inform the growth of
forensic science and the philosophical basis of
the forensic process as it looks today, a brief
history of the fundamentals in the development
of science is called for.
Fundamentals of the Philosophical
History of Forensic Science
Science and the Conjectural Model: The
Foundation of Forensic Science
Forensic science is intimately related to investigative inquiry through the forensic process.
Investigation comes from Latin investigationem
meaning “a searching into, a searching for.” The
noun stems from investigare “to trace out,
search after” which again is composited from
in “in, into” and vestigare “to track, trace”
from vestigium “footprint, track.” The etymology of the word “investigate” carries the seed to
3503
P
an understanding of the philosophical roots of
crime investigations and the forensic process, by
referring to “tracks and traces” and “footprints,”
since the roots of science and of medical, legal,
and forecasting sciences in particular, may
well be rooted in the first efforts of interpreting
footprints in the sand or the tracks of wildlife in
the forests, with the aim to hunt it down and
slay it. The point being that the philosophical
basis of the forensic process has a deep-rooted
connection to semiotics and other sciences
informed by what may be termed the “semiotic
paradigm” – which we shall get back to shortly.
Firstly, let us briefly return to etymology as
a starting point for the philosophical basis of
the forensic process stemming from the history
of science.
The root of the word “science” is the Latin
scire, to know, and hence scientia, or knowledge.
Before the late sixteenth and early seventeenth
centuries, the word denoted any kind of knowledge, undiscriminating between different forms
or types of knowledge. However, with the
Enlightenment and the introduction of scientific
method plus the resurgence of learning based on
classical theory during the Renaissance, the word
scientia came to mean knowledge of the natural
world and became increasingly associated with
natural science.
The Italian historian Carlo Ginzburg traces the
trajectory of what he terms the “conjectural or
evidential paradigm” throughout European history, attempting to excavate the historical lineage
of “disciplines based on reading the evidence”
(Ginzburg 1980, p. 14). Since this history is
rarely researched, the exposition of this part of
the philosophical history of the forensic process
is resumed from his seminal piece: Morelli,
Freud and Sherlock Holmes: Clues and Scientific
Method (1980). The roots of this paradigm,
Ginzburg argues, can be found well beyond the
civilizations of Ancient Greece with the earliest
hunters and the skills they developed for reading
signs in nature to track down the game and in the
Mesopotamian cultures different forms of divination. These skills he argues, played a significant
role in the invention of writing and the capacity
for abstract thinking, which is based on the use of
P
P
3504
purposefully crafted signs to represent inherent
meanings. The common characteristics of these
roots of skills leading onward to modern science,
are that, both traditions, hunting and divination,
“require minute examination of the real, however
trivial, to uncover the traces of events which the
observer cannot directly experience” (Ginzburg
1980, p. 13). The way hunters used their senses –
sniffing, listening, and observing – to give meaning and context to the slightest trace, “in the
shadowy and treacherous clearing,” Ginzburg
argues, endowed them with the ability to make
complex calculations in an instant. Hunters
inferred from part to whole, effect to cause. The
Mesopotamian model of divination gradually
intensified this tendency to generalize from
basic facts, thus strengthening this mode of inferring cause from effect. The model for explanation
and divination decipherable in Mesopotamian
texts is one which could be oriented toward
past, present, or future, depending on which
form of knowledge was called upon. Seeking
signs to form knowledge oriented toward the
future, was divination proper. Orientation toward
the past was related to jurisprudence or the body
of legal knowledge. Divination oriented towards
the past, present, and future in concert was that of
medical knowledge of symptoms. The latter
contained all of the orientations by way of its
double character of diagnosis, explaining past
and present, and prognosis, suggesting the likely
future following from the illness running its
course.
The conjectural paradigm lies at the root of
modern science in general. It is, however, the
development into scientific disciplines of these
original practical arts (hunting, divination,
medicine, and law) of deciphering various signs
and symptoms as well as individual cases to infer
generalized meanings, which serves to characterize the philosophical basis of the forensic process
in the Modern Age. It is however not well
researched nor generally acknowledged how
these artistic aspects of producing knowledge
inform the growth of science and ultimately
forensic science. Retracing this history is however the goal in the following.
Philosophical Basis of the Forensic Process
Ancient Greek Philosophy: The Cradle of
Modern Science
What characterized the Greek culture and the
city-states, of which Modern European legal culture and science are heirs, was the emergence of
a group of disciplines which all “depended on the
deciphering of various kinds of signs, from symptoms to writing” (Ginzburg 1980, p. 14). Passing
on to the civilizations of ancient Greece, these
disciplines changed considerably, giving rise to
new lines of study such as history and philology
while acquiring also independence from older
disciplines, such as medicine. In and through
these disciplines, one finds the seeds of the development of a general scientific stance or method.
Through these disciplines the “body” (e.g., medicine), “speech” (e.g., rhetoric, philology), and
“history” were all for the first time subjected to
dispassionate investigation, excluding the possibility of divine intervention. An important part of
this change can, according to Ginzburg, be
ascribed to a model of knowledge acquisition
based on symptoms and signs – the evidential or
“conjectural model” of knowledge.
This development is most evident in the case
of Hippocratic medicine. Hippocrates’ model of
medicine based its central methods on the concept of “the symptom”, in Greek semeion (hence
semiotics). The central tenet in the Hippocratic
model was the idea that by carefully observing
and registering every symptom, it was possible to
establish precise histories of each disease, despite
the fact that the disease as an entity would remain
intangible. The insistence on the evidential nature
of medicine, prevalent with the Hippocratic
model, probably stemmed from a distinction
between the certainty of divine knowledge and
the provisional, conjectural nature of human
knowledge, expounded by the Pythagorean doctor, Alcmaeon of Croton in his work Peri Physeos
(or On nature). This was the first scientific treatise
devoted to the art of medicine. The work itself
has not been preserved, but its contents and chief
findings have been passed on by commentators
and collectors of ancient knowledge such as Aristotle, Theophrastus, and Aetius. From these it is
known that Alcmaeon held that if reality was not
Philosophical Basis of the Forensic Process
necessarily clear – then by implication – it was
right to proceed by building up knowledge of the
whole from the parts, using conjecture.
Proceeding by using conjecture was in fact the
most common approach in a number of spheres of
activity in Ancient Greece, where physicians,
historians, politicians, potters, joiners, mariners,
hunters, fishermen, and women in general were
held to be adept in the vast areas of conjectural
knowledge. This conjectural or semiotic paradigm, continued however to be merely implicit
up through the ages, since it was completely
overshadowed by Plato’s theory of knowledge,
holding sway in more influential circles, hence
shrouded in more prestige. Plato’s theory of
knowledge emphasized the existence of an ideal
(i.e., ideational) world of universal untainted
truth existing independently of the human species
and lost to the human subject. To Plato, knowledge could only be regained or recollected
(anamnesis) through inspired contemplation and
philosophy.
The Science of the Universal and the
Problem of the Individual
While Ancient Greece and the developments of
disciplines of dispassionate investigation of various subjects and forms, may be the cradle of
modern science, the genesis of scientific method
is generally acknowledged to be the Middle
Ages. The emergence of a truly scientific paradigm is commonly associated with Galileo
(1564–1642). The scientific paradigm based on,
but outliving, Galileo, provided significant
changes in the concepts of “rigor” and “science.”
The model of Galilean science was fundamentally different from any of the disciplines using
conjecture to form knowledge of their subject.
None of these would meet the criteria of scientific
inference set out by the Galilean approach. The
ancient disciplines were concerned with the qualitative, the individual and the case particular, i.e.,
the individual body, document or situation itself.
Galilean physics were altogether different. It
used mathematics and experimental method to
gain knowledge of the universal order of nature.
This was in stark contrast to the conjectural
3505
P
model of the Ancient disciplines in the results of
which there was always an element of chance
(Ginzburg 1980, p. 15). In the science that grew
from the Galilean model, chance was sought
tamed by a dependence on measurement and
repetition of results. Creating knowledge by
induction from experimental design and observation of results, science increasingly became associated with the need to measure and repeat
phenomena in order to draw rigorous conclusions
about the phenomenal world. Paradigms following the Galilean tradition were concerned alone
with knowledge of the universal and generalizable, while being unable to say anything about the
individual and particular. The rise of the scientific
paradigm resulted in a division of fields of knowledge which would meet the criteria of scientific
inference set by and essential to the Galilean
approach, and those who could not. A host of
the classical disciplines could not meet or make
do with the criteria set for science by the new
scientific paradigm. However, one discipline in
particular diverged from the rest, proclaiming its
scientific status while drawing on a model of
inquiry stemming from the conjectural or semiotic model – medicine.
The philosophical history of forensic science
leads in many ways directly to an origin in medicine. This is generally acknowledged not least in
the fact that subjects of medicine such as the
autopsy and toxicology are some of the first
forms of knowledge enrolled in the service of
law to decide on the issues of identifying and
discriminating between a criminal act and
a legitimate or natural one. However, there are
more subtle philosophical roots to be discerned
from the relation between medicine and the
forensic process as a process of inquiry and
theory of knowledge.
The Rise of the Semiotic Model
The trajectories of thought and ideas informing
the move from hunter to the scientific detective
are manifold and diverse. It is therefore not possible to delineate them in detail here. However,
their different results may be indicated by
zooming in more directly on the common
P
P
3506
denominator through which they all at some point
pass – medical science and the paradigms
informing it. The roots of this paradigm may,
following Ginzburg, be traced to the contemporaries of Galileo and described through
a historical event in which the medical expertise
was enrolled to decide on the nature of a new
phenomenon: a two-headed calf born outside of
Rome in April 1625. The event became object for
discussion among a group of naturalists in the
Vatican Belvedere Garden, which was
documented in detail by Giovanni Faber (a friend
of Galileo and secretary of the Lincei Academy,
a center of a thriving intellectual group founded
in 1603 by Frederico Cesi). The discussion was
attended besides Faber, by another friend of Galileo, a Cardinal, Pope Urbano VIII, and the pope’s
chief medical man, a physician from Siena,
Giulio Mancini (Ginzburg 1980, p. 20).
Mancini was besides being a brilliant diagnostician renowned for his ability to divine the seat
of disease by a rapid look at the patient, also
a keen intellectual and the author of a book
about art recognition titled: Considerazioni sulla
pittura (or Some considerations concerning
painting as an amusement for a noble gentleman
and introducing what needs to be said)
(Mancini ed. Marucchi 1956–1957 ref. in
Ginzburg 1980). Part of this book was devoted
to the recognition of paintings, and set out
a method for telling originals from copies, by
recognizing the particular strokes of the hand of
the master in areas of the painting that were
swiftly executed. The underlying assumption of
Mancini’s views, were never stated, but were
nevertheless important to the later development
of connoisseurship in the century to follow and
its role in transmitting the semiotic model to
other areas of European cultural life and
knowledge, Ginzburg argues. The core assumption of deep relevance to the rise of the conjectural paradigm was that there are ineradicable
differences between a canvas painted by the
hand of Rafael, for example, and any copy of it
(Ginzburg 1980, p. 20).
The discussion between the parties participating in the event represents a historical example of
differing perspectives of different disciplines on
Philosophical Basis of the Forensic Process
the individual case and its relation to the general,
as well as a moment in the history of a converging
of differing paradigms, embodied in the practitioner of medicine and art connoisseur Giulio
Mancini. The first question addressed by the
group was whether the animal should count as
one animal or two. For the physicians the feature
distinguishing the individual was the brain,
whereas for followers of Aristotle it was the
heart.
To some extent, Mancini in the discussion of
the two-headed calf represents a point of contact
between the divinatory approach of the conjectural paradigm (as a diagnostician and
a connoisseur) and the generalizing model of
the Galilean approach to physics (as anatomist
and naturalist). He represents the emergence of
a generalizing paradigm within medicine while at
the same time being the skilled artistically
inclined diagnostician, reading the signs of the
disease in the bodies of the patient or identifying
the hand of the master versus that of the copyist in
a piece of art. By adopting the role of the physician, Mancini was ascertaining the character of
the individual as a “type” (i.e., within a scheme of
classification) rather than of the character of the
individual calf itself. The physician, in other
words, participated with a view to arrive at
a more accurate definition of what was normal –
and therefore repeatable for the individual of the
species, i.e., establishing the “common character” of it rather than the one peculiar to it
(Ginzburg 1980).
The medical paradigm that developed through
the eighteenth and nineteenth centuries thus
consisted of two paradigmatic models – the
anatomical model and the semiotic model
(Foucault and Sheridan 1973). The first aspired
for a system of knowledge which could be generalized and ascertained through empirical observation and classification (continuing the natural
history starting with Aristotle) and making use
of the organ associated with the sciences of
the time, which could not be reduced to
numbers and mathematical calculation: sight.
Vision and observations became synonymous
with “objectivity” – the defining character of
empirical science (Daston and Galison 2007).
Philosophical Basis of the Forensic Process
The second model of medicine aspired for an
effective diagnostics which were able to identify
and treat illnesses in individual cases (Foucault
and Sheridan 1973) and thus keep the subject
alive, healthy, and maintain a productive population to the benefit of the state (Foucault and
Sheridan 1979).
The conjectural or semiotic model of medicine
was, however, intimately connected with knowledge of the general, in the form evoked by the
anatomical model. The latter was based on the
dissection of corpses and the description of the
anatomy of the human body. This practice was
used to create knowledge of the general character
of the human body, which in turn could be used as
a general norm against which pathology could
measure the particular body’s deviance or abnormalities to achieve new knowledge of pathologies and their effects and signs in the human
body. The anatomical model, and its knowledge
of the general achieved from observations and
generalizations of many individual bodies was
promoted and spread through the medical atlases,
which became a significant artifact through
which medical science flourished (Daston and
Galison 1992). It did so in coexistence and
mutual development with a semiotic model of
medicine, based in the clinic, and the experience
and knowledge of the emergence, courses, symptoms, and effects of different illnesses, which
arose from the diagnostician’s effort to device
the proper treatment in each individual case
(Foucault and Sheridan 1973). In a similar way,
the discussions and the tiny incisions made in the
animal to reveal the internal organs representing
the roots of the modern scientific autopsy, which
developed during the Renaissance, represents the
meetings of divination and science in the anatomical model of medicine. In a review article of the
history of the autopsy, King and Meehan (1973)
find its roots in former practices of divination
using animal organs, the so-called haruspicy.
Autopsy, which literally means “seeing for oneself,” reflects the central role of sight referred to
above, as the key component of the scientific
method of “observation” and scientific objectivity. “Seeing,” however, here has roots of
a different kind, namely, that of ancient diviners
3507
P
looking for signs to foretell the future by the
reading of organs and tissues of dead animals.
The medical atlases of the nineteenth century
thus also had a predecessor in Babylonian models
of livers with diagrammatic markings used for the
instruction of diviners.
The predicament of medical science, that is, its
apparent lack of absolute scientific rigor and certainty while aspiring to be a science proper, can
be said to have two basic reasons. On the one
hand, there is a discrepancy between theory and
practice. Theoretically adequate descriptions of
particular diseases are not necessarily so in practice, where diseases can present itself in different
ways in different bodies. On the other hand,
knowledge of disease in the living body remains
indirect and conjectural, since it is per definition
out of reach (Ginzburg 1980, p. 21).
Medicine and the Birth of the Detective
It was however from the diagnostic branch of
medicine that the ideas giving birth to both detective fiction and forensic science originated. In the
late nineteenth century, there was a general
upsurge in the semiotic paradigm, giving birth
to the practice of art connoisseurship, the theories
of psychotherapy, and the detective novel. There
are actual biographical and inspirational links
between the key figures representing each of the
three, in the form of Giovanni Morelli, Sigmund
Freud, and Arthur Conan Doyle. The connecting
factor between them is medicine. Freud was
a doctor, Morelli had a degree in medicine, and
it is generally acknowledged that the inspiration
for Sherlock Holmes was Sir Arthur Conan
Doyle’s esteemed professor in medical school
(he was a doctor before he started writing) Professor Joseph Bell. Bell was renowned for his
diagnostic abilities, and there are key passages
in the stories of Sherlock Holmes, for example.
The Cardboard Box (1892), where Holmes
uses the actual methods of Morelli. Furthermore
the peculiar similarity of the activities of Holmes
and Freud have been discussed in literature (e.g.,
Marcus 1976).
At the same time however, they represent
a point in history where the semiotic model (or
the conjectural paradigm) resulted in different
P
P
3508
ideational developments, all however sharing
a common feature, namely, that they were based
on an interpretive method where one takes the
evidence at hand as “signs” which stands for
something else otherwise inaccessible for observation. In psychotherapy, art connoisseurship,
and detective fiction, methods were devised
which made tiny details the key to a deeper reality
or knowledge hereof that are inaccessible by
other methods. These details may be symptoms
for Freud, clues for Holmes, or features of paintings for Morelli, but in all three cases they are
seen as evidence to inferences about the nature
and characteristics of something hidden to the
inquirer: the psyche, the identity of the criminal
and the structure of the criminal event, or the
identity of the painter and the characteristics of
his paintings, behind a given piece of art. Sir
Arthur Conan Doyle’s character Sherlock
Holmes has been deemed inspirational for one
or more of the pioneers in forensic science in
the area of criminalistics, such as Bertillon and
Locard. Whether or not this is a historical fact, we
encounter also in the foundation of the field of
forensics the medical common factor, as
Locard’s teacher and mentor was Alexandre
Lacassagne, a physician and chair of Médecine
Légale de la Faculté de Lyon, and, Bertillon the
son of a medical professor.
However, developments within the natural
and social sciences occurring at the same time
were also influenced by the conjectural paradigm
and the problem of the individual. Thus the semiotic or conjectural paradigm generated forensic
science from several trajectories of thought,
which in concert gave rise to the idea of the
scientific detective and the forensic process.
The Social Sciences and the Rise of Technologies
of Government
The problem of the individual in the generalizing
Galilean model of science was encountered from
another field of knowledge, namely, that of the
emerging social sciences. During the eighteenth
and nineteenth centuries, a tendency evolved for
the state to impose a close-meshed net of control
on society: one based on knowing its individual
members and submitting them to forms of
Philosophical Basis of the Forensic Process
surveillance that became increasingly more subtle and that was based on knowledge of its population, statistical as well as biographical
(Foucault 1970; Foucault and Gordon 1980; Foucault et al. 2007, 2008; Foucault and Sheridan
1979).
With the emergence of a capitalist economy
based on production in the late eighteenth and
early nineteenth century, new concepts of property and the introduction of a still greater number
of punishable offenses and punishments of more
severity were introduced. Thus, prison systems
were built up to contain the resultant growing
population of criminals. With the rise in the number and types of crimes came the problem of
identifying repeat offenders. This problem
became a specific subset of the more general
one of keeping a complete and general track on
the whole of society, that is, keeping order
through knowledge. It also represents the point
where the models derived from Galilean physics
start to impinge on and interact with the semiotic
model of medicine, in the development of new
forms of control and methods for identification,
devised in the same period of the 1800s as the
detective novel, psychoanalysis, and connoisseurship evolved.
This was evident in two competing systems
for identification initiated or developed by
Francis Galton (1822–1911) and Alphonse
Bertillon (1853–1914), respectively. In both
cases, however, science meets conjecture to
give rise to the methods of scientific investigation
based on classifications of the individual and
interpretations of effects of specific causes in
particular cases.
In the first case, the system of Bertillionage,
a classification scheme for identifying repeat
offenders through anthropometric measures
plus the description and classification of physical characteristics, was based on the idea of
probability statistics used within astronomy
and physics to get a measure of errors, introduced to the study of society and crime by
Adolphe Quetelet (1796–1874) in his
conceptualisation of a “social physics”.
Quetelet’s work provided a major inspiration to
Bertillon and the anthropometric method, by
Philosophical Basis of the Forensic Process
suggesting that while single measures might be
just probative, the higher number of specific
measures made per individual the higher the
probability of identification.
In the second case, the system of fingerprint
identification introduced by Francis Galton, yet
not originally discovered by him (for a history of
fingerprint identification, see Cole 2001;
Sengoopta 2004), was based on the idea that the
individual has an identity that may be recognized
in his every characteristic, even the most imperceptible and the slightest, and that there will be an
internal norm or “typus,” which maintains the
variety of each species within its limits. The
philosophical input, as well as the study of fingerprints as the empirical example of the rule,
was made by a physiologist founding the subject
of study of organ tissues: Purkyné in a work
in 1923.
Bertillon, being the son of a medically educated statistician and inspired by an influential
figure in the emerging science of society especially criminology developed by Adolphe
Quetelet, represents the meeting point of the conjectural model derived from medical semiotics
and that of divining probability through the science developed to this end: statistics. Galton’s
discoveries of the individuality of the fingerprint
and its possible practical implications, on the
other hand represent the point of contact between
old divinatory traditions; the problem of identification of the complex society and the risks posed
by the transgressor of laws and regulations, and
the anatomical model of medicine and its attempt
at understanding the relation between individual
and general type, in order to diagnose individual
symptoms and secure the proper treatment in
each individual case.
Forensic science arose from the interconnections of these ideas and trajectories of thought
spanning detective fiction, graphology, art connoisseurship, statistics, sociology, criminology,
administrative practices, psychology, linguistics,
history, etc., as they were developed and
promoted in the works of Hans Gross, Alexandre
Lacassagne, Edmond Locard, Alphonse Bertillon,
Francis Galton, and many more in a range of subdisciplines to emerge within the field.
3509
P
New Philosophies of Science: Conjecture and
“The Scientific Man”
In the late nineteenth century and early twentieth
century, the philosophy of science saw new
developments which were influenced by processes rooted in the conjectural or semiotic paradigm. One of the ways in which semiotics was
developed into scientific theories and new
approaches to knowledge were through the
ideas of American Pragmatism. In addition to
being an exemplification of the broader upsurge
and explication of semiotics as a theoretical discipline, (which itself developed into a range of
different philosophical theories, for example,
critical theory, post-structuralism, hermeneutics,
reflexivity, etc.) this particular theoretical paradigm has overt links to the developments of
forensic science. The credited founder of Pragmatism Charles Sanders Peirce (1839–1914) was
deeply inspired by the detective figure in his
exposition of his thoughts on reasoning and the
“scientific man”, which again was part of a more
general theory of signs and sign processes as
a distinct theory of knowledge (Eco and Sebeok
1983). Besides, Peirce was and is generally
acknowledged as a natural scientist, above being
a philosopher, although his philosophy of science
is seminal.
Based among other on the figure of the scientifically reasoning detective fashioned by Conan
Doyle, Peirce developed a new concept of inference – abduction – which served to integrate the
approach to science initiated with Galileo and the
necessity of using incomplete knowledge and
conjecture – not least in face of lack of enough
data – to form hypotheses and predictive conditionals to be tested as a means of approaching
what could qualify as secure beliefs. Abduction is
Peirce’s designation of an inference which is
made from an observation to its possible explanation; that is, from evidence to a hypothesis or
conclusion – where the phenomenon to be
explained, through the evidence of it observed,
itself constitutes a significant part of the reason
to presume that the explanation is correct
(Lipton 2000).
Peirce held that all inquiry, including science,
is to be described as a struggle to move from
P
P
3510
irritating, inhibitory doubt born of surprise, disagreement, and the like, toward the reach of
a secure belief. Belief being defined as that on
which one is prepared to act, rather than as the
pursuit of truth per se. Scientific inquiry, to
Peirce, was thus just one part of a broader spectrum and spurred, like inquiry generally, by genuine doubt, rather than by mere verbal or
hyperbolic doubt, which he held to be fruitless.
Having grouped all forms of inquiry under the
same pursuit, he also devised a theory of inquiry
and logic, which introduced a new logic of inference to the description of science.
Peirce held that abduction, induction, and
deduction made a complementary whole which
together constitute scientific method. None of the
modes of inference could be dispensed with –
since abduction to Peirce is the only way new
knowledge can arise. Logic was based on all the
modes of inference, including abduction, which
he perceived as a precondition for the development of science. Peirce also replaced the conception of science as a collection of certified truths or
systematized knowledge with a definition of science as an inquisitorial attitude, the defining characteristics of science to be found in the approach
to inquiry and the controls put on it rather than in
the contents of its canons.
Being deeply inspired by Aristotle, whose concept of apagoge seems to entail all the things that
Peirce describes as retroduction or abduction and
who may be the original inspiration for his thinking (Aliseda 2006), he picks up on the same theme
that Vico (as mentioned earlier) raised during the
Renaissance: that all actions, whether practical
decisions, medical diagnosis or scientific inquiry
make use of the principle of the topics (i.e., those
culturally formed beliefs, habits of experience and
expectations to the order and workings of the
empirical world) to form judgments and beliefs
on which to act whether practically or theoretically
(Miner 1998). One can legitimately speak of pragmatism as a reinterpretation of Greek philosophy
about knowledge and thus a rekindling with the
roots of the development of the sciences and their
conjectural roots.
While Peirce is not a prominent figure in the
canons of forensic science – if at all present, the
Philosophical Basis of the Forensic Process
practices, which inspire his philosophical positions, are indeed those of the forensic practitioner. It therefore seems in keeping with the
micro-history of science and the conjectural paradigm delineated by Ginzburg, to point to the
pragmatist philosophy as an example of trends
in the philosophy of science, which are informed
by the conjectural and semiotic paradigm running
from Ancient Greek philosophy to some of the
developments in thinking about science making
way from the nineteenth century and onward, and
which are in keeping – however subtle and
mutely – with the philosophical basis of the
forensic process.
Conclusion and Further Perspectives
The Forensic Process: Conjecture and Science
or Conjectural Science?
The philosophical basis of the forensic process is
not to be discerned as a linear development of
thought but rather as a meshwork of ideas. However, the influence of the ideas delineated is clear
considering the principles brought to bear on the
forensic process by the founding fathers and key
figures in forensic science, for example: (a) the
principle of transfer: that two physical objects
coming into contact with each other will result
in minute traces, imprints or effects of this contact (also known as the principle of exchange)
commonly attributed to Edmond Locard (1928,
1930), but by him attributed to the inspiration
from Alexandre Lacassagne; (b) the goal of individualization based on the identification and/or
classification of the physical or chemical nature
of the evidence (Saferstein 1998); (c) the idea of
classification/individualization and the attempt to
determine the source of evidence (Kirk 1963);
(d) the idea of association linking person to
crime scene and leading to inferences of source
and target (Osterburg 1968), and (e) the purpose
of reconstruction of past events (DeForest et al.
1983) (for an overview description, see Inman
and Rudin 2001). All the central ideas of forensic
analysis are based on a combination of topical
inspiration (i.e., attempts at answering the 6W’s
essential to legal deliberation) and the
Philosophical Basis of the Forensic Process
foundational concepts of the semiotic paradigm
that infers causes from effects and which on the
basis of classifications of the nature of the physical world discerns individual characteristics of
the particular situation, object, or document.
Characteristics which in turn serves to discern
the history, meaning, and causes of the phenomenon under scrutiny, (i.e., of the physical evidence left over from an event (possibly
criminal) passed).
In much the same way as the clinician of early
modern medicine, the forensic practitioner may
be said to embody a point of contact between two
differing yet cognate paradigms – that of semiotics and that of Galilean natural science. Each has
differing goals and criteria for valuing truth
claims: one related to abstract disinterested truth
and one related to practical effects and rationality. Like the physicians of the late nineteenth
century, the forensic practitioner is a “diagnostician” and “expert connoisseur” with respect to
assisting the investigative inquiry and translating
the topics and questions of law into scientific
questions. He is also a “naturalist” and “empiricist scientist” with respect to analyzing artifacts
believed to adhere to an incident under legal
scrutiny and providing conclusions regarding
their identity based exclusively on knowledge
obtained through observable and verifiable data.
The first form of knowledge tends to be unspoken
and does not easily lend itself to formalization,
since nobody learns how to be a connoisseur or
diagnostician simply by applying the rules.
The allusion to one of the key works in Modern scientific philosophy by Karl R. Popper: Conjectures and refutations: The growth of scientific
knowledge (1963) in the heading above is
intended as a pun on words which may serve as
the key points of further perspectives, since the
question is not whether conjecture is part of what
informed the growth of forensics but rather: How
much conjecture is allowed and under what circumstances for forensic science to qualify as
science? The key concern of the forensic sciences
and thus the key philosophical problem with
which it grapples is that of individualization
(Kirk and Thornton 1974). Forensic science –
arising out of the conjectural paradigm and the
3511
P
semiotic model of medicine – is in many ways the
science of the individual par excellence. Seeking
from beginning to end of the forensic process and
through the fields of knowledge informing the
practical process of forensic inquiry (i.e., the
forensic process) answers, which may provide
individuation of both people and cases – of
“who” and “what” of the Ciceroean topics. In
this way, the forensic process is heir not only of
the key topics identified by Cicero for the presentation of a legal case but also of the quest of the
sciences build on the Galilean model and the
problems facing these in their attempt to explain
and describe the particular and the characteristics
of the individual, and the ways that phenomena
known through theoretical classifications may
play out in particular situations, hereby
ascertaining the span of characteristics to be associated with the phenomena in a given situation
and thus ascribed to it.
This problem is becoming increasingly acute
to the forensic sciences, due to emergence of
debates in scientific circles of is scientific status.
Yet in the philosophical history of the forensic
process of inquiry, can also be identified ideas
and principles stemming from a semiotic paradigm, which has grown large in the human and
social sciences, and which holds promise of
a new formulation of a model of science, which
may provide an alternative route to the problems
faced, by introducing a concept of scientific rigor
which is more “elastic.”
Forensic sciences is finding itself in a position
where new developments are imminent, if not
already in process. Developments which may
prove decisive for the future role and status of
forensic evidence, and which springs exactly
from the problem having faced sciences developed
in the image of Galilean physics, since the
beginning of the Modern Era. This dilemma is
coined by Ginzburg (1980) as that of choosing
between: (a) achieving significant results from
a scientifically weak position, or (b) to establish
a strong scientific position, but get meager results.
Rethinking and revisiting the philosophical
basis of the forensic process, and turning to key
thinkers within this paradigm for new approaches
to the dilemma may provide an alternative route
P
P
3512
bypassing the deadlock of the debates. The theories, models, and scientific content of the conjectural model that might inspire new ways of
theorizing and conceptualizing the foundational
and defining principles of the forensic process are
still however largely unexcavated, even though it
provides the philosophical basis and seed of
development for a broad range of contemporary
disciplines in all areas of science.
Related Entries
▶ Forensic Science and Criminal Inquiry
▶ Forensic Science Culture
▶ Forensic Science Effectiveness
▶ Forensic Science in the Nineteenth and
Twentieth Centuries
▶ History of Criminal Investigation
▶ History of Forensic Science in Policing
▶ Philosophy of Forensic Identification
▶ Prosecutorial Discretion
▶ Scientific Basis of the Forensic Process
▶ Scientific Evidence in Criminal Prosecutions
Recommended Reading and References
Aliseda A (2006) Abductive reasoning: logical investigations into discovery and explanation. Springer,
Dordrecht
Aristotle, Lawson-Tancred H (1991) The art of
rhetoric. Penguin Books, London
Aristotle (384–322 BC) Topics: (trans: Gaza T, CR
Editors)
Balkin JM (1996) A night in the topics: the reason of
legal rhetoric and the rhetoric of legal reason.
In: Brooks P, Gewirtz P (eds) Law, stories: narrative
and rhetoric in the law. Yale University Press,
New Haven
Cole SA (2001) Suspect identities: a history of fingerprinting and criminal identification. Harvard University
Press, Cambridge/London
Daston L, Galison P (1992) The image of objectivity.
Representations, No. 40, Special issue: seeing science.
pp 81–128
Daston L, Galison P (2007) Objectivity. Zone Books,
New York/Cambridge
DeForest P, Lee H, Gaensslen R (1983) Forensic science:
an introduction to criminalistics. McGraw Hill,
New York
Eco U, Sebeok TA (1983) The sign of three: Dupin,
Holmes, Peirce. Indiana University Press, Bloomington
Philosophical Basis of the Forensic Process
Foucault M (1970) The order of things: an archaeology of
the human sciences. Tavistock, London
Foucault M, Gordon C (1980) Power-knowledge: selected
interviews and other writings, 1972–1977. Harvester
Press, Brighton
Foucault M, Sheridan A (1973) The birth of the clinic: an
archaeology of medical perception. Tavistock Publications, London
Foucault M, Sheridan A (1979) Discipline and punish: the
birth of the prison. Penguin, Harmondsworth
Foucault M, Senellart M, Davidson AI (2007) Security,
territory, population lectures at the Collège de
France, 1977–78. Palgrave Macmillan, Basingstoke/
New York
Foucault M, Senellart M, Ewald F, Fontana A, Burchell G,
Collège de France (2008) The birth of biopolitics:
lectures at the Collège de France, 1978–79. Palgrave
Macmillan, Basingstoke/New York
Giambattista V (transl. Marsh D 1999) 1744 New Science.
Penguin Books, London/New York/Victoria/Ontario
Ginzburg C (1980) Morelli, Freud and Sherlock Holmes:
clues and scientific method. Hist Workshop
J 9(1):5–36
Hookway C (1985) Peirce. Arguments of the Philosophers
series. Routledge, London/New York
Inman K, Rudin N (2001) Principle and practice of criminalistics: the profession of forensic science, Protocols
in forensic science. CRC Press/Taylor & Francis
Group, Boca Raton/London/New York
King LS, Meehan MC (1973) A history of the autopsy:
a review. Am J Pathol 73(2):514–544
Kirk PL (1963) The ontogeny of criminalistics. J Crim
Law Criminol Police Sci 54:235–238
Kirk PL, Thornton JI (1974) In: Thornton JI (ed) Crime
investigation, 2nd edn. Wiley, New York/London
Leff MC (1983) The topics of argumentative invetion in
Latin rhetorical theory from Cicero to Boethius.
Rhetorica J Hist Rhetoric 1(1):23–44
Lipton P (1991) Inference to best explanation. Routledge,
London/New York
Locard E (1928) Dust and its analysis. Police J 1:177
Locard E (1930) The analysis of dust traces Am J Police
Sci 1: pp. 276–298, 401–418, 496–514
Markus S (1976) Introduction to A. Conan Doyle. In:
Conan Doyle A (ed) The adventures of Sherlock
Holmes: a facsimile of the stories as they were first
published in the Strand magazine. Schocken Books,
New York, pp x–xi
Miner RC (1998) “Verum-factum” and practical wisdom
in the early writings of Giambattista Vico. J Hist Ideas
59(1):53–73
Mnookin JL et al (2011) The need for a research culture in
the forensic sciences. UCLA Law Rev 58:725–779
National Academy of Sciences (2009) Strengthening
forensic science in the United States: a path forward.
National Academies Press, Washington, DC
Osterburg JW (1968) The crime laboratory: case studies of
scientific criminal investigation. Indianna University
Press, Bloomington
Philosophy of Forensic Identification
Popper KR (1963) Conjectures and refutations: the growth
of scientific knowledge. Routledge, London
Robertson DW (1946) A note on the classical origins of
“Circumstances” in the medieval confessional. Stud
Philol 43(1):6–14
Saferstein R (1998) Criminalistics: an introduction to
forensic science, 6th edn. Prentice Hall, Englewood
Cliffs/New York
Schoeck RJ (1983) Lawyers and rhetoric in sixteenthcentury England. In: Murphy J (ed) Renaissance
eloquence: studies in the theory and practice of renaissance rhetoric. University of California Press, Berkeley
Sengoopta C (2004) Imprint of the Raj: how fingerprinting
was born in colonal India. Pan Books, Basingstoke/
Oxford
Tilstone W, Hastrup M, Hald C (eds) (2012) Fisher’s
techniques of crime scene investigation. First international edition. CRC Press/Taylor & Francis Group,
Boca Raton
Toulmin SE (1958) The uses of argument. Cambridge
University Press, Cambridge
Victorinus M (1863) Explanationum in rhetoricam
M. Tullii Ciceronis libri duo. In: Halm C (ed) Rhetores
Latini minores. Teubner, Leipzig
Philosophy of Forensic Identification
Ton Broeders
Faculty of Law, Institute for Criminal Law
and Criminology, Leiden University, Leiden,
The Netherlands
Maastricht University, Maastricht,
The Netherlands
Overview
In most if not all criminal investigations, the
collection, examination, and interpretation of
physical evidence plays a major role. Material
traces, whether they are of a physical, chemical,
biological, or digital nature, may serve to suggest
or support plausible scenarios of what
might have happened at a possible scene of
crime. Ultimately, they may be instrumental in
distinguishing between the rival scenarios that
the judge or jury will have to consider in arriving
at the final verdict: Was the crime committed by
the suspect in the way described by the prosecution, was it someone else who did it, or was no
crime committed in the first place?
3513
P
Individualization of physical traces, like
DNA, handwriting, or fingerprints, logically
amounts to a process of inference of identity of
source between a crime scene trace and some
reference material whose origin is known.
While its aim is to uniquely identify the source
of the trace and many criminalists, most notably
the dactyloscopists, have long proceeded as
though individualization of traces is a routine
affair, there is no scientific basis for this claim
or for the methods by which individualization or
source attribution is supposedly achieved. It is
largely as a result of the advent of forensic DNA
analysis and the conceptual framework associated with the interpretation of DNA evidence
that the forensic community has come to appreciate the “individualization fallacy” (Saks and
Koehler 2008) and the consequent need to adopt
what has come to be known as a logical or
Bayesian approach to the interpretation of
technical and scientific evidence.
A comprehensive model for the assessment
and interpretation of forensic evidence along
Bayesian lines was developed in Britain by Ian
Evett and colleagues at the Forensic Science Service (Cook et al. 1998a, b; AoFSP 2009; Evett
2011). It defines a hierarchy of propositions to be
addressed in casework which is composed of
three levels and extends over two domains. The
first domain, that of the forensic expert, involves
propositions at the level of source and at the level
of activity. Propositions at the level of source
address the question of the origin of the trace,
while propositions at the activity level relate to
the question how and when the trace arose. Propositions formulated at the third level, termed the
level of offense, belong exclusively to the domain
of the trier of fact and relate to the ultimate, legal
issue whether an offense was committed and if
so, whether it was committed by the defendant.
While the Bayesian approach provides an
excellent framework for the evaluation and interpretation of expert evidence in that it helps define
the relevant questions and draws a sharp line
between the domain of the expert and that of the
trier of fact, it is not an intuitively easy approach.
A clear and transparent exposition of the method
used by the expert to determine the weight of the
P
P
3514
scientific evidence is therefore of the essence.
Recent research as well as court decisions suggest
that if this information is lacking judges and
juries will be hard put to assess the evidence at
its true value.
Criminalistic Processes:
Individualization Versus Identification
In order for material traces like DNA to be able to
contribute to criminal investigations in the form
of technical or scientific expert evidence, there
are various questions that need to be addressed.
One question seeks to determine the nature or
substance of the trace material. Are we dealing
with blood, gunshot residue, MDMA, or hair?
The process by which this type of question is
answered is called identification.
A second process, classification, takes the first
process one step further. It is undertaken to determine whether the crime scene trace can be
assigned to a particular group or class of similar
objects or substances. Examples of classification
processes in the forensic sciences are: the classification of blood into human versus animal blood;
of hair into animal versus human hair and,
beyond that, the (sub)classification into body,
head, and pubic hair; the classification of fibers
into man-made versus natural fibers; the classification of firearms and cartridges in terms of their
make and caliber, or that of shoes according to
size, make, model, sole pattern, etc.
A third process seeks to identify the origin or
source of the trace material. Is it possible to relate
the trace or the carrier of the trace to a particular
person or object? This process, the determination
of the unique source of a trace, is technically
called individualization and is often seen as the
process that lies at the heart of the criminalistic
effort. Although, in common parlance, the term
“identification” may also be used for the process
of source determination, as in a phrase like fingerprint identification or indeed in the title of this
entry, the proper technical term for this process is
individualization.
A fourth question concerns the relation of the
trace material with the crime by trying to
Philosophy of Forensic Identification
establish how and when the trace was generated.
This process, which focuses on the nature of the
activity by which the trace material arose, is often
referred to as association.
All four processes, identification, classification, individualization, and association, are
essential elements of the larger process of reconstruction (Inman and Rudin 2000): the attempt to
create coherent scenarios of the chain of events
that constitutes the incident under examination
on the basis of the available evidence.
Particularly in the context of DNA evidence,
both “source level” questions of the type “Whose
DNA is it?” and “activity level” questions of the
type “How and when did the cell material get
there?” are of paramount importance, as it is
only on the strength of the combined answers to
these questions that a link can be established
between a particular person or object and
a possible crime.
Conclusions in Individualization
Criminalistics is the science of individualization
(Kirk 1963, p. 235)
Individualization may be defined as the process
undertaken to establish the unique source or origin
of a particular material trace. Put simply, it tries to
answer questions like “Whose DNA/handwriting/
fingerprint, etc. is this?” According to Paul Kirk
(1902–1970), one of the founding fathers of criminalistics, it is individualization that marks off
criminalistics as a science from other sciences.
While the goal of criminalistics is always the
same, there are considerable differences in the
methodology used and the conclusions formulated
by the practitioners of the various forensic disciplines to achieve this aim. By and large, three
groups may be distinguished in terms of the way
in which the conclusions of a comparative trace
examination are expressed. First, the conclusions
may directly address the probability of the source
hypothesis given the evidence. This tends to be the
case in dactyloscopy and typically also still holds
for most other types of source level trace examinations. An important exception is DNA, where the
conclusions expressed relate to the probability of
the evidence under a particular hypothesis rather
Philosophy of Forensic Identification
3515
than to the probability of the hypothesis given the
evidence. Second, unlike dactyloscopists, who will
generally express their conclusions in categorical
terms, forensic experts in most other fields of trace
examination will tend to express their conclusions
in probabilistic terms, in either a verbal format or in
a quantitative format, as in DNA typing.
Schematically, where C stands for categorical,
P for probabilistic, E for statement of the probability of the evidence, and H for statement of the
probability of the hypothesis, the following distinctions may be noted:
Dactyloscopy (fingerprints)
Other physical traces (e.g.,
handwriting, fibers, shoe prints, tool
marks)
DNA (biological traces)
C
H
P (and verbal) H
P (and
quantitative)
E
For example, the dactyloscopist will conclude
that a finger mark does or does not originate from
a particular (person’s) finger, thereby making
a categorical statement of the probability of H,
the source attribution hypothesis. But the handwriting expert will – traditionally anyway –
typically conclude that the questioned writing
probably/very probably/with a probability bordering on certainty does – or does not, as the
case may be – originate from the writer of the
reference material, thereby making a (verbal)
probabilistic rather than a categorical statement
of H, the source hypothesis.
By contrast, the DNA expert will first determine
whether the source material could originate from the
person whose reference profile is being compared
with that of the crime scene sample. If the profiles of
the crime scene material and the person under investigation match, the expert will state that the crime
scene material may originate from this person. To
indicate the significance of this finding, the expert
will add how likely the evidence, i.e., a matching
profile, is if the crime scene material originates from
the “matching” person under investigation as
opposed to a random member of the population.
Alternatively, the expert may report the estimated
frequency of the profile in the relevant population.
A third way to report the weight of the evidence is in
terms of the so-called random match probability of
P
the profile: The probability that a randomly chosen
member of the population who is not related to the
donor of the cell material or the matching suspect
will match the crime scene profile.
It appears therefore that while the experts in
the diverse fields are answering the same question, that of the origin of the trace, they will tend
to use different conclusion formats to express
their findings. The notions underlying these differences are explored below.
The Classical Approach to Individualization:
The Underlying Principles
The traditional forensic approach to trace individualization is based on four – partly implicit –
principles:
1. The principle of transfer of evidence
2. The principle of the divisibility of matter
3. The uniqueness assumption
4. The individualization principle
The Transfer Principle: “Every Contact
Leaves a Trace”
The first of these, the transfer principle, is captured in the phrase “Every contact leaves a trace.”
It expresses the notion that in the commission of
criminal acts, invariably some traces will be left
on the scene. It provides a theoretical basis for the
generation of traces and a principled argument
for the examination of the crime scene for traces.
This insight was formulated most clearly by the
Frenchman Edmond Locard (1877–1966), who
was the first director of the Laboratoire de Police
Scientifique in Lyon, France, in 1912 and is universally acclaimed as one of the fathers of criminalistics. While Locard was primarily thinking
of the transfer of microscopic traces like dust,
dirt, nail debris, or fibers, as left in the commission of the more violent type of crime, Locard’s
exchange principle, as the principle is also commonly referred to, is also applicable to traces that
arise in the context of less violent crimes as well
as to latent (hidden), patent (visible to the naked
eye), or plastic impression evidence, such as
finger marks, tool marks, footwear marks, or
striation marks on bullets and cartridge cases. In
the latter case, it is not so much the physical
matter that is deposited at or taken away from
P
P
3516
the crime scene that is of interest but the patterns
or shapes that are transferred from donor (object)
to recipient (object). In addition, the principle
increasingly applies to so-called contact traces,
as in DNA evidence. Here, terms like “trace
DNA” or “touch DNA” are used to refer to cell
material and debris transferred through skin
contact, which may arise as a result of regular
use, as on a watch, from a single, firm contact,
as on a tie wrap or strangulation cord, or from
a single touch, as on a glass surface
(Raymond et al. 2004).
As Locard observed at the time, transfer may
go either way. Traces like glass fragments, blood
stains, hairs, or fibers are left at the crime scene
by a donor and may be picked up from there by
a receptor. If (part of) the material collected is
subsequently left by the receptor and picked up
by a second receptor, we speak of secondary
transfer. For example, fibers or DNA material
picked up by A from B’s clothing may be transferred from B’s clothing to chair C, and eventually end up on the clothing of D. In principle,
forms of tertiary and quarternary transfer might
also occur.
The Divisibility of Matter: “Matter
Divides Under Pressure”
The second underlying principle of criminalistics, that of the divisibility of matter, was explicitly defined only relatively recently by the
American DNA experts Keith Inman and
Norah Rudin (2000). It explains why transfer
can play such an important role in the generation
of traces. Although the principle as such is fairly
obvious, it is of considerable importance for
a proper understanding of the relation between
the way traces arise and their interpretation. In
an article entitled “The origin of evidence,”
Inman and Rudin (2002: p. 12) describe the
process of the division of matter and its results
as follows:
Matter divides into smaller component parts when
sufficient force is applied. The component parts
will acquire characteristics created by the process
of division itself and retain physico-chemical properties of the larger piece
Philosophy of Forensic Identification
This mechanism has important implications
for the relation between traces and their sources.
They are:
Corollary 1 Some characteristics retained by the
smaller pieces are unique to the original item or to
the division process. These traits are useful for
individualizing all pieces to the original item.
Corollary 2 Some characteristics retained by
the smaller pieces are common to the original as
well as to other items of similar manufacture. We
rely on these traits to classify them.
Corollary 3 Some characteristics from the original item will be lost or changed during or after the
moment of division and subsequent dispersal; this
confounds the attempt to infer a common source.
(Inman and Rudin 2002, p. 12)
While the two principles discussed so far
relate to the creation or generation of traces, the
following two principles are central to the interpretation of traces, as viewed in the traditional
approach to trace individualization.
The Uniqueness Assumption: “Nature
Never Repeats Itself”
The first of these is the uniqueness assumption. It
is nicely captured in the phrase “Nature never
repeats itself” and essentially simply states that
no two objects are identical. Or, as Kirk and
Grunbaum (1968, p. 289) put it:
Now most students believe that all items of the
universe are in some respect different from other
similar items, so that ultimately it may be possible
to individualize not only a person but any object of
interest. This effort is the heart of criminalistics.
The uniqueness assumption was probably
most vigorously championed by the fingerprint
fraternity. However, before the fingerprint was
discovered as a means to verify a person’s identity and subsequently came to be used for forensic
purposes, the same principle provided a basis for
forensic anthropometry, which was developed by
the Frenchman Alphonse Bertillon (1853–1914).
Anthropometry. Anthropometry was developed by Bertillon primarily to identify repeated
offenders. The method is based on an assumption
derived from the Belgian astronomer and statistician Adolphe Quetelet (1796–1874) that no two
Philosophy of Forensic Identification
human bodies are equal. A founder of modern
quantitative sociology, Quetelet is not only
believed to be the inspiration for the frequently
cited phrase that nature does not repeat itself but
must also be credited with the definition of the
Quetelet index, which, since 1972, has come to be
more widely known as the BMI or Body Mass
Index.
The anthropometric method consisted in
recording the dimensions of an arrestee’s body in
terms of seven and later 12 measurements of
a fixed set of parts of the body, including total
physical height, the length and width of the head,
the right ear, and the left foot, which were believed
to be constant for adult members of the human
race. In 1883, Bertillon succeeded in identifying
a repeated offender by comparing his measurements with anthropometric data recorded earlier.
Later in his life, Bertillon was the first to make
a fingerprint identification in a murder case on the
European Continent (Thorwald 1965, p. 83).
The assumption of uniqueness, together with
the temporal stability of fingerprints, or friction
ridge patterns as they are more technically called,
is often adduced as a theoretical ground for the
justification of the use of categorical conclusions
of origin, as has typically been common if not
universal practice in dactyloscopy. However,
as Saks and Koehler (2005, p. 892) put it, in
formulating these categorical conclusions,
dactyloscopists in fact rely on a flawed notion of
“discernible uniqueness.” The real issue in source
attribution is not whether all possible sources can
be distinguished from each other in principle,
which is what the uniqueness assumption –
presumably correctly – implies. The crucial question is whether a trace, which, due to the factors
described under the divisibility principle, will
inevitably differ to some extent from its particular source, can be attributed – with certainty, or,
failing that, with any reliable degree of probability – to that source.
The Individualization Principle: “That Can’t Be a
Coincidence”
The fourth principle in the traditional approach to
trace individualization states that a conclusion of
3517
P
(probable) common origin of a trace and reference material – as in a comparative examination
of a questioned handwriting sample and
a reference sample from a known person – may
be arrived at if there are so many similarities of
such significance that their occurring together by
chance may be practically excluded. As the
American handwriting expert R.A. Huber
(1959–1960, p. 289) put it in his definition of
what he called “the principle of identification”:
When any two items have characteristics in common of such number and significance as to preclude
their simultaneous occurrence by chance, and there
are no inexplicable differences, then it may be
concluded that they are the same, or from the
same source.
The suggestion here is that a criterion, i.e., “of
such number and significance,” may be defined
which will provide a principled and objective
way to determine that the possibility (or probability) that two objects meet this criterion by
chance can be excluded. Such a criterion is not
only not feasible in practice, as it begs the question what number and what degree of significance
is required, but it also lacks a theoretical basis in
that it ignores the essentially inductive nature of
the individualization process.
The reason why individualization is problematic from a theoretical point of view is that any
attempt to identify the unique source of physical,
biological, or pattern evidence like finger marks,
footwear marks, DNA, or handwriting is typically frustrated by the induction problem. We
cannot, solely on the strength of even an extreme
degree of similarity between trace and reference
material, conclude that a particular trace must
have originated from some specific reference
material to the exclusion of all other possible
sources, unless we have been able to examine
all these alternative sources and eliminate them
categorically.
To begin with, the population from which
a finger mark or a questioned handwritten text
actually originates is typically indefinite in size
and frequently largely unavailable for examination. But even if it were possible to examine
a large number of potential writers or fingers,
P
P
3518
we could not exclude finding one or more whose
reference handwriting or fingerprint would show
a similar or even greater degree of similarity with
the questioned handwriting or finger mark than
did the reference sample of the suspect. Since this
possibility cannot be excluded, the individualization problem tends to be impossible to solve and
in that sense is strongly reminiscent of that of
Popper’s white swans: We cannot conclude that
all swans are white unless we have been able to
examine all swans (Popper 1959). Nor for that
matter can we conclude that the trace probably
originates from the reference material with which
it shares many features. Indeed, although the
observed degree of similarity between trace and
possible source will tend to make the hypothesis
of identity of source more probable in relative
terms than it was before the comparative examination was carried out, similarity is neither
a sufficient nor even a necessary condition for
identity of source in absolute terms.
It is interesting that the individualization criterion as captured in the phrase “That cannot be
a coincidence” is essentially similar to that used
in traditional, standard textbook statistical significance testing. In this approach, the result of
a hypothesis test is termed significant if it is
unlikely to have occurred by chance. More specifically, the so-called null hypothesis that the result
is due to chance may be rejected if the obtained
result is less likely to occur under this hypothesis
than a predetermined threshold probability of –
frequently – 1 or 5 %. Like the traditional identification paradigm, this approach is also coming in
for more and more criticism, partly because it also
fails to take account of the probability of the result
under the alternative hypothesis.
Class Characteristics Versus Individual
Characteristics
In an attempt to overcome the induction problem
traditional forensic identification, experts frequently rely on the distinction between class
characteristics and individual characteristics.
For example, all firearms of a particular make
and caliber may leave the same markers on
a cartridge or bullet, thereby making it possible
to identify the type of weapon on the basis of the
Philosophy of Forensic Identification
class or system characteristics that the particular
type of firearm is known to leave. However,
a certain configuration of striation marks left on
a bullet may be distinctive for a particular
weapon. As Thornton and Peterson (2002) put it:
Class characteristics are general characteristics that
separate a group of objects from a universe of
diverse objects. In a comparison process, class
characteristics serve the very useful purpose of
screening a large number of items by eliminating
from consideration those items that do not share the
characteristics common to all the members of that
group. Class characteristics do not, and cannot
establish uniqueness.
Individual characteristics, on the other hand,
are those exceptional characteristics that may
establish the uniqueness of an object. It should be
recognized that an individual characteristic, taken
in isolation, might not in itself be unique. The
uniqueness of an object may be established by an
ensemble of individual characteristics. A scratch
on the surface of a bullet, for example, is not
a unique event; it is the arrangement of the
scratches on the bullet that mark it as unique.
Unfortunately, the definition of individual
characteristics is circular. They are defined as
characteristics that are – collectively – capable
of establishing uniqueness. But whether an
ensemble of individual characteristics is unique
is itself an inductive question: We can never be
sure that a feature or combination of features is
unique, until we have observed all relevant
objects, which is impossible.
What practitioners of traditional forensic identification sciences really do is perhaps best
described by Stoney (1991), who used the
image of the “leap of faith” as the mechanism
whereby the forensic scientist actually establishes individualization, as in dactyloscopy:
When more and more corresponding features are
found between the two patterns scientist and lay
person alike become subjectively certain that the
patterns could not possibly be duplicated by
chance. What has happened here is somewhat analogous to a leap of faith. It is a jump, an extrapolation, based on the observation of highly variable
traits among a few characteristics, and then considering the case of many characteristics. [] In fingerprint work, we become subjectively convinced of
identity; we do not prove it.
Ultimately then, in traditional identification
disciplines, in reaching a conclusion about the
Philosophy of Forensic Identification
probable or categorical origin of the trace material, the expert delivers an essentially subjective
opinion. When informed by adequate levels of
training, experience, and expertise, this conclusion will frequently be correct, but it must be
clear that the expert becomes convinced of the
(probable) origin of the trace. He does not
“prove” it: There is no logical basis for the
conclusion.
The Logical Approach to
Individualization: The Concept
of the Likelihood Ratio
By contrast, in the logical approach, the expert
does not primarily seek to determine the probability of the source or activity hypothesis. Instead,
the purpose of the comparative examination is to
determine the likelihood of the evidence under
two mutually exclusive propositions or hypotheses. Suppose we find a size 14 shoeprint at a crime
scene and a suspect emerges who takes size 14. It
will be clear that the shoe size information by
itself gives us insufficient basis to say that it was
the suspect who left the print rather than one of
the other shoe size 14 wearers in the area (or
beyond). The mere fact that the suspect wears
size 14 shoes does not make him more suspect
than anybody else with this size shoes. At the
same time, it is clear that the finding looks
incriminating. Or is it a mere coincidence?
The likelihood ratio (LR) provides
a principled way to address this question. To
calculate it, we need to determine the ratio of
the likelihood of a size 14 turning up at the
crime scene under two rival hypotheses: (1) The
shoe mark was left by the suspect versus (2) the
mark was left by a random member of the relevant potential donor population. We know that
the likelihood of finding a size 14 mark if the
suspect left it (and assuming he does not occasionally or otherwise wear a different size shoe)
is 1, or 100 %. To assess the likelihood of finding
a size 14 shoe mark under the alternative hypothesis that a random member of the relevant population left the print, we need to know what
percentage of that population takes a size 14
3519
P
shoe. Suppose we know this figure to be 4 %.
We can now calculate the likelihood ratio of the
evidence under these two competing hypotheses,
which in this case would amount to 100/4 ¼ 25.
We can paraphrase this result by saying that
the footwear evidence is 25 times more likely if
the suspect left the mark than if a random member
of the population left it. We can also say that the
evidence makes it (25 times) more likely that the
suspect left the mark than we believed was the
case before we obtained the evidence. But we
cannot on the basis of the shoe mark evidence
alone pronounce upon the probability of the
hypothesis that the suspect left the mark in absolute terms. Alternatively, what we can say in
a case like this is that an LR of 25 reduces the
group of potential suspects by a factor 25, leaving
just one potential suspect on average for every 25
potential suspects considered before the shoe
print evidence became available.
Similarly, when applied to DNA evidence, the
likelihood ratio is a measure of the weight of the
evidence and may be seen as an indication of the
extent to which the uncertainty about the source
hypothesis is reduced by the evidence. If, for
example, the matching profile is known to have
a frequency of 1 in a million in the relevant
population, the LR of the matching DNA evidence may be reported as one million: It is the
ratio of the likelihood of the evidence under the
hypothesis that the person under investigation is
the donor of the DNA material, i.e., 1 (or 100 %)
divided by the likelihood of the evidence under
the alternative hypothesis that the DNA originates from a random member of the population,
i.e., 1/1,000,000. The conclusion implies that the
evidence is a million times more likely if the
crime scene sample originates from the donor of
the reference material than if a random member
of the population were the donor of the crime
scene sample. It is only if there is no match that
the DNA expert may make a – negative – categorical statement of the probability of the
hypothesis.
Discrete Versus Continuous Variables
In the examples involving DNA and shoe sizes
discussed above, the variables of interest are
P
P
3520
discrete or categorical in nature and the similarity
between trace and reference material could be
said to be either complete or to be entirely absent:
The DNA profiles and the shoe sizes are the same
or they are different. If they are different, the
suspect’s shoe or body can be eliminated as the
source of the trace. However, in many cases, the
correspondence between trace and source will not
be perfect. Many variables are not discrete (like
DNA markers or shoe size) but continuous.
Examples are quantitative variables such as
length, weight, or shape. For these variables,
there will always be a difference between trace
and reference materials because in both cases, we
are dealing with samples that will at best only
approximate the “true,” i.e., average, population
value for the feature of interest.
Again, handwriting analysis may serve as
a case in point, but the same holds for any other
type of trace with continuous properties like tool
marks, fibers, speech, glass, or paint. Both the
questioned writing and the reference writing
must be seen as samples taken from the indefinitely large population of handwriting productions that the writers of these samples are
capable of producing. This means that any
marker that is examined, like the shape and
execution of the letter t or the figure 8, will
exhibit a certain degree of variability even within
samples originating from the same source.
However, traces originating from a single source
will typically show a relatively small degree of
within-source variation, while samples of material originating from different sources will
typically show a relatively larger degree of
between-source variation. As a result, in these
cases, the numerator of the likelihood ratio will
not be 100 % or 1, but say 80: A degree of
similarity as great as that found between trace
material and reference material will then be
found in 80 % of cases if both originate from
the same source and in say only 10 % of cases if
trace material and reference material originate
from different sources. The likelihood ratio
would then be 80/10 ¼ 8.
If, however, the degree of similarity found
would be expected in only 20 % of cases under
the hypothesis of common origin and in 40 % of
Philosophy of Forensic Identification
cases under the hypothesis of different origin, the
likelihood ratio would be 20/40 ¼ ½, and the
evidence would actually weaken the common
source hypothesis rather than strengthen it. In
the first instance, the likelihood ratio would
exceed 1 and the evidence would (weakly) support the hypothesis of common origin; in the
second instance, the likelihood ratio would be
less than 1 and the evidence would – again
weakly – support the alternative hypothesis, that
the trace did not originate from the same source
as the reference material.
The Likelihood Ratio and
the Diagnostic Value
The concept of the likelihood ratio is similar to
that of the diagnostic value, a measure which has
found wide acceptance in fields like medicine and
psychology as a way to express the value of
a diagnostic test result. It is arrived at by dividing
the relative number of correct (positive or negative) results of the test in question by the relative
number of false (positive or negative) results of
the test. In more technical language, by dividing
the sensitivity of the test by (1 – the specificity).
The sensitivity of the test is the percentage of
correct positives it produces, the specificity the
percentage of correct negatives. Applied to an
HIV test, if the sensitivity of the test is 98 %
and the specificity is 93 %, this would mean that
98 % of those infected would (correctly) test
positive and 7 % of those not-infected would
(incorrectly) also test positive. The diagnostic
value of a positive result for the test would then
be 98/(100 93) ¼ 14 and the probability of the
patient being infected would be 14 times greater
now that the test result is known than whatever it
was estimated to be before the test result was
known. The diagnostic value of a negative
test result would in this case be 93 (the relative
number of correct negatives) divided by
(100 98) ¼ 2, the relative number of false
negatives, i.e., 93 divided by 2 or 46.5.
Scientific evidence may be said to have diagnostic value in much the same way as a medical
test such as an HIV test, which will not provide
absolute proof of infection or otherwise but,
depending on the result and its diagnostic value,
Philosophy of Forensic Identification
will make infection more or less probable. Similarly, evidence may be more or less likely under
one of two rival hypotheses. To the extent that the
evidence favors, or better fits one hypothesis
rather than another, it may be said to lend more
support to that hypothesis.
Bayes and the Prior Probability
The likelihood ratio may be seen as a measure of
the extent to which the hypothesis of interest is
more probable or less probable after the scientific
evidence is known than it was before the evidence was known. Another way of putting it is
to say that we can update what we saw as the prior
probability of the hypothesis with the evidence
that has become available to arrive at the posterior probability of the hypothesis, in which the
weight of the new evidence has been taken into
account. By means of the Bayes’ Rule, the odds
form of Bayes’ theorem, so called after the
Rev. Thomas Bayes (1702–1761), we can calculate the posterior probability of a hypothesis by
multiplying its prior probability with the likelihood ratio:
Prior probability Likelihood ratio
¼ Posterior probability
A simple example may illustrate the application of the rule. Suppose a prisoner is found dead
and we may safely assume that one of the 99
remaining fellow prisoners in the ward is the
perpetrator. As it happens, one of these 99 prisoners, S, admits to being the killer. His DNA
profile is obtained and found to match with
a partial profile obtained from the nail debris
secured from the victim. The estimated frequency
of the partial profile in the relevant population is
1 in 1,000.
If the cell material in the nail debris originates
from S, the likelihood of the evidence, i.e., the
matching profile, would be 1, or 100 %. If, on
the other hand, the cell material belonged
to a random member of the population, the
likelihood of a match would be 1 in 1,000 or
0.001. The likelihood ratio of the DNA evidence
would therefore be 1 divided by 0.001 ¼ 1,000.
3521
P
To determine the prior probability, we assume
that all 99 remaining prisoners are equally likely
to be the perpetrator. In that case, the prior probability of any one of them being the donor is 1/99,
or, expressed in odds, 1 to 98. We can now
calculate the prior probability of S being the
donor of the cell material in the nail debris by
applying Bayes’ rule:
Prior probability LR ¼ Posterior probability
ð1 to 98Þ 1; 000 ¼ 1; 000 to 98:
The odds form of the posterior probability
may be converted to a fraction, i.e., 1,000/1,098,
and from that into a percentage, i.e., 91.1 %. This
means that on the basis of the DNA evidence
alone, the probability of the suspect being the
donor of the cell material is 91.1 %, assuming
that the donor is one of the remaining 99
prisoners.
Of course, apart from the DNA evidence, the
fact that S admitted to killing his fellow inmate is
also relevant to the determination of the ultimate
issue whether S is the perpetrator. However, this
type of evidence as well as possible eyewitness
accounts from fellow prisoners or prison staff,
and any other non-DNA evidence, clearly extends
beyond the domain of the DNA expert. Other
experts might be able to assign a particular weight
to a spontaneous confession or an eyewitness
identification, in the form of a likelihood ratio.
For example, empirical research may be available
on which an estimate may be based of the diagnostic value (¼ likelihood ratio) of a confession.
This could take the form of a statement that
a confession is on average seven times more
likely if the suspect is the perpetrator than if he
is not. This evidence in turn could then be used to
further update the hypothesis of guilt.
When Numbers Are Lacking
It is worth noting that the concept of the likelihood ratio may also be applied to evidence types
where the frequency of relevant markers cannot
be estimated in numerical terms, for example,
because there are no suitable reference databases.
P
P
3522
This would in fact hold for many types of trace
evidence, including handwriting, glass, paint,
fibers, toolmarks, and firearms. The logical
framework may then be applied to the verbal
statements that are used in these fields. In
response to the logical objections raised to the
use of traditional probability scales in forensic
identification, various proposals have been
made in recent years for the introduction of logically correct verbal probability scales. One such
format, developed at the Netherlands Forensic
Institute (NFI) and used primarily in the various
forensic identification disciplines, looks as
follows:
The findings of the comparative examination are
equally likely/
more likely/
much more likely/
very much more likely
under the prosecution hypothesis that the suspect is
the source of trace material as/than under the
defence hypothesis that a random member of the
population is the source of the trace material
Note that, true to the logical format, it is the
likelihood of the evidence that is addressed.
While the verbal phrases used clearly have
a probabilistic basis, the probabilities are not
based on quantitative empirical data but informed
by the analyst’s experience and may be seen as
internalized frequency estimates.
In those exceptional cases where the forensic
scientist arrives at a subjective conviction that the
trace material originates from a particular source
(as in physical fits of torn paper, or qualitatively
superior shoe prints or tool marks), he or she
might express his or her subjective conviction,
emphasizing that this is precisely that –
a subjective conviction, not a scientific fact.
Words Versus Numbers
In the United Kingdom as well as in some
Continental European countries, conclusions in
forensic identification are increasingly expressed
in a slightly different logical format. According
to the “standards for the formulation of evaluative forensic science expert opinion,” compiled
by the Association of Forensic Science Providers
(AoFSP 2009), “the evidential weight (. . .) is the
Philosophy of Forensic Identification
expression of the extent to which the observations
support one of the two competing propositions.
The extent of the support is expressed to the client
in terms of a numerical value of the likelihood ratio
(where sufficiently robust data is available) or
a verbal scale related to the magnitude of the
likelihood ratio when it is not.” (AoFSP 2009:
p. 63) For this purpose, the following scale is used:
Value of likelihood ratio
>110
10100
1001,000
100010,000
10,0001,000,000
>1,000,000
Verbal equivalent
Weak support for proposition
Moderate support
Moderately strong support
Strong support
Very strong
Extremely strong
In the case of the DNA evidence with
a likelihood ratio of 1,000, the expert would
either report the numerical value of the likelihood
ratio as such or would report that the DNA evidence provides “moderately strong” to “strong”
support to hypothesis 1 (the suspect is the donor).
The decision to use the numerical or the verbal
form would depend on the extent to which the
expert considers the data underlying the calculation of the likelihood ratio to be robust. This
sounds like a sensible criterion. However, there
are those who advocate the blanket use of verbal
terms, i.e., even if perfectly valid qualitative
empirical data are available, arguing that the use
of a uniform reporting format is vastly preferable
and adding that quantitative data are generally
too complex for nonscientists to grasp.
The Prosecutor’s Fallacy
Regardless of the use of words or numbers, statements of the probability of the evidence under
a particular hypothesis, as made in the Bayesian
approach, are often prone to misunderstanding
and may strike the recipient as counterintuitive.
The finding that the similarities observed
between the handwriting of the writer of an anonymous letter and the reference material produced
by the suspect are much more likely if the suspect
wrote the questioned sample than if it was written
by a random member of the population is often
taken to mean the converse: that it is much more
Philosophy of Forensic Identification
likely that the suspect wrote the questioned
handwriting sample than that it was written by
a random member of the population.
However, if we do this, we are guilty of
making a fundamental logical error which, in
the judicial context, has come to be referred to
as the prosecutor’s fallacy (Thompson and
Schumann 1987). Although the term would
seem to suggest that prosecutors are particularly
prone to this fallacy, it is in fact an example of
a more general type of error that is often made in
the context of probability statements or inverse
reasoning, where it is known as the “fallacy of the
transposed conditional.” In its simplest form, it is
easily spotted: if an animal is a cow, it is very
likely to have four legs. However, the converse
clearly does not hold: if an animal has four legs, it
is not very likely to be a cow.
Transposed conditionals or prosecutor’s fallacies are very frequently encountered in the
context of DNA evidence. Suppose a partial profile is obtained from a crime scene sample whose
frequency in the relevant population is estimated
to be smaller than 1 in 100,000. If the expert
subsequently reports the probability that
a random member of the population has the
same profile as the crime scene sample as
being smaller than, say, 1 in a 100,000, this
statement is frequently understood to mean that
the probability that the DNA does not originate
from the suspect is smaller than 1 in a 100,000.
However, the former is clearly a statement of the
probability of the evidence (i.e., a match with
the partial profile), while the latter is a statement
of the posterior probability of the source hypothesis. Mathematically, the latter statement would
be correct in a situation where the prior probability was set at 50 %, or 1 to 1. This would be
the case if, in addition to the matching suspect,
only one person would equally qualify as
a possible donor, which will not frequently be
a reasonable assumption to make. If, for the sake
of the argument, the prior probability were set at
200,000 (e.g., the size of the adult male population of a large town), application of the odds
form of the Bayes’ rule would yield a posterior
probability of 33.3 %: (1 to 200,000) 100,000 ¼ 1 to 2, or 33.3 %.
3523
P
Transposed conditionals may also occur in the
context of cause and effect arguments. While it is
correct to say that the street will be wet if it has
been raining, the converse is clearly not necessarily true. The single observation that the street
is wet does not allow us to infer that it must have
been raining, or even that it has probably been
raining. Alternative explanations are possible:
The street may have got wet when the police
used water cannon to break up a demonstration,
or it is wet because somebody has just been
washing his car. So we can make a statement
about the likelihood of a particular finding (e.g.,
a wet street) under a particular hypothesis (“it has
been raining”) but not about the probability of
this same hypothesis merely on the basis of the
finding that the street is wet. To determine the
posterior probability that it has been raining, we
need to combine the evidence of the wet street
with the prior probability of rain. In England, the
prior would be high, but in a country like Dubai, it
would presumably be very low. The same evidence combined with vastly different priors may
lead to very different posterior probabilities.
Controversy
The Bayesian approach is not uncontroversial. Its
opponents frequently view its advocates as
“believers” (Risinger 2012). More specifically,
critics of the Bayesian approach object to the
subjective nature of the prior probability, as well
as to the use of likelihood ratios which lack an
empirical basis. While the former criticism is
clearly valid, it might be argued that the explicit
consideration of the prior probability and the formulation of alternative hypotheses is a virtue in
that it helps identify the relevant questions and
prevents tunnel vision. The conclusion format
propagated by the predominantly British Association of Forensic Science Providers is not
unproblematic either, in that a phrase, like “the
examination provides strong support for the proposition that X originates from Y,” will almost
invariably be interpreted to mean that it is very
likely that X originates from Y. Without due warning, logically correct conclusions of this type will
P
P
3524
almost inevitably tend to be mistaken for the logically flawed ones they are meant to replace.
A further problem is highlighted by a decision
of the English Court of Appeal in R v T (2010). In
it, the judges express sharp criticism of the report
of a footwear expert, who, quite in line with the
policy of the (now defunct) UK Forensic Science
Service, had phrased his conclusion in verbal
terms without making it clear to the judge or the
jury that the verbal conclusion was based on
a quantitative estimate of the frequency of the
characteristics size, pattern, and wear as
exhibited by the shoe mark found at the crime
scene. The expert’s conclusion was formulated
as: “(. . .) there is at this stage a moderate degree
of scientific evidence to support the view that [the
Nike trainers recovered from the appellant] had
made the footwear marks.” However, as later
appeared from case notes that were not added to
the original report, the expert had calculated
a likelihood ratio of 100, which number he had
subsequently converted into the verbal phrase
“moderate support,” in accordance with the
scale presented above. In addition, counsel for
the appellant pointed out to the Court of Appeal
that in his testimony in court in response to questions by the defense, the expert had mentioned
estimates of the relevant characteristics of the
shoe mark which, when combined, would lead
to a likelihood ratio of no less than 26,400, in
which case the footwear evidence would have
been much more incriminating (in verbal terms
expressed as “very strong support”). The Court of
Appeal’s judgment is perfectly clear:
The process by which the evidence was adduced
lacked transparency. (. . .) it is simply wrong in
principle for an expert to fail to set out the way in
which he has reached his conclusion in his report.
The Court of Appeal continues:
(. . .) the practice of using a Bayesian approach and
likelihood ratios to formulate opinions placed
before a jury without that process being disclosed
and debated in court is contrary to principles of
open justice.
The ruling in R v T has led to various reactions,
many (Berger et al. 2011; Evett 2011; Redmayne
et al. 2011) but not all (Risinger 2012) in defense
Philosophy of Forensic Identification
of the Bayesian approach. While one may disagree
with some of the views expressed in the Court of
Appeal judgment, recent research conducted in the
Netherlands suggests that there is at least one other
problem with the use of these “logically correct”
conclusion formats. It appears that evaluative
opinions expressed in the Bayesian format are
likely to be misunderstood not only by defense
lawyers and judges but also by forensic experts
themselves. Participants in the study were asked to
indicate for a variety of statements whether they
were correct paraphrases of the Bayesian style
conclusions that were used in a fictitious report.
The study shows that a proper understanding of
statements involving likelihood ratios by jurists is
alarmingly poor. In order to be able to compare
actual versus supposed understanding, participants
were also asked to indicate how well they understood the Bayesian style conclusions of the reports
on a scale from 1 (“I do not understand it at all.”)
to 7 (“I understand it perfectly.”). The most worrying finding to emerge from the study is no doubt
that not only did judges, defense lawyers, and
forensic experts alike tend to interpret the conclusions of the submitted reports incorrectly but they
combined their lack of understanding with a high
degree of overestimation: They believed they
understood the conclusions much better than in
fact they did. This suggests that the continued
use of Bayesian style conclusion formats or likelihood ratios requires a major educational effort if
structural miscommunication between experts and
triers of fact is to be avoided. Aitken et al. (2010)
and Puch-Solis (2012) may be seen as attempts to
meet this demand.
Conclusion
The findings of a comparative examination
undertaken with a view to establishing the source
of a particular trace or set of traces do not strictly
allow the type of probabilistic source attributions,
be they of a quantitative or verbal nature, that
until recently were used by the vast majority of
forensic practitioners. With the advent of forensic
DNA analysis over the last decades and the widespread use of the conceptual framework
Philosophy of Forensic Identification
associated with the interpretation of DNA evidence, awareness among forensic practitioners
of other identification disciplines of the inadequacies of the traditional evidence evaluation paradigm has grown rapidly. Increasingly, this is
leading to attempts to apply a logically correct
way to express the value of the findings of
a source attribution examination of trace material
other than DNA by expressing the weight of the
trace evidence in a way similar to that used in
forensic DNA analysis. This takes the form of
a so-called likelihood ratio. The concept is similar to that of the diagnostic value, a measure
which has found wide acceptance in fields like
medicine and psychology as a way to express the
value of any diagnostic test result.
The concept of the likelihood ratio requires the
consideration of the probability of the evidence
under two competing hypotheses, one based on
a proposition formulated by the police or the
prosecution, the other based on an alternative
proposition which may be based on a scenario
put forward by the defense. As such, the Bayesian
approach may be seen as a remedy against
suspect-driven investigations, in which the police
tend to focus on collecting evidence that will
confirm the suspect’s involvement in the crime
and ignores alternative explanations. By contrast,
in a crime-driven investigation, the investigators
base the direction of the investigation on the clues
provided by the crime rather than by the person of
the suspect and develop one or more scenarios
based on the evidence rather than make the evidence fit a particular scenario.
Technical or scientific evidence derived from
material traces such as DNA, finger marks, handwriting, fibers, footwear marks, or digital data
derived from a harddisk or a mobile telephone
may be incompatible with a particular hypothesis
that is central to a larger scenario and then effectively eliminate that scenario. More frequently,
scientific evidence may be more or less likely to
be found in one scenario than another and in this
way may help discriminate between various scenarios. It is the consideration of the totality of
evidence, both direct, witness and scientific evidence if available, which forms the basis for the
ultimate decision made by the trier of fact.
3525
P
Related Entries
▶ Automated and Manual Forensic Examinations
▶ Bayesian Updating and Crime
▶ Cognitive Forensics: Human Cognition,
Contextual Information, and Bias
▶ Crime Science
▶ Criminal Investigative Analysis
▶ DNA Profiling
▶ DNA Technology and Police Investigations
▶ Fingerprint Identification
▶ Forensic Science in the Nineteenth and
Twentieth Centuries
▶ Identification and the Development of Forensic
Science
▶ Identification Technologies in
Policing and Proof
▶ Scientific Basis of the Forensic Process
▶ Scientific Evidence in Criminal Prosecutions
Recommended Reading and References
Aitken C, Roberts P, Jackson G (2010) Communicating
and interpreting statistical evidence in the administration of criminal justice: fundamentals of probability
and statistical evidence. In criminal proceedings –
Guidance for judges, lawyers, forensic scientists and
expert witnesses, Royal Statistical Society, London
Association of Forensic Science Providers (2009) Standards for the formulation of evaluative forensic science expert opinion. Sci Justice 49:161–164
Berger CEH, Buckleton J, Champod C, Evett IW, Jackson G
(2011) Evidence evaluation: a response to the court of
appeal judgment in R v T. Sci Justice 51:43–49
Broeders APA (2006) Of earprints, fingerprints, scent
dogs, cot deaths and cognitive contamination: a brief
look at the present state of play in the forensic arena.
Forensic Sci Int 159(2–3):148–157
Cook R, Evett IW, Jackson G, Jones PJ, Lambert JA
(1998a) A model of case assessment and interpretation. Sci Justice 38(3):151–156
Cook R, Evett IW, Jackson G, Jones PJ, Lambert JA
(1998b) A hierarchy of propositions: deciding which
level to address in casework. Sci Justice
38(4):231–239
de Keijser J, Elffers H (2010) Understanding of forensic
expert reports by judges, defense lawyers and forensic
professionals. Psychol Crime Law 18(2):191–207
Evett IW (2011) Expressing evaluative opinions:
a position statement. Sci Justice 51:1–2
Inman K, Rudin R (2000) Principles and practice of criminalistics: the profession of forensic science. CRC,
Boca Raton
P
P
3526
Inman K, Rudin R (2002) The origin of evidence. Forensic
Sci Int 126:11–16
Kirk PL (1963) The ontogeny of criminalistics. J Crim
Law Criminol Police Sci 54:235–238
Kirk PL, Grunbaum BW (1968) Individuality of blood and
its forensic significance. Leg Med Annu 289–325
Popper KR (1959) The logic of scientific discovery.
Hutchinson, London
Puch-Solis R, Roberts P, Pope S, Aitken C (2012) Communicating and interpreting statistical evidence in the
administration of criminal justice: 2. Assessing the
probative value of DNA evidence. Royal Statistical
Society, London
Quetelet A (1870) Anthropométrie ou Mesure des
Différentes Facultés de l’Homme. Muquardt,
Bruxelles
R v T (2010) EWCA Crim 2439
Raymond JJ, Walsh SJ, Van Oorschot RA, Roux C (2004)
Trace DNA: an under-utilized resource or Pandora’s
box? – A review of the use of trace DNA analysis in the
investigation of volume crime. J Forensic Identif
54(6):668–686
Redmayne M, Roberts P, Aitken CGG, Jackson G (2011)
Forensic science evidence in question. Crim Law Rev
5:347–356
Risinger DM (2012) Reservations about likelihood ratios
(and some other aspects of forensic ‘Bayesianism’).
Law Probab Risk 0:1–11
Saks MJ, Koehler JJ (2005) The coming paradigm
shift in forensic identification science. Science
309:892–895
Saks MJ, Koehler JJ (2008) The individualization fallacy
in forensic science evidence. Vanderbilt Law Rev
61:199–219
Stoney DA (1991) What made us ever think we could
individualize using statistics? J Forensic Sci Soc
31(2):197–199
Thompson WC, Schumann EL (1987) Interpretation of
statistical evidence in criminal trials: the prosecutor’s
fallacy and the defence attorney’s fallacy. Law Hum
Behav 11:167–187
Thornton JI, Peterson JL (2002) ‘The general
assumptions and rationale of forensic identification’ }
24–2.0–2.1. In: Faigman DL, Kaye DH, Saks MJ,
Sanders J (eds) Modern scientific evidence: the
law and science of expert testimony. West Publishing
Co, St. Paul
Thorwald JT (1965) The century of the detective.
Harcourt, Brace & World, New York, first published
in German as Das Jahrhundert der Detektive,
Droemersche Verlagsanstalt: Z€
urich 1964
Physical Punishment
Place-Based Randomized Trials
Robert Boruch1, David Weisburd2,3,
Richard A. Berk4 and Breanne Cave5
1
Graduate School of Education, University of
Pennsylvania, Philadelphia, PA, USA
2
Department of Criminology, Law and Society,
George Mason University, Fairfax, VA, USA
3
Faculty of Law, The Hebrew University,
Mt. Scopus, Jerusalem, Israel
4
Department of Criminology, University of
Pennsylvania, Philadelphia, PA, USA
5
George Mason University, Fairfax, VA, USA
Overview
Place-randomized trials are an important vehicle
for generating evidence about “what works” in
criminology. Place-based randomized trials are
a form of cluster randomization that involves
identifying a sample of places (for instance,
crime hot spots) and randomly allocating these
locations to different police or community interventions. Random allocation assures a fair comparison among the interventions, and when the
analysis is correct, a legitimate statistical statement of confidence in the resulting estimates of
their effectiveness can be made.
This entry provides basic definitions and practical counsel about the use of such trials in generating evidence in crime prevention. It also
identifies issues, ideas, and challenges that might
be addressed by future research. Finally, it discusses the special analytic difficulties that may
occur in the development of place-randomized
trials as opposed to more traditional trials where
individuals are the units of allocation and analysis.
Places that Are Randomized: Theory and
Units of Randomization
Physical Punishment
▶ History of Corporal Punishment
Place-randomized trials depend on a clear understanding of the role of place. A place can be an
entity in itself; for instance, a business establishment that has a legal status separates from that of
Place-Based Randomized Trials
owners or employees. Or a place can be an organizational convenience for smaller units within it.
One may, for instance, become interested in the
effect of neighborhood context on the likelihood
of individual victimization. Inasmuch as individual habits can also contribute to risk of victimization, attention may also be directed toward both
the attributes of neighborhoods as well as to the
individuals who live within the neighborhood.
Whether the place is an entity per se or a
receptacle for lower-level observational units
can dramatically affect how the randomized trial
is designed and analyzed. The key definitional
point in a “place-randomized trial,” however, is
that the random assignment occurs at the place
level. The implication is that the methodological
benefits of random assignment are realized across
places, not the units within them. Statistical
analysis at the place level therefore conforms
to well-understood and accepted statistical
practices. Analysis attempted at the level of the
nonrandomized units which are nested within
places can be complex and controversial.
The “units of randomization” in a placerandomized trial may vary considerably.
Weisburd and Green’s (1995) study, for instance,
operationalized “drug hot spots” as street segments rather than institutions such as housing
developments, schools, or business units. The
broad theory underlying the trial posited that
focusing police and other resources on these hot
spots would reduce crime, rather than leading to
no effect or a migration of criminal activity into
other areas. An earlier crime hot spot trial by
Sherman and Weisburd (1995) in Minneapolis
similarly defined the unit of analysis as a single
street segment from intersection to intersection.
Other places have been targeted for different
types of interventions, such as saloons in the
context of preventing violence and glasswarerelated injuries (Warburton and Sheppard 2000).
Private properties, including apartment houses
and businesses, have been targeted in a study
of the effects of civil remedies and drug
control (Mazerolle et al. 2000). Convenience
stores, crack houses, and other entities have
also been randomly allocated to different
interventions.
3527
P
These examples invite a number of basic questions. How can such trials be deployed well? Can
better theory be developed about what should
happen as a consequence of an intervention
at high levels of units: province or county, city
or village, institution or housing development, or
crime hot spot? Can theories be developed to
guide thinking about change or rate of change at
the primary aggregate level – the places – and
below it? What new statistical problems emerge
from randomization of places? These examples
also invite questions about how to learn about
other trials of this sort, involving yet other units
of allocation and analysis.
Relationships and Agreements
People get place-randomized trials off the ground
through agreements between the trialist’s team
and prospective partners in the place-based trial.
Here, “partners” mean individuals or groups
whose cooperation and experience are essential
in deploying both the intervention and the trial.
Weisburd (2005) emphasizes the need to develop
personal relationships that lead to trust and willingness to experiment on innovations that might
work better than conventional practice. In his
Jersey City experiment, for instance, the strong
involvement of the senior police commander as
principal investigator in the study played
a critical role in preventing a breakdown of the
experiment after 9 months.
In Jersey City, the Deputy Chief who administered the interventions was strongly convinced
of the failures of traditional approaches and the
need to test new ones. The commander took personal authority over the narcotic unit and used his
command powers to carefully monitor the daily
activities of detectives in the trial. This style of
work suggests the importance of integrating
“clinical” and research work in criminal justice,
much as they are integrated in medical experiments. It also reinforces the importance of practitioner “belief” in the necessity of implementing
a randomized study. The Kingswood experiment
described by Clarke and Cornish (1972)
illustrates how doubts regarding the application
P
P
3528
of experimental treatment led practitioners to
undermine the implementation of the study.
Place-randomized trials do not only involve
the practitioners and researchers who are directly
involved with implementation and analysis. For
instance, in policing, federal funding agencies,
such as the National Institute of Justice and the
Bureau of Justice Assistance; accreditation
agencies; and professional organizations such as
the International Association of Chiefs of Police,
the Police Executive Research Forum, or the
Major Cities Chiefs Police Association can all
influence the willingness of individual police
departments to adopt place-based practices
and implement place-based trials. The capacity
of researchers and practitioners to carry out
place-randomized trials is therefore often
influenced by the preferences and priorities of
these institutional stakeholders.
Developing relationships in place-randomized
trials as in many other kinds of field research
depends on reputation and trust, of course. The
topic invites attention to questions for the future.
How can better contracts and agreements in networks of organizations, public and private, ones
that permit us to generate better evidence about
the effects of an innovation, be developed?
How can “model” contracts and memorandums
of understanding be developed and made available to other trialists and their potential
collaborators?
Justifications for a Place-Randomized
Trial
For many social scientists, an important condition for mounting a randomized trial on any intervention that is purported to work is that (a) the
effectiveness of a conventional practice, policy,
or program is debatable and (b) the debates can be
informed by better scientific evidence. In the
crime sector, police of course are local theorists,
and they often disagree about what could work
better. Crime experts have also disagreed about
what approaches might be effective in high-crime
areas. More generally, of course, people disagree
with one another about what might work in the
Place-Based Randomized Trials
policy sector, and there is, at times, some agreement that better evidence would be helpful.
For instance, Weisburd (2003) points out that
one of the major justifications for random trials is
disagreement among experts about the effectiveness of an intervention. This is an important
factor in justifying a randomized trial using individuals or places as the units of random allocation. For instance, the Cambridge-Somerville
Youth Study, one of the most famous experiments in youth crime prevention, illustrated that
even well-planned and implemented intervention
programs may have no effect on offending
behavior or even have a “backfire” effect
(McCord 2003). Experimental studies of the
effectiveness of random beat patrol and foot
patrol likewise provided surprising results
regarding the effectiveness of widely accepted
standard practices in criminal justice in reducing
crime (Kelling et al. 1974). In these cases,
experimental research concerning the effectiveness of commonly accepted practices spurred
criminologists and criminal justice agencies to
revise and improve intervention strategies.
The scientific justification for placerandomized trials is the assurance that if the
trial is carried out properly, there are no systematic differences between groups of places randomized, which in turn carries a guarantee of
statistically unbiased estimates the intervention’s
effect. It also assures that chance and chance
imbalances can be taken into account and
that a legitimate statistical statement of one’s
confidence in the results can be made. Further,
Weisburd (2005) and others have pointed out that
the simplicity and transparency of the idea of fair
comparison through a randomized trial has
a strong appeal for policy makers and practitioners who cannot understand and do not trust
complex model-based analyses of data from
nonrandomized studies. For the abiding statistician, the crucial aspect of simplicity is that
the statistical inferences as to the effect’s size
relative to chance need not depend on econometric, statistical, or mathematical models. The
randomization feature permits and invites less
dependence on such speculation. And modern
methods permit the use of randomization tests.
Place-Based Randomized Trials
The empirical evidence on the vulnerability of
evidence from nonrandomized trials in comparison to the evidence from randomized trials has
been building since at least the late 1940s. Assuring that one does not depend on weak and easily
assailable evidence when stronger evidence can
be produced is an incentive at times in parts of the
policy community. Assuring that one does not
needlessly depend on heroic assumptions to produce good estimates of effect, assumptions often
required in the nonrandomized trials, is an incentive for the scientific and statistical community.
Boruch (2007) summarizes these concepts in the
field of health care, employment and training, and
economics.
Shadish and colleagues (2008) provided a persuasive illustration that is especially compelling
because comparisons between a randomized
experiment and an observational study were anticipated as part of their research design and before
the data were collected. Their results suggest comparability, as opposed to major differences, if the
quasi-experiment is designed well in a particular
domain. Theirs is an interesting and potentially
important specific case. More generally, the biases
in estimating an intervention’s effect based on the
quasi-experiments can be very large, small, or
nonexistent (Weisburd et al. 2001). The variance
in estimates of effects appears to be typically
larger in the quasi-experiments than in the randomized tests. So far, and with some narrow
exceptions, there is no way to predict the directionality or magnitude of such biases or in the
variances of the estimates of the intervention’s
effect based on a nonrandomized trial.
It remains to be seen whether similar methodological studies on aggregate level analyses using
randomized versus nonrandomized places, clusters of individuals, or groups yield similar results,
that is, uncover serious biases in estimating
effects or the variance of the estimates or both.
But it is reasonable to expect biases here also.
Bertrand, Duflo, and Mullainathan (2002), for
instance, focused on biases in estimates of the
standard error of effects assuming no effect at
all using conventional difference in differences
methods and found type I error rates that were
nine times the error rate presumed (0.05) in using
3529
P
conventional statistical tests. This was partly on
account of serial correlation. More methodological research, however, needs to be done on the
quasi-experimental approaches to aggregate level
units so as to understand when the biases in
estimates of effect appear, when the biases in
estimates of their standard errors appear, and
how large the mean square error is, relative to
place-randomized trials.
It is important to identify dependable scenarios
in which bias and variance of estimates generated
in nonrandomized trials are tolerable. Doing so
can reduce the need for randomized trials (Boruch
2007). It is also not easy, as yet, to identify particular scenarios in which bias in estimates of
effect or variance will be small, as was the case
in some work by Shadish et al. (2008). Berk’s
(2005) more general handling of the strengths
and weaknesses of randomized controlled trials
at the individual level is pertinent to the analysis
of data generated in place-randomized trials also.
The scientific justifications that are identified
here are important in the near term. In the long
term, it would be good to understand what other
incentives are and to make these explicit at different levels, for example, policy, institution (agencies), and individual service-provider levels.
Incentives for better evidence may differ depending
on whether the stakeholders are members of the
police force at different levels, the mayor’s office,
and the community organizations that have a voice,
and so on. Many police executives, for instance,
want to improve policing and produce evidence on
whether things do improve and also usually want to
keep their jobs. The two incentives may not always
be compatible if a city council or mayoral preferences are antagonistic toward defensible evidence
generated in well-run field tests. Sturdy indifference to dependable evidence of any kind is, of
course, a problem in some policy sectors.
Deploying the Intervention:
Implementation, Dimensionalization,
and Measurement
Justifications and incentives are essential for
assuring that places, and influential people within
P
P
3530
them, are willing to participate in a randomized
trial. Understanding how to deploy a new program requires expertise at a ground level, as well
as substantial cooperation between researchers
and mid- and upper-level management. It also
requires constant communication between
researchers and street-level practitioners to
ensure that their daily practices conform to the
treatment that they intend to occur.
The Drug Market Analysis Program, which
fostered a series of randomized experiments on
crime hot spots, suggests that “ordinary” criminal
justice agencies can be brought on board to participate in experimental study if there is strong
governmental encouragement and financial support that rewards participation (see Weisburd
et al. 2006). A similar experience in the Spouse
Assault Replication Program (SARP) reinforces
these observations. Joel Garner (2002), who
served as program manager for SARP, noted
that he knew that the program was a success the
“day that we got 17 proposals with something like
21 police agencies willing to randomly
assign offenders to be arrested” (Weisburd
2005, p. 232).
Day-to-day implementation of placerandomized trials also requires significant
forethought on the part of both the researchers
and middle managers who are involved with
the experiments. For instance, during Lum and
colleagues’ (2010) trial of the effectiveness of
license plate recognition systems, the research
team made an effort to directly involve both
the officers who would be implementing the
intervention and the middle managers who
would be responsible for directly supervising
the officers in the process of identifying
hot spots, developing mechanisms for reporting
officer patrol, and overseeing implementation.
Given that the intervention involved the participation of two separate police departments, it was
also important to understand and account for
different practices in supervising officers, in the
responsibilities of officers to continue to respond
to calls for service, and other functions.
Gaining an understanding of the experience and
obligations of these practitioners allowed the
researchers to revise plans for patrol deployment
Place-Based Randomized Trials
and ensure that the experiment could realistically
be implemented given the normal practices of the
two police departments.
The reports on deploying programs and the
randomized trials that are published in peerreviewed journals can be excellent, but they are
typically brief. The brevity invites broad questions
about how the authors’ experience in detail can be
shared with others, that is, through web-based
journals, reports without page limits, workshops,
and so on. It invites more scientific attention to the
question of how one can dimensionalize implementation and the engineering questions of how
to measure or observe implementation level inexpensively and how to establish a high threshold
condition for implementation.
Resources for the Trial’s Design,
Statistical Problems, and Solutions
Some of the important technical references for
trial design and model-based data analysis
include Hayes and Moulton (2009) on clusterrandomized trials in different countries, mainly
in the health arena. Murray’s (1998) book on
group-randomized trials focuses more on individuals within groups and considers diverse applications in the USA. Bryk and Raudenbush’s (2002)
text considers in detail model-based statistical
analyses of multilevel data that may or may
not have been generated in randomized trials.
The models are complex and entail assumptions
that the analyst may not find acceptable.
In the context of place-randomized, clusterrandomized, or group-randomized trials, there
are few publications covering the simplest and
least model-dependent approaches to analyzing
data from such a trial. Such approaches fall under
the rubric of randomization tests or permutation
tests. A simple randomization test, for instance,
involves computing all possible outcomes of the
trial, ignoring the actual allocation to intervention
or control conditions, and then making
a probabilistic judgment about the dependability
of the effect detected based on the distribution of
possible outcomes so generated. There is no
dependence on linear or other parametric models.
Place-Based Randomized Trials
Given the relatively small sample sizes
involved in place-randomized trials (typically
less than 200 are allocated to one or another
intervention) and given contemporary computing
capacity, generating the distribution of possible
outcomes in these approaches is relatively
straightforward. The basic concepts were
described by Sir Ronald Fisher in the 1950s for
trials in which individuals or plants were the
independent units of random allocation and analysis. The idea is directly relevant to trials in
which places are randomly allocated and inferences are made about the effects of intervention
in places.
The Campbell Collaboration organized conferences on place-randomized trials in 2001 and
2002 to update researchers on the design and
conduct of place-randomized trials. A special
issue of the Annals of the American Academy of
Political and Social Sciences (Vol. 599, 2005)
resulted from these. After 2002, the William T.
Grant Foundation funded workshops and software development to enhance the technical
capacity of researchers to design such trials and
analyze results. At the time of this writing, there
was no information on more specialized efforts
sponsored by the National Institute of Justice or
private foundations in criminology.
One of the major scientific challenges in
designing a place-randomized trial is assuring
that the size of the sample of places is large
enough to detect the relative effects of interventions. A statistical power analysis is essential to
this planning process and allows the researcher to
determine the number of places that are needed
given assumptions about the expected effect size,
randomization levels, and specific statistical tests
of hypotheses and related procedures. Userfriendly software for estimating the modelbased statistical power of a trial under various
assumptions about sample size and other factors
is available on the William T. Grant Foundation’s
website (http://www.wtgrantfdn.org).
There are a number of common mistakes made
during analysis of place-randomized trials. First,
researchers have often wrongly employed the
number of people in places for hypothesis testing
and power analysis rather than employing the
3531
P
number of places randomly allocated (see, e.g.,
Kelling et al. 1974). Second, simple linear regression models are often used to analyze data even
though many of the modeling assumptions are not
necessarily true, as the “uncertainty” in randomized trials stems from the assignment process,
rather than unobservable disturbances that are
assumed to be the source of variation in regression models (see Freedman 2008). It is very difficult to determine how misleading the results of
such regression analyses are likely to be in
multilevel contexts. Analyses based on regression models with categorical, count, or time-tofailure outcomes create additional problems
(Freedman 2008).
Place-based trials often involve the study of
people within places, and thus, a very small number of independent units (places) may be present
at the randomized level and a large number of
related units (people) may be nested within each
place. A key issue in this scenario is that individuals or other units within a place were not randomly assigned to the intervention or control
conditions. Randomization at the place level
ensures that preexisting differences between
treatment and control groups in the study are not
systematic; this is due to the matching and subsequent randomization process. However, since
all individuals within each place are automatically assigned the same condition, intervention,
or control, the analysis is riskier to carry out at the
individual level (Raudenbush and Bryk 2002;
Spybrook et al. 2006).
Prior techniques for clustered data involved
either aggregating all information to the group
level as in a randomization test or a “t-test” of
mean differences between place level outcomes
or have involved disaggregating all group-level
traits and confining attention to all the individuals
within the intervention and control conditions.
For Raudenbush and Bryk (2002), the problem
with the first method is that all within group
information is wasted; it is omitted from the
analysis. The problem with the second method
is that the observations are no longer independent, as all individuals within a certain intervention or control condition will have the same value
on a certain variable and will not be independent
P
P
3532
from one another on account of the intra-class
correlation among people within place. To
account for the dependence that arises when
using samples of individuals nested within
places, researchers can use several statistical
methods to correct for clustering, such as hierarchical linear modeling techniques (i.e., HLM)
and generalized least squares estimation techniques with robust standard errors.
There are several options for analyzing the
results of place-based randomized experiments
in a largely model-free manner. Small and colleagues (2008), for instance, emphasize randomization/permutation tests with different kinds of
adjustments for covariates, not relying on any
form of HLM. Imai, Kind, and Nall (2009)
advanced the state of the art by showing how
matched pair designs in place-randomized trials
can often enhance precision in estimates of the
effects of interventions, increase statistical power
of analysis in detecting effects, and better assure
unbiased estimates of variability of the effect
with a small number of places. Matched pair
designs can also be used in this context, but the
concern is that matching may unnecessarily
reduce degrees of freedom and statistical power
of the studies.
As a practical matter, the work by Imai and his
colleagues (2009) shows that the number of units
within a place are an important matching variable
that can reduce costs and increase statistical
power in a trial. Bloom and Riccio (2005)
advance the state of the art in another important
respect by coupling conventional comparison of
randomized places in a time series analysis of
data from places. While the coupling of randomized and nonrandomized approaches to estimating effects of interventions is not a new concept
(see, e.g., Boruch 1985), incorporating these
ideas thoroughly into design and analysis is
unusual.
Another novel contribution relating to designing place-randomized trials lays in the idea of
“step wedge” and “dynamic wait list” designs.
The wait list design assigns a random half sample
of eligible places to the intervention for some
period of time. The remaining half sample
receives the intervention after this period has
Place-Based Randomized Trials
ended, under the assumption that all units will
eventually receive the intervention. While this
approach satisfies ethical concerns that may
arise if a beneficial treatment is not administered
to certain places, it may be problematic if there
are treatment consequences for a delay in treatment. To address such concerns about the
sequence of treatment, places may also be
divided into subsets which are periodically
given treatments of a given intensity at a given
time period. In a related area of work, Hussey and
Hughes (2006) develop the idea of step wedged
designs, with the “steps” being a point at which a
subset of places is randomly assigned a treatment.
The step wedge and dynamic wait list designs
are innovative and have attractive features. However, the power and design analysis of these
approaches have relied on hierarchical linear
models with unusual and sometimes untestable
assumptions. There may be alternatives that are
less model dependent.
Registers of Place-Randomized Trials
Learning about place-randomized trials can be
difficult. Reports may be published in a variety
of scientific journals and may appear in
unpublished (gray) literature. Relying on webbased searches may not be as effective as hand
searches in terms of locating randomized trials in
social sciences. Taljaard et al. (2009), for
instance, found that fewer than 50 % of published
place-randomized trials in health are appropriately classified in titles and abstracts.
There are a number of data sources concerning
randomized trials that can help to reduce these
difficulties. For instance, the Cochrane and
Campbell Collaborations, which are concerned
with the fields of health care and social sciences,
respectively (including psychology, education,
criminal justice, and sociology), publish reviews
of results of randomized trials in which references to original sources can be accessed through
their websites. In the United States, one can learn
about trials in the health sector, including placerandomized trials, at http://www.clinicaltrials.
gov. Often the interventions tested or the
Place-Based Randomized Trials
outcome variables examined in these trials are
pertinent to criminologists as well as health-care
researchers and practitioners. Finally, in 2007
David Greenberg and Mark Schroder developed
a new register of randomized trials oriented
toward tests of economic interventions under
the auspices of the Social Sciences Research Network (SSRN).
Standards for reporting randomized trials
where individuals are the unit of allocation were
the product of the late 1990s. The Consolidated
Standards of Reporting Trials (CONSORT) statement, which provided guidance on the composition of a report on randomized trial in health care,
has been modified to provide guidance on
reporting place-randomized trials. For instance,
studies should provide information on the rationale for a randomized design and measure outcomes on all levels of sampling and statistical
inference. As of this writing, there are no uniform
standards for reporting randomized trials in criminal justice. These problems have been raised by
scholars of criminology who have emphasized
that “reporting validity” is critical for the development of experimental science in criminology
(Perry et al. 2009; Gill 2009; Farrington 2003).
A lack of clear and uniform standards in this area
of research means that reanalysis of experimental
studies is difficult and often impossible. While
some funding agencies have tried to make policy
to assure that independent analysts have access to
record data in research that they sponsor, clear
criteria such as that included in the CONSORT
statement are critical for advancing experimental
criminology.
Ethics and the Law
As of this writing, no professional society or
government agency has promulgated explicit
statements about the ethical propriety of placerandomized trials. Applying contemporary standards to place-randomized trials can be awkward
and imperfect (see, e.g., Boruch 2007), and at
times, standards of ethics and government regulations are not clearly relevant to placerandomized trials (see, e.g., Taljaard et al. 2009,
3533
P
for a discussion of these issues in health care). For
instance, ethical standards established for the
treatment of human subjects are arguably inappropriate for randomly allocated treatment in
a city’s crime hot spots. The ethical issues surrounding place-based trials in criminology are
complex. For instance, researchers may confront
concerns about equality of treatment, as in
research surrounding hot spot policing, wherein
residents were concerned that concentrated patrol
in one area would result in reduced services in
other locations (see Weisburd 2005). Or officers
could feel that sitting for extended periods of time
reduces their capacity to protect the public
(Weisburd et al. 2006). Finally, residents of
areas that are designated as crime hot spots may
feel as if they are being unfairly targeted by the
police. In these cases, the implementation of
a place-randomized trial can undermine the relationship between the practitioners involved in the
trial and the clients whom they serve (see also
Clarke and Cornish 1972).
Regardless of the ethical dilemmas engendered by place-based trials over the last decade,
there appear to have been no serious challenges in
the US courts to the conduct of place-randomized
trials. For instance, the former director of the
Institute for Education Sciences (IES), Russ
Whitehurst, reported in a personal communication (2008) that he had encountered no court
challenges as a consequence of the IES’s sponsoring many such trials in education during
2002–2008. Similarly, we are aware of no
judicial challenges in the context of placerandomized trials in the crime sector, such as
the trials on crime hot spots, bar room
violence, convenience store vulnerability to
holdups, and so on.
Conclusions and Implications for the
Future
Place-randomized trials have become important
not only in criminology but also in a broad range
of other disciplines because they employ substantive theory about the effects of intervention at the
place level. The experience of people who have
P
P
3534
been involved in the design, execution, and analysis of place-randomized trials is an important
source of intellectual and social capital. This
experience must be exploited by future
researchers and incorporated into graduate and
postgraduate education, as should the knowledge
of researchers in disciplines such as medicine and
education who have substantially advanced the
state of knowledge about more generally clusterrandomized trials. The statistical armamentarium
for design and analysis of place-randomized trials
is fundamental and lies on the simple idea
of randomization rather than complex
statistical models. Future development of placerandomized trials should better explore the implications of randomization of different interventions when the expected effects of the
interventions are in “equipoise” and advance
investigations into the ethics of this type of
research. Further, standardized reporting of
design, execution, and analysis of placerandomized trials should be advanced in criminology to allow for reanalysis and secondary
analysis, particularly of controversial studies.
Walter Lippmann, an able social scientist and
newspaper writer, had a strong interest in cops
and crimes by adults and adolescents and was
familiar with political ambivalence about or
opposition to sound evidence. He was a streetlevel criminologist, remarkable writer, and good
thinker. In the 1940s, Lippmann said: “The problem is one for which public remedies are most
likely to be found by choosing the most obvious
issues and tackling them experimentally. . .the
commissions of study are more likely to be
productive if they can study the effects of
practical experimentation.” Nowadays, trialists
in criminology would have little difficulty in subscribing to Lippmann’s counsel.
Related Entries
▶ Cambridge-Somerville Youth Experiment
▶ Criminology of Place
▶ History of Geographic Criminology Part I:
Nineteenth Century
Place-Based Randomized Trials
▶ History of the Statistics of Crime and Criminal
Justice
▶ Hot Spots and Place-Based Policing
▶ Randomized Block Designs
▶ Randomized Experiments in Criminology and
Criminal Justice
Recommended Reading and References
Berk R (2005) Randomized experiments as the bronze
standard. J Exp Criminol 1(4):417–433
Bertrand M, Duflo E, Mullainathan S (2002) How much
should we trust difference in differences estimates?
Working paper 8841. National Bureau of Economic
Research, Cambridge
Bloom HS, Riccio JA (2005) Using place random assignment and comparative interrupted time-series analysis
to evaluate the jobs-plus employment program for
public housing residents. Ann Am Acad Polit Soc Sci
599:19–51
Boruch RF (2007) Encouraging the flight of error: ethical
standards, evidence standards, and randomized trials.
N Dir Eval 133:55–73
Boruch RF, Wothke W (1985) Seven kinds of randomization plans for designing field experiments. N Dir Prog
Eval 28:95–113
Bryk AS, Raudenbush SW (2002) Hierarchical linear
models. Sage, Thousand Oaks
Clarke RVG, Cornish DB (1972) Controlled trial in institutional research – paradigm or pitfall for penal evaluators. The Home Office, London
Farrington DP (2003) Methodological standards for evaluation research. Ann Am Acad Polit Soc Sci
587(1):49–68
Freedman DA (2008) Randomization does not justify
logistic regression. Stat Sci 23:237–249
Gill C (2009) Reporting in criminological journals. Report
for seminar on advanced topics in experimental
design. Graduate School of Education and Criminology Department, University of Pennsylvania,
Philadelphia
Hussey MA, Hughes JP (2006) Design and analysis of
stepped wedge cluster randomized trials. Contemp
Clin Trials 28(2):182–191
Imai K, King G, Nall C (2009) The essential role of pair
matching in cluster randomized experiments, with
application to the Mexican universal health insurance
evaluation. Stat Sci 24:29
Kelling GL, Pate T, Dieckman D, Brown CE (1974)
Kansas City preventative patrol experiment: a technical report. Police Foundation, Washington, DC
Lippmann W (1963) The young criminals. In: Rossiter C,
Lare J (eds) The essential Lippmann. Random House,
New York, Originally published 1933
Plea Bargaining
Lum C, Merola L, Willis J, Cave B (2010) License plate
recognition technology (LPR): impact evaluation and
community assessment. Accessed from http://gemini.
gmu.edu/cebcp/lpr_final.pdf
Mazerolle L, Price J, Roehl J (2000) Civil remedies and
drug control: a randomized trial in Oakland, California. Eval Rev 24(2):212–241
McCord J (2003) Cures that harm: unanticipated outcomes of crime prevention programs. Ann Am Acad
Polit Soc Sci 587(1):16–30
Murray D (1998) Design and analysis of group randomized trials. Oxford University Press, New York
Perry AE, Weisburd D, Hewitt C (2009) Are criminologists reporting experiments in ways that allow us to
access them? Unpublished manuscript/report
Raudenbush SW, Bryk AS (2002) Hierarchical linear
models: applications and data analysis methods.
Sage, Thousand Oaks
Shadish WR, Clark MH, Steiner PM (2008) Can
nonrandomized experiments yield accurate answers?
A randomized experiment comparing random and
nonrandom assignments. J Am Stat Assoc
103(484):1334–1356
Sherman LW, Weisburd DL (1995) General
deterrent effects of police patrol in crime “hot spots”:
a randomized, controlled trial. Justice Q 12(4):
625–648
Spybrook J, Raudenbush SW, Liu X, Congdon R (2006)
Optimal design for longitudinal and multilevel
research: documentation for the “Optimal Design”
software. Accessed from http://rmcs.buu.ac.th/
statcenter/HLM.pdf
Taljaard M, Weijer C, Grimshaw J, Bell Brown J, Binik A,
Boruch R, Brejhaut J, Chaudry S, Eccles M, McRae A,
Saginur R, Zwarenstein M, Donner A (2009) Cluster
randomized trials: rational design of a mixed methods
research study
Warburton AL, Sheppard JP (2000) Effectiveness
of toughened glass in terms of reducing injury in
bars: a randomized controlled trial. Injury Prev
6:36–40
Weisburd D (2003) Ethical practice and evaluation of
interventions in crime and justice: the moral
imperative for randomized trials. Eval Rev 27(3):
336–354
Weisburd D (2005) Hot spots policing experiments and
lessons from the field. Ann Am Acad Polit Soc Sci
599:220–245
Weisburd DL, Green L (1995) Policing drug hot spots: the
Jersey City drug market analysis experiment. Justice
Q 12(4):711–735
Weisburd D, Lum C, Petrosino A (2001) Does research
design affect study outcomes? Ann Am Acad Polit Soc
Sci 578:50–70
Weisburd D, Wycoff L, Ready J, Eck JE, Hinkle JC,
Gajewski F (2006) Does crime move around the corner? A controlled study of spatial displacement and
diffusion of crime control benefits. Criminology
44(3):549–591
3535
P
Plea Bargaining
Ronald F. Wright
Wake Forest University, Winston-Salem,
NC, USA
Synonyms
Negotiated guilty pleas
Overview
As criminal justice systems increase in volume,
they turn to negotiated guilty pleas as a primary
method to dispose of cases. The prosecution
offers reduced sentencing risk to defendants in
exchange for more certain and inexpensive
convictions. Although such negotiated outcomes
are unpopular with the general public, the
full-time practitioners who operate the system
think of plea bargains as a necessary practice in
busy courts. Negotiated guilty pleas have become
by far the most common disposition of criminal
charges in the United States, and similar practices
are becoming more prevalent in other countries.
Plea bargains create some uncertainty about the
accuracy of convictions in the system and alter
the balance of power among criminal justice
actors by making judges and juries less influential
by comparison to prosecutors. While external
legal limits on plea bargains have mostly proven
ineffectual, prosecutor offices routinely place
internal administrative limits on the use of plea
bargains.
Introduction
Criminal defendants in the United States usually
decide to plead guilty. Very often, defendants
negotiate with the prosecution to receive
particular benefits as a precondition to entering
a guilty plea – so often, in fact, that these
so-called plea bargains have become more
P
P
3536
common and more important than criminal trials.
In the United States, more than 9 out of every 10
felony convictions (and more than 99 out of every
100 misdemeanor convictions) derive from guilty
pleas, and the overwhelming majority of guilty
pleas occur after the parties negotiate the terms of
that plea. Plea negotiations have been slower to
arrive in other countries, but criminal court
systems around the world often deal with
increased volume by allowing some form of
discounted sentence in return for a negotiated
settlement. Plea bargains, in some form or
another, are critical to the operation of
high-volume criminal justice systems around the
world today.
The dominance of plea bargaining has
predictable effects on the institutions of criminal
justice in the United States. Most immediately, it
makes criminal juries less influential. Negotiated
pleas also shift power away from the trial judge
toward the parties; they particularly strengthen
the hand of the prosecutor. The prosecutor does
not merely present legal and factual arguments to
the judge on an equal footing with the defense
attorney. Instead, the prosecutor hears arguments
from the defense and then selects from the available charges and evidence thus substantially
impacting the ultimate sentence. The judge, in
effect, reviews these prosecutor decisions for
possible error in exceptional cases. This shift of
power in favor of prosecutors is more pronounced
in “guideline sentencing” states that link the
sentence more closely to the charge of conviction
and to particular factual findings about the crime.
Plea bargaining also affects outcomes in the
criminal courts. For one thing, system volume
goes up. A system that relies heavily on
negotiated pleas can process more cases than
a jurisdiction that resolves more of its cases
through trials. Negotiated plea bargains also
probably trade the quality of convictions for
a higher quantity of cases. As systems get busier,
the risk of erroneous convictions increases.
External limits on the terms of negotiation
between prosecution and defense are available
in some systems but are still unusual and have
shown only limited practical effects. Internal
limits on bargaining practices imposed within
Plea Bargaining
the bureaucratic hierarchy of the prosecutor’s
office are still the most common limits on individual sentencing discretion of a prosecuting
attorney.
This entry begins with a review of the most
common topics for bargaining between the
defense attorney and the prosecutor in criminal
cases. It then discusses the prevalence and
systemic effects of those party negotiations,
both domestically and internationally. The entry
closes with a review of the external and internal
controls that legal systems use to structure
prosecutorial discretion.
Topics for Plea Negotiations
During plea negotiations, the prosecution can
offer the defendant two categories of benefits in
exchange for a plea of guilty. First, under
a “sentence bargain,” the prosecutor agrees to
recommend to the judge a sentence below the
maximum available, without amending the
charges. A variation on this type of agreement
involves a “fact bargain.” In this situation, the
prosecution and defense agree to represent to
the court that certain facts were either present or
absent in a case. Under the sentencing rules of
some jurisdictions, these facts trigger specific
sentencing consequences.
The second major category of negotiated plea
aims for a “charge bargain.” In this situation, the
prosecutor agrees to reduce the most serious
charge to some lesser offense or to reduce the
total number of counts in the indictment or
information. The effect of amending the charges
is to reduce the defendant’s potential exposure to
more severe sentence outcomes.
Sentence Bargains
The parties develop detailed knowledge about the
strength of the evidence, the expectations of the
victims, the most likely outcomes of any legal
issues presented, and the normal sentences
imposed in similar cases. The judge, meanwhile,
does not get involved in the details of the case as
early as the parties and never becomes as familiar
as the parties with the specifics of the crime
and the offender’s background. When the
court faces a crowded docket and the parties
Plea Bargaining
assure the judge that a given sentence is
appropriate – particularly if that recommendation
falls roughly into a normal range of sentences
for this crime – the judge usually accepts the
recommendation (Scott and Stuntz 1992).
Sometimes the parties want a level of certainty
about the sentencing outcome that is not possible
based
on
sentencing
recommendations
alone. The criminal procedure rules in some
jurisdictions allow the defendant to enter a plea
of guilty conditioned on the judge’s acceptance
of the agreed-upon sentence. In other words, the
defendant may withdraw the plea of guilty if
the court later rejects the specified
sentence. Some scholars favor sentence
bargains that give the judge something more
than a “take it or leave it” option; more
substantial input for the judge in the
selection of the sentence leads, according to this
view, to a more balanced sentencing system
(Alschuler 1976; Wright and Miller 2002).
Before accepting a plea of guilty, the trial
court judge must be convinced that the
defendant’s plea is “knowing and voluntary”
and that the prosecution holds a sufficient
“factual basis” to prove each element of the
crime beyond a reasonable doubt. Other facts
may be relevant to the sentence, as well, even if
they are not elements of the offense. For instance,
in the federal system, a defendant who plays
a “minimal role” or “minor role” in a group
offense receives a lower sentence.
Given this opportunity to control the available
sentence through the facts they plan to present,
the parties naturally turn to negotiations about the
proof of these non-element facts; they agree in
some cases to tell the judge that the fact is present
or absent from the case. The parties cannot lie to
the judge and cannot compel the judge to make
the factual finding that forms the basis for their
agreement. But the prosecutor and defense
attorney do exercise serious practical influence
when they present a united front to the judge.
This subtype of sentence bargaining is
sometimes known as “fact bargaining.” It is
especially pertinent in jurisdictions with sentencing guidelines or other statutes that attach specific
sentencing consequences to particular facts.
3537
P
For instance, some state codes authorize
a 5-year increase in a sentence based on the use
of a weapon during commission of the crime.
The prosecution and defense might negotiate an
agreement to stipulate to the judge that no
weapon was used during the crime.
The prosecutor sometimes goes beyond the role
of advocate and plays a strong gatekeeper role for
some non-element facts. In these special situations, the prosecutor does not merely request
a factual finding from the judge; instead, the prosecutor herself exercises the power to increase or
decrease a sentence automatically. For example, in
the federal system (and in some states), the
offender is eligible for a sentence discount after
cooperating in the government’s investigation of
further crimes only if the government certifies that
the defendant provided “substantial assistance.”
The government’s willingness to file such
a motion is one of the important bargaining chips
available to the prosecutor. Some states give the
prosecutor a similar gatekeeper function over certain sentencing enhancement facts, such as the
proximity of a narcotics sale to school grounds or
the fact that a defendant’s prior record makes him
or her a “habitual felon.” The defense attorney and
prosecutor routinely bargain to determine whether
the prosecutor will file the allegations that trigger
such a sentencing enhancement.
P
Charge Bargains
Criminal codes in both the federal and state
systems give prosecutors a generous menu of
options in the selection of charges. Many
common fact scenarios could support criminal
charges under multiple sections of the criminal
code, each leading to different potential
sentencing outcomes.
After the initial filing of charges and the assignment of defense counsel to a case, the parties often
negotiate over possible amendments to those
charges. The defense attorney might request
a dismissal of the most serious charge to be
replaced by a less serious charge (a “vertical”
charge bargain). Alternatively, the defense lawyer
might ask for a dismissal of some charges in
a multi-count indictment (a “horizontal” charge
bargain). Both types of amendments reduce the
P
3538
maximum sentence that a judge could legally
impose on the defendant.
Negotiations between the parties about revised
charges take on even greater importance in the
context of more highly structured sentencing systems. Sentencing laws that structure or reduce
judicial discretion, including “mandatory minimum sentences” and “presumptive sentencing
guidelines,” make the charge of conviction
a more important predictor of the sentence. Once
the defendant is found guilty of a particular crime,
these highly structured sentencing laws give
judges relatively few options for how to sentence
the defendant (and perhaps no options at all).
Because a structured sentencing environment
makes visible the sentencing consequences of
charging decisions, the parties can isolate the
impact of a charge reduction and negotiate based
on a more certain prediction of the outcome.
Mandatory minimum sentencing laws offer
a clear-cut example of the impact of charge
bargains on sentencing outcomes. These laws create “cliff effects” that dramatically affect the
potential sentence, depending on whether the
parties agree to the mandatory sentence versus a
similar crime with no mandatory penalty attached.
Mandatory penalties could lead to wholesale
increases in sentence severity in a jurisdiction.
Empirical studies confirm, however, that prosecutors selectively mitigate the impact of specific
mandatory sentencing laws by dismissing or
reducing charges in some cases (Bjerk 2005).
Presumptive sentencing guidelines tie
sentencing options to the charge of conviction,
much like mandatory minimum sentencing laws
do. Systems that link the charge of conviction to
a narrow range of sentence outcomes limit the
options for the sentencing judge and empower
the prosecutor, who controls both the filing of
charges and the proof of the defendant’s conduct.
Empirical studies have examined what
prosecutors in guideline jurisdictions actually do
about charge reductions. For the least serious
cases, studies found only small shifts from
sentence bargains to charge bargains.
Apparently, busy prosecutors in high-volume
systems do not change their customary
negotiation practices across the board.
Plea Bargaining
Charge bargains increased more markedly,
however, in more serious cases. In cases where
conviction on the original charge would result in
a presumed prison sentence and conviction on
a lesser charge would allow probation or a
shorter jail sentence, charge bargains happened
more often after the arrival of sentencing
guidelines in a jurisdiction. The amended
rules gave the parties greater control over
the sentencing outcomes in those cases, and
they took advantage of that new power
(Miethe 1987; Frase 2005). Research on the
charging and plea bargaining practices of
federal prosecutors under the federal
sentencing guidelines also suggests that
prosecutors frequently exercise their charge
bargaining discretion to reduce sentences,
particularly in drug and weapon possession
cases (Bowman and Heise 2001). On the whole,
the empirical evidence suggests that sentencing
guideline systems make charge bargains
somewhat more attractive than they are in
discretionary sentencing regimes.
There are also questions about the equal
distribution of charge bargain benefits. The
presence of legally relevant facts is the most
important determinant of the charge of
conviction (i.e., the charge that ultimately forms
the basis for the conviction). Some nonlegal
factors, however, also have some effect on the
charge of conviction. Factors such as the race of
the defendant appear to have a relatively small
but persistent impact on the outcomes
(Steffensmeier et al. 1998).
Defense Concessions
While the prosecutor can offer reduced sentences
to the defendant, the defense attorney also holds
certain bargaining chips during plea negotiations.
Each of the defendant’s concessions involves
the removal of procedural hurdles from the
prosecutor’s path.
The defendant’s power to make the government’s work easier begins with the investigation.
If the defendant cooperates in an ongoing
investigation, the government might be able to
convict additional defendants. This form of
assistance is especially important in the federal
Plea Bargaining
Plea Bargaining,
Fig. 1 Percent of
convictions obtained by
guilty plea in federal court,
1950–2010
P
3539
Percent of Convictions Obtained by Guilty Plea
in Federal Court, 1950-2010
100%
98%
96%
94%
92%
90%
88%
86%
84%
82%
80%
2010
2005
2000
1995
1990
1985
1980
1975
1970
1965
1960
1955
1950
system. The discounts available for “substantial
assistance” lead to some difficult anomalies in the
sentences among defendants who commit their
crimes as part of a group, with the largest
discounts awarded to the most blameworthy
(and knowledgeable) organizers of the criminal
enterprise (Maxfield and Kramer 1998).
Another procedural hurdle that the defendant
can remove for the prosecution is discovery and
disclosure. The prosecution has a constitutional
duty to disclose all material exculpatory
information in its possession and a duty under
state statutes or procedural rules to respond to
any discovery requests for certain types of
inculpatory evidence. The defendant, however,
can remove most of those obligations from the
prosecution by pleading guilty.
The most valuable concessions that
defendants make in plea negotiations are waivers
of pretrial hearings and the trial itself. These
waivers might extend to all the procedural rights
at trial, including the right to a jury, confrontation
of adverse witnesses, counsel, and so forth. In
some jurisdictions, defendants can obtain some
benefits by offering to waive the jury trial in favor
of a bench trial (Schulhofer 1984).
The defendant can also offer the government
certainty and finality of outcomes by waiving the
right to appeal or to file post-conviction collateral
attacks on the conviction. Most state and federal
courts have concluded that a defendant may
explicitly waive the right to appeal a conviction
as part of a plea agreement. One empirical study
of federal cases found that nearly two-thirds of
the cases settled by plea agreement included
a waiver of appeal rights and three-quarters of
the defendants who waived appeal also waived
collateral review (King and O’Neill 2005).
Prevalence of Plea Negotiations
The concessions that the prosecutor and the
defense attorney offer each other during plea
negotiations have become the norm in criminal
justice: plea agreements account for the great
majority of convictions in every jurisdiction in
the USA. The proportion of negotiated pleas
changes over time and across different systems.
Negotiated outcomes were not at all common in
the mid-nineteenth century in state systems (Fisher
2003). In the federal system, less than 8 % of
convictions were obtained through guilty pleas in
the early 1970s; that number rose inexorably
through the decades to reach the current level,
above 97 % (Wright 2005). See Fig. 1.
A high volume of cases within a court
system – both the civil and criminal
dockets – creates the conditions for plea
bargaining to thrive (Bibas 2012). Courts that
resort to plea bargaining, in turn, increase their
caseloads even further. Routine practices and
expectations also play a role in sustaining the
institution of plea bargaining: some courts with
relatively uncrowded dockets still depend almost
entirely on negotiated pleas.
The United States is unusual in an
international context in its heavy reliance on
P
P
3540
party negotiations to resolve criminal
proceedings and the tight connection it promotes
between plea negotiations and sentencing
outcomes. For example, only a generation ago,
Germany was known as the “land without plea
bargaining” (Langbein 1979). Over time,
however, many European nations have faced
a higher volume of criminal cases; crowded
court dockets have led to various innovations
that allow the prosecutor to designate some
cases for summary treatment.
In some countries, party negotiations
to achieve these more streamlined criminal
adjudications have become more common.
Parties not only negotiate guilty pleas but also
dismissals and diversion into alternative
punishment or restitution programs (Jehle and
Wade 2006; Luna and Wade 2010). In the past,
these negotiated summary dispositions were
restricted to less serious cases, but that limitation
is disappearing over time (Langer 2004). Plea
bargaining is somewhat more prevalent in Israel,
Canada, Australia, England, and Wales than in
European Continental countries.
Negotiation Effects
It is difficult to say whether plea bargains reduce
the typical sentence imposed on a defendant over
the long run. On the one hand, defendants
who plead guilty to a crime receive a definite
“discount” for waiving trial rights, when
compared to defendants facing the same charges
during the same time period. On the other hand,
as the courts rely more heavily on plea bargains
and prosecutors file more charges, the
legislature tends to pass new criminal laws that
increase the risk for defendants who refuse to
plead guilty. That is, the new laws
authorize higher sentences that defendants
potentially face if they insist on a trial and lose.
Over the long run, reliance on plea bargains
does not correlate with reduced sentences
(Pfaff 2011).
Other potential effects of plea bargains go
beyond the sentences imposed, raising issues
about the accuracy of system outcomes and the
influence of different system actors.
Plea Bargaining
Questions About Accuracy
Established constitutional doctrine declares that
the courts may not punish a criminal defendant
simply for exercising the constitutional right to
a trial. Daily practice in every American criminal
court, however, contradicts this doctrine. It is
clear that defendants routinely receive
more severe sentences after trial than they
would receive if they were to plead guilty
(Brereton and Casper 1981). Judges resolve this
apparent conflict between constitutional doctrine
and daily reality by declaring that they reduce
sentences in guilty plea cases based on the
defendant’s cooperative attitude and prospects
for rehabilitation, and not based simply on the
fact that the defendant waived trial rights.
The exact size of the “trial penalty” – or to put
it more politely, the “guilty plea discount” – is
difficult to measure. Studies of the trial penalty
that attempt to control for the seriousness of the
offense and other variables find a substantial
gap
between
posttrial
sentences
and
post-plea sentences. One study found a wide
range of differences – between 13 %
and 461 % – depending on the crime and the
jurisdiction involved (King et al. 2005).
When defendants face such a large trial
penalty, concerns start to mount that some
defendants with valid defenses nevertheless
plead guilty. Although there is a reasonable
prospect that such defendants would be acquitted
at trial, they dare not risk the large increase in the
sentence that happens after a conviction at trial
(Wright 2005). Behavioral economics offers
some reasons to believe that defendants will
undervalue the long-term impact of a felony
conviction, leading defendants to accept a guilty
plea too easily (Bibas 2004). On the other hand,
there is evidence that some innocent
defendants resist highly attractive plea offers,
even when it might be rational to accept the
offer in light of the risk of a wrongful conviction
(Gazal-Ayal and Tor 2012).
The risk of coercing innocent defendants into
pleading guilty is built into some basic structural
features of American criminal justice. Criminal
codes that offer more bargaining options to the
Plea Bargaining
prosecutor make it possible to increase the trial
penalty and to pressure some defendants into
waiving potentially effective defenses. The
same holds true for increased sentence severity.
When the law authorizes higher maximum
sentences for a wide range of offenses and retains
low potential sentences for lesser-included
offenses, defendants face an enormous range of
risk. Particularly in those states with sentencing
guidelines or other limits on judicial sentencing
discretion, the defendant can control some of that
risk through a charge bargain. It would be an
overstatement to claim that sentencing guidelines
lead directly to more inaccurate convictions.
They can, however, contribute to a coercive
environment for defendants.
Finally, the structure of criminal justice
institutions increases the risk of coercive and
inaccurate guilty pleas. The public does not
invest in enough judges, courtrooms, prosecutors,
and public defenders to try a substantial
proportion of the cases filed each year. The
“working group” dynamic that develops in most
courtrooms places the highest value on
agreements that will move cases more quickly
through the system (Nardulli et al. 1985).
3541
P
from judges to prosecutors since judges might
find other methods to influence sentences
(Engen 2008).
At the very least, fears that prosecutors would
entirely usurp the judge’s sentencing authority
in presumptive guideline jurisdictions appear to
have been overstated. For one thing, judges in an
indeterminate sentencing system do not, in reality, control the sentence that an offender actually
serves. The judge announces one sentence, but
parole authorities could later reduce that sentence. Since the judge in an indeterminate system
never held actual power over the sentence to be
served, guidelines do not take that power away
from the judge. It is also true that presumptive
sentencing guidelines leave important zones of
discretion available to judges in the selection of
sentences.
Limits on Prosecutorial Discretion
Legal institutions place some controls on the
negotiation of guilty pleas. Those constraints
originate both from outside the office of the
prosecutor and from the internal workings of
a local prosecutor’s office.
External Constraints
Balance of Power Among Sentencing Actors
Sentencing guidelines and other structured
sentencing laws that began to proliferate in the
1970s were designed to regulate judicial
discretion in sentencing. As a result, some critics
of these laws expressed concern that the reforms
would transfer power from judges to prosecutors.
Compared
to
traditional
indeterminate
sentencing schemes, the more structured systems
would concentrate sentence authority in one
branch rather than allowing one institution to
check and balance the other (Alschuler 1991).
The transfer of power hypothesis finds some
tentative empirical support. For instance, studies
of guideline systems confirm that charge bargains
become more common for at least some crimes,
and charge reductions determine an important
component of the sentence actually served
(Frase 2005). But this shift in charging practices
may not result in a provable transfer of authority
If the legislature defines crimes narrowly and sets
penalties at modest levels, it reduces the risks of
inaccurate convictions and increases the power of
different sentencing actors to check each another.
This approach to crime legislation, however, does
not thrive in the American political climate.
Voters expect prosecutors to take the lead in
addressing crime, and they expect legislators to
give them the legal tools to do the job. Legislators
respond with broadly worded criminal laws and
multiple statutes (each with a different
corresponding punishment) that could apply to
a single common factual scenario.
Although legislators do not seriously constrain
prosecutors through the terms of the substantive
criminal law, statutes in a few jurisdictions do
limit the timing of plea negotiations or limit the
size of the charge reduction that a prosecutor can
offer to dispose of a case (at least for some high
priority categories of crime). While these
P
P
3542
legislative directives can be meaningful, their
current impact is small. Statutory limits on the
timing of negotiations simply push plea bargains
into earlier phases of the proceedings.
Judges also refuse, for the most part, to monitor
and control the negotiation of guilty pleas.
Granted, judges hold the power to accept or reject
guilty pleas, along with the plea agreements that
the parties present to them. These judicial powers,
however, operate within a system of mass justice.
The caseload would become overwhelming if
judges balked regularly at proposals to remove
a case from the trial docket, or even took the
time regularly to investigate this possibility.
The law in some jurisdictions also limits the
role of the judge during plea negotiations.
Statutes, rules of criminal procedure, rules of
judicial ethics, and judicial opinions in more
than half of the states instruct the judge not to
“participate” at all in the plea discussions. The
judge’s only role is to evaluate the legitimacy of
the guilty plea after the parties finalize
their agreement. A smaller number of states
discourage judges from participating in plea
negotiations, but they do not ban the practice
outright. The laws in these states allow the
judge to comment on the acceptability of charges
and sentences that the parties themselves propose
or to participate in the negotiations only upon the
invitation of both parties.
The rules of professional responsibility as
enforced by state licensing authorities are
also a potential source of limits on the plea
bargaining behavior of prosecutors. For instance,
a prosecutor who withholds discoverable
information from the defense during plea
negotiations might violate the specialized ethical
obligations of prosecuting attorneys. Again,
however, we get limited accountability from
these regulators. State bar authorities rarely
discipline prosecutors, and the penalties are
usually not severe (Zacharias 2001).
Although institutions external to the
prosecutor’s office do not exert much power
over the discretion of line prosecutors, the chief
prosecutor is directly accountable to the voters.
Prosecutors in the United States are normally
Plea Bargaining
elected. Given the distaste among voters for the
practice of plea bargaining, elections in theory
should limit the prosecutor’s ability to reduce
charges or to recommend lower sentences as
part of a plea negotiation. In practice, however,
the influence of voters over the plea bargaining
policies of the prosecutor’s office is limited.
The heavy advantage of incumbents in
prosecutorial elections makes this a weak
accountability mechanism (Wright 2009).
Internal Constraints
While legal institutions outside the prosecutor’s
office do not fully meet the need for
checks and balances, internal regulation has
a substantial constraining effect. Forces within
the prosecutor’s office can produce plea
policies that remain true to declared sources of
law, in keeping with current public
priorities in the enforcement of that law,
applied with reasonable consistency across
cases (Bibas 2012; Miller and Wright 2008).
These internal regulations take several forms.
First, the managers in a prosecutor’s office sometimes arrange the flow of cases to encourage line
attorneys to interact dispose of cases through plea
negotiations. This might involve the use of
“horizontal” prosecution for some crimes, with
different attorneys or units in the office making
decisions as a file moves up through the system. It
is also common, particularly in larger offices, to
require supervisor approval for any plea
agreement that dismisses a charge in priority
cases, such as homicide and domestic violence.
Second, chief prosecutors can promote
consistency and fidelity to public values among
their line attorneys by creating written guidelines
for the disposition of cases (Podgor 2012). The
United States Attorney’s Manual is one such
resource. The guidelines typically declare that
they do not carry the force of law and are not
enforceable in judicial proceedings, but they
nevertheless exert some control over the
behavior of prosecutors (Abrams 1971).
Through the use of such guidelines,
prosecutors have from time to time banned
the use of plea bargains for certain classes of
Plea Bargaining
cases, although such guidelines require serious
monitoring and enforcement to remain effective.
While there is much promise in the power of
chief prosecutors to hold their line prosecutors
accountable, the system also depends on
professional tradition, informal office culture,
peer pressure, and individual conscience to
achieve just results. At the end of the day, each
prosecutor must remain individually committed
to the ideal of responsible prosecution.
Policy Challenges
The most significant checks on the use of plea
bargains come from inside prosecutors’ offices.
Prosecutors in the United States are profoundly
decentralized: the state courts operate over 2000
separate prosecutors’ offices, with no effective
hierarchical control over the local offices in
most states. As a result, the internal controls on
plea bargains are often nontransparent and
inconsistent from place to place. Policy makers
who hope to promote enforcement of the criminal
law that is consistent with legal values and with
current popular priorities must find ways to make
these local internal policies and practices more
visible and subject to evaluation by the public and
by other legal actors. A balance of power, in
a transparent environment, would lead to the
most responsible use of negotiated guilty pleas.
Related Entries
▶ Prosecution and Wrongful Convictions
▶ Prosecutorial Discretion
Recommended Reading and References
Abrams N (1971) Internal policy: guiding the exercise of
prosecutorial discretion. UCLA Law Rev 19:1–58
Alschuler AW (1976) The trial judge’s role in plea
bargaining (pt. 1). Columbia Law Rev 76:1059–1154
Alschuler AW (1991) The failure of sentencing
guidelines: a plea for less aggregation. Univ Chic
Law Rev 58:901–951
Bibas S (2004) Plea bargaining outside the shadow of trial.
Harv Law Rev 117:2463–2547
3543
P
Bibas S (2012) The machinery of criminal justice. Oxford
University Press, New York
Bjerk D (2005) Making the crime fit the penalty: the role
of prosecutorial discretion under mandatory minimum
sentencing. J Law Econ 48:591–625
Bowman FO, Heise M (2001) Quiet rebellion? Explaining
nearly a decade of declining federal drug sentences.
Iowa Law Rev 86:1043
Brereton D, Casper J (1981) Does it pay to plead guilty?
Differential sentencing and the functioning of criminal
courts. Law Soc Rev 16:45–70
Engen RL (2008) Have sentencing reforms displaced
discretion over sentencing from judges to prosecutors?
In: Worral JL, Nugent-Borakove ME (eds) The changing role of the American prosecutor. State University
of New York Press, Albany
Fisher G (2003) Plea bargaining’s triumph: a history of
plea bargaining in America. Stanford University Press,
Palo Alto
Frase RS (2005) Sentencing guidelines in Minnesota,
1978–2003. Crime Justice: A Rev Res 32:131–219
Gazal-Ayal O, Tor A (2012) The innocence effect.
Duke Law J 62:339–401
Jehle JM, Wade M (2006) Coping with overloaded
criminal justice systems: the rise of prosecutorial
power across Europe. Springer, Berlin
King NJ, O’Neill M (2005) Appeal waivers and the future
of sentencing policy. Duke Law J 55:209–261
King NJ, Soule DA, Steen S, Weidner RR (2005)
When process affects punishment: differences in
sentences after guilty plea, bench trial, and jury trial
in five guidelines states. Columbia Law Rev
105:959–1009
Langbein JH (1979) Land without plea bargaining: how
the Germans do it. Mich Law Rev 78:204–225
Langer M (2004) From legal transplants to legal
translations: the globalization of plea bargaining and
the Americanization thesis in criminal procedure. Harvard Int Law J 45:1–64
Luna E, Wade M (2010) Prosecutors as judges. Washington Lee Law Rev 67:1413–1532
Maxfield LD, Kramer JH (1998) Substantial assistance: an
empirical yardstick gauging equity in current federal
policy and practice. United States Sentencing
Commission, Washington, DC
Miethe TD (1987) Charging and plea bargaining practices
under determinate sentencing: an investigation of the
hydraulic displacement of discretion. J Crim Law
Criminol 78:155–176
Miller ML, Wright RF (2008) The black box. Iowa Law
Rev 94:125–196
Nardulli PF, Flemming RB, Eisenstein J (1985) Criminal
courts and bureaucratic justice: concessions and
consensus in the guilty plea process. J Crim Law
Criminol 76:1103–1131
Pfaff JF (2011) The myths and realities of correctional
severity: evidence from the National Corrections
Reporting Program. Am Law Econ Rev 13:491–531
P
P
3544
Podgor ES (2012) Prosecution guidelines in the United
States. In: Luna E, Wade M (eds) The prosecutor in
transnational perspective. Oxford University Press,
New York
Schulhofer SJ (1984) Is plea bargaining inevitable? Harv
Law Rev 97:1037–1107
Scott RE, Stuntz WJ (1992) Plea bargaining as contract.
Yale Law J 101:1909–1968
Steffensmeier D, Ulmer JT, Kramer JH (1998) The interaction of race, gender, and age in criminal sentencing:
the punishment cost of being young, black and male.
Criminology 36:763–798
Wright RF (2005) Trial distortion and the end of
innocence in federal criminal justice. Univ PA Law
Rev 154:79–156
Wright RF (2009) How prosecutor elections fail us. Ohio
State J Crim Law 6:581–610
Wright RF, Miller ML (2002) The screening/bargaining
tradeoff. Stanford Law Rev 55:29–118
Zacharias FC (2001) The professional discipline of
prosecutors. North Carol Law Rev 79:725–743
Plural Policing
▶ Family Engagement Strategies to Reduce
Crime
▶ Legal Frameworks for Third-Party Policing
Poaching
▶ Crimes Against Animal Life
Police
▶ Privatization of Policing in an International
Context
▶ Police and the Excessive Use of Force
Police Accountability
▶ Control of Police Misconduct
Plural Policing
Police and the Excessive Use of Force
Robert J. Kane
Program in Criminal Justice, Drexel University,
Philadelphia, PA, USA
Synonyms
Brutality; Coercion; Force; Police
Overview
The excessive use of force by police is often
a difficult phenomenon to identify, as well as
a difficult concept to define. Unlike corruption
that is motivated by profit – which is generally
impermissible by its very nature – excessive use
of force often resides at the marginal end of the
acceptable use of force continuum. That is,
because society authorizes the police to use physical coercion to carry out their mandate, it is often
unclear when, along the use of force continuum,
the acceptable use of force becomes excessive.
Perhaps this is why excessive force, or brutality,
often seems to occur unchecked in communities
that are unable to advocate for themselves, making it difficult to hold police officers accountable
for the use of unauthorized violence.
Several classic police scholars have developed
occupational templates to help describe the circumstances under which some police officers
engage in brutality. These perspectives try to
account for police-suspect interactions, the police
working environment, and dangers therein; while
others locate the causes in the personalities of the
officers, or the agencies in which they work. By
understanding some of the causes of police brutality, society and policy-makers have a better
chance of developing prevention strategies that
can minimize the use of excessive force, particularly in the most vulnerable communities. Currently, these prevention strategies are rooted in
the areas of recruitment and training, as well as
the proper supervision of officers in the field.
Police and the Excessive Use of Force
Introduction
When two officers of the Los Angeles Police
Department were shown on tape beating the
motorist Rodney King with batons – after their
field supervisor had already exposed King, with
little effect, to the Taser, a conductive energy
device designed to “stun” suspects into submission – even while he was rolling on the ground
and seemingly defenseless – few observers were
left with any doubt that they had just witnessed an
example of police brutality. Some experts (e.g.,
Skolnick and Fyfe 1993; Toch 1996), however,
who viewed the entire tape – as opposed to simply its last 7 s – saw a more nuanced escalation of
force that ultimately became excessive. Rodney
King had broken a primary subcultural rule in
policing: When initially contacted by the team
of California Highway Patrol officers who
attempted to stop his vehicle for speeding on
Interstate 210 in the San Fernando Valley area
of Los Angeles, he failed to pull over for them.
Indeed, King would later report that he fled from
the CHP cruiser because he was driving while
intoxicated and feared that an arrest would violate the provisions of his parole (the result of an
earlier robbery conviction) (Skolnick and Fyfe
1993). For his failure to submit to police authority
in ways that satisfied the five LAPD officers at the
scene, King was “Tazed” once, struck with police
batons 56 times (several of which were blows to
the head), and kicked six times by officers
(Independent Commission 1991). To date, the
beating of Rodney King represents the archetype
of police brutality, and it encapsulates multiple
conceptual perspectives – from the levels of the
social interactionist, workgroup, and cultural –
aimed at explaining why some police officers
resort to excessive force during their encounters
with members of the public.
Excessive use of force by police is a tricky
phenomenon to identify and explain. Even in the
most “obvious” and infamous examples of police
brutality, it is frequently the case that at some
point during the violent encounter, some force on
the part of police was justified – perhaps even
necessary – to preserve officer and/or public
3545
P
safety. As such, excessive force by police is substantively distinct from other forms of misconduct, such as profit-motivated corruption
because, while the later is usually wrong on its
very face (e.g., police officers are not permitted to
extort money from drug dealers or motorists; and
they are not permitted to steal jewelry off corpses
they encounter during health-and-welfare
checks), they are allowed, and in many cases
even expected, to use coercive force as part of
their role in society. In fact, it as a virtual axiom
in policing that the use of force is a “tool,” whose
use should be regulated but not taken away.
Several leading scholars have developed
frameworks to explain why the excessive use of
force occurs and how it may foster in some police
departments. Van Maanen (1978) identifies the
“asshole” as one lens through which to view the
use of force encounter. He argues that when
a suspect commits an “affront” to a police officer’s authority, the officer – under certain circumstances – may respond with violence as a way of
(re)establishing control over the police-suspect
event. Muir (1977) takes a more individualistic
approach, observing that some police officers
have an “enforcer”-type personality borne out of
their inabilities to (1) empathize with people
whom they contact under the worst of circumstances, and (2) communicate in ways that may
de-escalate potentially violent encounters.
Skolnick and Fyfe (1993) view the excessive
use of force from an organizational-level, noting
that certain police departments develop and foster a so-called “siege mentality” in which officers
are made to feel outnumbered, outgunned, and
unwelcome in the communities they serve.
Indeed, the siege mentality – which stems from
the top levels of police department administration – views officers as a collective occupying
force deployed in dangerous communities
inhabited by dangerous people. As a result, officers in such locations are often made to believe
that the only way for them to maintain order is
through the use of highly aggressive enforcement
strategies that may translate into excessive force.
This entry examines the use of excessive force
by police through the perspectives offered above
P
P
3546
while also noting the difficulty and complexities
involved in defining police brutality. As alluded
to above, identifying excessive force is made allthe-more difficult by virtue of the use of physical
coercion that is a large part of how the police
fulfill their mandate in society. In addition to
discussing these conceptual issues, the entry
also describes how legal mechanisms, such as
judicial review and the US Code, have made the
excessive use of force difficult to prove at an
acceptable legal standard, which may inadvertently stifle efforts to redress abusive police practices and reform police departments in need of
reforming. The entry closes with a discussion of
how to prevent and control the excessive use of
force.
The Excessive Use of Force Prism
Before trying to define the excessive use of force
by police, the reader might first start with a more
fundamental question: Why do we have the police
in our society? It is common knowledge that at
times, some police officers use more force than is
necessary on suspects they encounter; some
engage in profit-motivated misconduct and then
lie on the witness stand to protect their corrupt
enterprises; and these practices often take place
in the most socially and economically vulnerable
communities (Kane 2002). So, again, while most
people understand that some police officers, as
well as some departments, have the propensity to
engage in brutality, the question is again begged:
Why have the police in the first place?
American citizens enjoy – perhaps more than
those of other developed nations – almost unfettered access to their court system. They have the
ability to report crimes directly to their local
District Attorney’s Offices, and they have the
right to request protective orders and other
forms of injunctive relief directly from magistrates in open court. In short, American citizens
under the best of circumstances have the ability to
mobilize government on their behalf in ways that
could theoretically bypass the police entirely, and
yet, most Americans – even those who have had
bad experiences with them – would scarcely
Police and the Excessive Use of Force
advocate dissolving the police as a public
institution. Why? As cited by Klockars (1985),
we maintain a collective police force
because of what Egon Bittner wrote in 1970:
“something-ought-not-to-be-happening-aboutwhich-something-ought-to-be-done. . .NOW!!”
That is, although it is possible to mobilize the
government without using the police, it is often
not feasible to do so. In highlighting this reality, Klockars (1985) notes that if a crazy resident goes into the yard of his neighbor to chop
down a prized apple tree with an axe, seeking
an injunction against such an intrusion likely
will not prevent the destruction of the apple
tree. In such events, therefore, the point of the
police is to stop – or at the very least – freeze
a moment in time in order for the courts to gain
the opportunity to properly adjudicate the case.
The police accomplish this mandate by exercising their “general right to use coercive force”
(Klockars 1985) – a right bestowed upon them
by the American public through the Executive
Branch of government. And it is this general
right to use coercive force that makes possible
both the best and worst practices of policing. At
their best, the police use their coercive authority to improve the life chances of people they
contact – often under the most abject social and
economic conditions imaginable. At their
worst, police use their coercive authority in
ways that abuse, denigrate, and otherwise
“tread on the human dignity” (Carter 1985,
p. 322) of people they encounter.
Although the American public grants police
the general right to employ coercive force, it does
so with few guidelines as to the boundaries of its
proper use, making it difficult for evaluators of
police behaviors to determine – for example –
when appropriate force may have become excessive, or when excessive force was used when no
force was necessary. The following serve as
instructive examples:
• Incident 1: Two police officers on foot see
a man on a public sidewalk who matches the
description of a robbery suspect. They attempt
to contact the man, but as they summon him
their way, the man turns and runs. Both officers engage the suspect in chase. The man
Police and the Excessive Use of Force
leads them on a 5-min foot pursuit through
alleys, over fences, and across a vacant lot
until they finally catch him in an open field.
At the point of contact, the man physically
resists the officers’ efforts to subdue and handcuff him. He swings and kicks at them, making
glancing contact, until one of the officers
makes two baton strikes to the man’s body,
sending the suspect hard to the ground. One of
the officers is bleeding slightly from a scratch
on his face, the other suffered a hyperextended knee. As the suspect writhes on the
ground from the pain of the baton strikes, one
of the officers kneels to handcuff him, while
the other officer kicks the man twice in the
ribs. Even after the suspect is securely cuffed
and clearly incapacitated, the other officer
punches him with a closed fist in the kidney
two times just after the other officer pulled him
from the ground.
• Incident 2: A police officer working traffic
duty sits in his marked cruiser at mid-day.
The car is parked in a somewhat concealed
position relative to the corner of an intersection governed by a four-way stop. When he
observes a pickup truck roll confidently
through the intersection without stopping, the
officer ignites his engine, activates the light
bar, and pulls quickly away from the curb to
initiate a vehicle stop on the truck. After a few
hundred yards, and two siren “chirps,” the
driver of the pickup truck finally pulls to
the curb. The officer exits his vehicle, walks
to the truck, and instructs the driver to put
down his window. The officer observes that
the driver and his one passenger are teenage
boys, most likely between the ages of 16 and
18 years. The officer waits with expectancy
while the driver – who is wearing a smirk on
his face – finally begins to slowly roll down his
window. The officer, losing patience, asks for
the driver’s operating license. After a pause,
the driver says, “Why do you need that? I am
just taking my buddy home.” The officer again
requests the young man’s license, and
although the driver complies, he does so
slowly. Suddenly, the officer reaches in
through the window, grabs the young man by
3547
P
the hair, and swiftly pulls him through the
window and out onto the street. The officer
then drags the young man – still by the hair –
around the front of the truck to the sidewalk,
where he slams the young man’s head against
the hood of the truck. The young man screams
in surprise and pain, while the officer repeats
the motion of bashing his head against the
hood. The officer then drags the young man
back around the front of the truck, opens the
driver’s side door, and roughly deposits the
young man back into the truck, slamming the
door closed. The officer then advises, “The
next time a police officer asks for your driver’s
license, I suggest you give it to him the first
time.” Then officer then turns, walks back to
his patrol car, enters, and drives away.
The above scenarios indicate two types of
situations during which excessive force may be
used by police, and both have their own definitional difficulties associated with them. In the
case of incident #1, it is clear that the two officers
involved in the foot chase were justified – even
compelled – to use force in the apprehension and
control of the suspect. The suspect, in fact, used
physical force to prevent his apprehension,
swinging, kicking, and striking the officers as
they attempted to subdue him. The officers were
probably justified in the use of their batons while
the suspect was fighting them, but what about the
subsequent kicks and blows once the suspect was
handcuffed and effectively restrained? At what
point did the appropriate use of force by the
police become excessive? Different people –
e.g., police officers, police administrators, community residents, and promoters of law and
order – may have different answers to this question. For some, the baton strikes may have been
regarded as excessive, considering there were
two officers and the suspect was not at all armed
with a weapon. To others, the kicks and punches
delivered by officers at the conclusion of the
encounter were not excessive, but rather “just
desserts” visited upon a “victim” who should
have cooperated with the police in the first place.
The events described in incident #2 suggest
perhaps a less ambiguous evaluation of the use of
excessive force. Although the driver of the
P
P
3548
pickup truck was not physically combative with
the officer who stopped him, he was defiant, and
he failed to comply with the lawful order of
police officer, namely, to produce his operator’s
license. In doing so, the driver of the pickup truck
left the officer with two categories of choice:
(1) make a tactical retreat, perhaps to his vehicle
where he could have radioed for a supervisor to
join him at the scene in an effort to avoid the use
of force; and (2) increase the dosage of coercion
in order to secure the driver’s compliance. From
the description of events in incident #2, the officer clearly chose the latter by using physical
coercion to neutralize the young man’s defiance.
And again, commentators of police behaviors
may disagree as to the appropriateness of the
officer’s actions. Some would conclude that the
force used by the officer was unnecessary, excessive, and perhaps even illegal; others may
conclude that the officer rightly punished
a smart-aleck kid who needed to be taught to
respect authority.
Despite potential differences in assessments of
the excessive use of force across incidents 1 and
2, the hypothetical events share a common link:
From the perspective of responding officers, both
victims may have “deserved” it. That is, in both
cases described above, an argument can be made
that
both
suspects-turned-victim
forced
a confrontation with the police, which led to the
excessive force. This is another feature of many
use of force incidents that makes identifying brutality or excessive use of force difficult. As
Walker (2010) has argued in his criticism of
drug-war policing, when police officers use
excessive force during their encounters with the
public, they usually do so on socially marginalized and/or otherwise “undesirable” suspects,
often rendering it unlikely that the victim of brutality can successfully hold officers accountable
for their actions. Thus, despite the procedural and
legal constraints (discussed in the following section) placed on officers’ authority on the use of
physical coercion, the sociology of use of force
encounters often challenges society’s definition
of the excessive use of force, particularly when
the force may have been used on someone who at
least initially defied the police.
Police and the Excessive Use of Force
The following section discusses several of the
legal issues associated with defining the excessive use of force, which adds further difficulty of
successfully defining excessive force and
brutality.
Excessive Force Through Legal and
Social Lenses
Moving more thoroughly into an evaluation of
the excessive use of force by police requires some
definitional clarity. In general, police use of force
can be categorized into four types: (1) appropriate, (2) deadly, (3) excessive, and (4) unnecessary. Once these types of force are defined, the
entry focuses on excessive force, also commonly
referred to as police brutality (e.g., Skolnick and
Fyfe 1993).
The appropriate use of force occurs when
police officers use physical coercion that is proportional to a suspect’s resistance to gain compliance and/or diffuse a violent or potentially
violent encounter. Though perhaps commonly
believed, police officers are not required to
match a suspect’s resistance incrementally with
force that follows a use of force continuum
(Graham v. Connor 1989). Rather, police officers
are legally (and usually administratively) permitted to “skip” steps on the use of force continuum
in order to overwhelm a suspect – within reason –
to achieve custody. Society and the courts do not
require police officers to engage in fist fights with
suspects in order to gain compliance. Indeed,
most police officers are trained to control many
suspects – whether armed or not – from a safe
distance, often via the threat or use of chemical
spray, a conductive energy device (e.g., Taser), or
a firearm. To provide a framework for assessing
the appropriateness of the use of force, the
Supreme Court applied an “objective reasonableness” standard (via the Search and Seizure Clause
of the Fourth Amendment) in the case Graham v.
Connor (1989), holding that the legality of
the use of force must be evaluated based on the
judgment of a “reasonable” officer at the scene of
the event rather than through a “retrospective”
lens. The Court reasoned that when police
Police and the Excessive Use of Force
officers become involved in violent or potentially
violent encounters, they frequently have just seconds to decide how to apply coercion, often in
self-defense and/or in the defense of others. As
such, the fairest assessment of the legality (and
administrative permissibility) of the application
of force must be made from the perspective of
someone at the incident who would have had to
make a force decision in the heat of the moment –
just as the actual officer did.
Deadly force – i.e., “force that kills or is likely
to kill” (Fyfe 1979) – and which is usually, but
not always, the result of a firearm discharge – is
governed largely by the same objective reasonableness standard as nonlethal force (the standard
appears to change to a test of “deliberate indifference,” as governed largely by the Eighth
Amendment, if deadly force is used on a suspect
while in a custody facility), though with an extra
restriction. In the case Tennessee v. Garner
(1985), the Supreme Court abolished the common law custom that had historically allowed
police officers to use deadly force against fleeing
suspects who posed no threat to life. In abolishing
this practice – which was known as the “fleeing
felon rule” – the Court reasoned that using deadly
force to apprehend a person suspected of having
committed a felony, but who seemed to pose no
imminent threat to life – constituted an unreasonable seizure of the person under the Fourth
Amendment. Despite this landmark decision
that made the “defense of life” standard the law
of the land, the Court in Garner failed to outline
circumstances under which deadly force was permissible. Thus, Garner allows officers to determine the meaning of “life” threatening: The risk
to life need not be real at the time of the incident;
it must be apparent.
Among all the dimensions of the use of force
construct, “unnecessary” force is perhaps the
most difficult to assess and the most controversial
to condemn. Unnecessary force usually represents the proper use of force at the immediate
moment of application, but which could have
been avoided had officers not allowed, or caused,
the incident to escalate (Fyfe 1986). Thus, rather
than representing the product of malice or the
inclination to “punish,” unnecessary force
3549
P
generally results from police officer’s professional incompetence. A typical example of this
occurs when police respond to a man with a gun
call. They might arrive on the scene to find an
agitated man pacing in his front yard with a pistol
in hand. Because officers neglected to find cover
and concealment (i.e., failing to position themselves behind parked cars or out of direct line of
fire), they shoot the man when he makes a furtive
movement. When focusing on the last point in the
encounter – i.e., the instance immediately preceding the shooting – the force used may be justifiable since the police were compelled to protect
their lives and perhaps the lives of bystanders.
When considering, however, the events that led to
the shooting, professional incompetence left the
officers at risk of being shot in the first place. Had
they found proper cover, they might have been
able to wait and persuade the man to drop the gun,
which would have averted the use of violence.
Although the examination of unnecessary force is
largely beyond the scope of this entry, it is important to at least identify because (1) it represents
a component of the overall use of force construct,
and (2) likely happens more frequently than does
excessive force, often without the subsequent
application
of
accountability
remedies
(Fyfe 1986).
As the above discussion should suggest, police
brutality is physical force that breaches the confines of legally permissible physical coercion.
Because police brutality is often committed out
of malice, disrespect, and/or officers’ desires
to reassert their authority over a suspect
(e.g., Skolnick and Fyfe 1993), several academic
researchers often regard excessive force as an
example of police abuse of authority (e.g., Carter
1985) rather than a mere “excess” of force. This is
because police brutality often stems from the
same theoretical processes that lead to verbal
chastising and psychological abuse during police
interrogations (e.g., Carter 1985) – both of which
are abusive but not physical.
Although most state criminal codes in the
United States include statutes that specifically
make police brutality under “color” of state
authority illegal, the most common legal remedy
for parties attempting to redress the alleged
P
P
3550
excessive use of force by police is Title 42,
Section 1983 of the US Code. This is the Federal
civil rights statute that allows persons alleging
harm by police (and any representative of the
State who may have acted under color of authority) to seek damages from both the individual
officer(s) and the government agency that
employed them. The reason why Section 1983
is likely the most common legal mechanism
invoked in excessive use of force claims is
because the statute relaxes the usual requirement
placed on plaintiffs that they exhaust all lower
administrative and legal outlets before filing
a claim in Federal court. Although the statute
does not require proof that a police officer
intended to harm the victim (that is, the plaintiff
need not prove an officer’s psychological state of
mind), it does require the plaintiff to successfully
demonstrate that the government entity (e.g.,
municipal police department) that employed the
officer maintained official policies and directives
that directly resulted in the harm caused, or that
the entity had established informal customs and
practices that led to the harm. For this reason,
Section 1983 lawsuits are often colloquially
referred to as “patterns and practices” suits and
generally allege that the brutality (or other problematic behaviors) occurred because the police
agency failed to train and/or supervise the officers
properly and/or negligently retained the officers
as employees.
Thus, integrating the Section 1983 statute with
Supreme Court decisions results in plaintiffs
gaining ready access to Federal court for the
purposes of attempting to hold police officers
and the agencies/governments that employed
them accountable for brutality, though trial courts
must adhere to Graham’s “objective reasonableness” requirement of evaluating the use of force
from the standpoint of a reasonable officer at the
scene. The net effect has been that, while it is
relatively easy to sue police officers and police
departments, it is highly difficult to win such
cases due to the “reasonable officer at the
scene” requirement (Skolnick and Fyfe 1993).
Moreover, as research has shown, even when
plaintiffs are generally successful in proving
police brutality at an acceptable legal threshold,
Police and the Excessive Use of Force
juries historically have been reluctant to decide
against police officers due to generalized public
support for the police, and the fact than many
plaintiffs who allege brutality occupy dubious
social positions relative to the officers they are
accusing. Perhaps the best example in modern
history was the Rodney King case in which all
officers accused of brutality were found not guilty
in the first trial.
Classic Templates to Explain Excessive
Use of Force
As noted in the opening pages of this entry, police
brutality – or excessive use of force – occurs for
myriad reasons across multiple levels of analysis:
the psychological, sociological, social contextual, and organizational. Since William Westley
(1970) published his classic ethnography of
police in Indiana (which was actually published
20 years after his dissertation work that produced
the findings) in which he described for the first
time the importance officers and police organizations placed on making felony arrests, several
researchers have replicated Westley’s methodology in an effort to highlight several aspects of the
police occupation that previously had been generally unknown. The result was the publication of
classic statements on police coercion that
described the processes by which police officers
may commit, and why police departments sometimes attempt to hide, police brutality. This entry
considers these perspectives as occupational templates and describes the most important of these
below.
Starting at the individual-level, in his largely
ethnographic study of the police in “Laconia” (a
pseudonym), Muir (1977) interviewed and then
observed a group of police officers in an effort to
understand how they responded to “critical” incidents in their work settings. It is now commonly
known that “Laconia” was Oakland, CA, a large
city that employed 800 police officers at the time
of Muir’s research, and which was characterized
by high levels of racially concentrated economic
resource deprivation and crime. During his
research, Muir observed four police officer
Police and the Excessive Use of Force
personality “types” that were differentiated by
two elements: the abilities to (1) empathize with
those whom they encountered on the street, and
(2) morally reconcile need to use force in certain
circumstances. For example, the archetypal officer in Muir’s typology was the “professional” –
i.e., a collective group of officers who could place
themselves psychologically in the life circumstances of the suspects they encountered, and
although professionals did not care to use force
to resolve most conflicts during their engagements with members of the public, they used
force when necessary to stop conflict, protect
lives, and restore order. Once the situational exigencies were neutralized, professionals ceased
their use of physical force. A key attribute of
professionals was their ability to communicate
effectively with suspects in ways that often minimized, de-escalated, or even eliminated the need
to use force during many incidents.
In contrast to the professionals were the
“enforcers:” Officers who had the highest propensity to use excessive force against suspects
because they (1) generally lacked the capacity
to fathom the often “tragic” conditions that characterized the lives of many people they encountered, and (2) were quick to resort to violence
when they perceived minimal threats to their
safety or when they felt compelled to assert
police authority. Muir (1977) noted that, unlike
the professionals, enforcers usually lacked the
communication skills necessary to de-escalate
and/or otherwise avoid the need to use force in
certain circumstances. As a result, enforcers often
created situations where other officers near them
were forced to use violence against suspects
because of the degree to which enforcers provoked violence during their contacts with the
public. Thus, and as Muir notes, enforcers were
problematic to police workgroups partly because
they frequently engaged in excessive force, and
partly because they often caused potentially violent encounters to escalate into actual violent
encounters.
Perhaps the most interesting and important
element of Muir’s (1977) typology was the recognition that officers might slide between one
personality type and another based on cues in
3551
P
the social environment. For example, officers
regarded as professionals – and who usually
conducted themselves appropriately during most
police-citizen encounters – sometimes became
enforcers during certain incidents. Muir, for
example, identified one “professional” officer
among those he studied who was highly communicative and restrained in virtually all street
encounters except when he responded to sexual
or domestic assaults against female victims.
These events tended to trigger an uncharacteristic
response in the officer – perhaps due to family
history or other personal experience – leading
him to take on the attributes of the enforcer: He
was quick to use violence as an initial response
when contacting an alleged domestic or sexual
assault offender. Thus, although Muir’s typology
was largely psychological, it recognized the
importance of social context when analyzing
street-level behavior of officers.
Although no identified studies have found reliable empirical support for Muir’s typology (see,
for example, Hochstedler (1981)), none has
tested the typology in a setting that replicates
that in which Muir conducted his research. It
may be that Muir’s findings hold primarily for
postindustrial port cities with large urban
African-American populations that live predominantly in poverty conditions, socially isolated
from the larger urban matrix. Such are the conditions that most often allow officers to engage in
abuse of authority – including excessive force –
with little fear of recourse (e.g., Kane 2002;
Kubrin and Weitzer 2003). Moreover, despite
scant support for the typology itself, at least one
study has found that – like Muir’s findings –
officers who exhibit certain personality characteristics (diffuse though they may be) often
“slide” between profiles as the social contexts in
which they work vary (Hochstedler 1981). This
finding supports Muir’s observation that even socalled professional police officers may become
“enforcers” under the circumstances that elicit
particularly strong responses, and perhaps where
local social constraints are wanting. Such findings demonstrate the importance of the social
ecological contexts in which officers work as
determinants of their behaviors. At the very
P
P
3552
least, Muir’s typology provides a useful taxonomy for efficiently describing certain officer
characteristics that are easily recognized – i.e.,
enforcer.
Residing one level up from Muir (1977) on the
social ecological progression is Van Maanen’s
(1978) “asshole” perspective. This framework is
so-named for the colloquial usage of the term
by police officers as a shorthand for suspects
who – through speech or gesture – “deny”
a police officer’s “definition” of the power structure that characterizes their interactions with the
public – namely, that officers are in charge. Van
Maanen argued that the line officers he studied
tended to classify members of the public into
three groups: “suspicious persons” (suspected
criminal offenders), “assholes” (people who challenge police authority), and “know nothings”
(people who do not fit the previous two categories
and who are also not police officers). Whereas
police officers usually have little or no cause to
use coercion against know nothings, and they
generally use coercion only as necessary to control and/or gain the compliance of suspicious
persons, they may use coercion – i.e., force –
against those labeled assholes as a retaliatory
mechanism to redress an affront to an officer’s
authority (Van Maanen 1978).
Van Maanen (1978) observed that a suspect’s
initial response to a police officer’s inquiry determined the officer’s categorization of him (Van
Maanen never clarifies as to whether the template
applies to female suspects), as well as the subsequent police reaction. Suspects who appeared to
reject the officer’s definition of the situation –
thereby challenging the officer’s authority and
legitimacy – were labeled assholes and were
often subjected to a violent response as a form
of punishment or reassertion of authority. From
an officer workgroup standpoint, police officers
who labeled a suspect an asshole were justified,
and perhaps even expected, to use violence to
reestablish control.
Recall Incident #2 – the dramatized use of
force scenario – in the “Excessive Force Through
Legal and Social Lenses” section of this entry.
The officer’s response to the youthful driver’s
verbal resistance was a quintessential example
Police and the Excessive Use of Force
of how a person may become labeled an asshole
under Van Maanen’s model. Importantly – and as
shown in Incident #2 – when officers label
a suspect an asshole, and to the extent that the
officers use brutality to redress the affront, they
often do so to impose police authority typically at
the expense of formal sanctioning (e.g., arrest).
As such, some suspects – particularly those
engaged in illegal behavior when contacted by
the officer – may accept the brutality in lieu of
a formal sanction because if they reported the
brutality, they would also have to report their
activities that led to their initial contact with the
police.
Though it is the case that occupational
templates-perspectives often fail to yield empirical support, at least one recent study has offered
some empirical validation of Van Maanen’s
(1978) “asshole” thesis. In a study of how the
occupational goals of police may sometimes conflict with rule of law practices, Kane and Cronin
(2011) examined the extent to which police officers used force in situations that may have been
characterized by the “asshole” process. Through
the use of discriminant function analysis, Kane
and Cronin reported that during coercive
exchanges with suspects (i.e., situations that
ultimately ended in arrest), suspects who demonstrated “verbal antagonism” toward officers – but
not physical resistance – were likely to be
“punched” and “kicked” during the incident.
Interestingly, when suspects demonstrated physical resistance, but not necessarily verbal antagonism, they were usually subjected to more
“normal” types of force, such as chemical sprays,
baton strikes, or basic joint locks (Kane and Cronin 2011). Verbal antagonism appeared to elicit
“asshole” responses in the form of punches and
kicks – use of force techniques that usually do not
exist on a police use of force continuum.
Whereas Van Maanen (1978) operationalized
the “asshole” perspective as one rooted primarily
in the social interactionist tradition, there is evidence that the asshole process also may be driven
in part by police subcultural pressures. In his
observational study of how the Los Angeles
Police Department controls space within its vast
jurisdictional boundaries, Herbert (1998, p. 347)
Police and the Excessive Use of Force
adds a conceptual dimension to Van Maanen
(1978) with his identification of normative
orders: a collective subcultural belief system
“. . .oriented around a common value.” Herbert
argues that police officers working in small
groups develop informal work and productivity
standards within the context of six normative
orders that include law, bureaucratic control,
adventure/machismo, safety, competence, and
morality. Herbert (1998, p. 347) observes that
officer workgroups place differential values on
the normative orders in ways that “provide guidelines and justifications for” workgroup enforcement activities. For present purposes, Herbert
notes that normative orders – such as law and
safety or law and competence – may conflict
with one another as officers use their coercive
authority to satisfy one normative order (e.g.,
competence), perhaps at the expense of another
(e.g., law or bureaucratic control).
Herbert’s normative order system, which he
referred to a police subculture, may be integrate
with occupational templates because it explains
how situational circumstances may trigger police
responses designed to satisfy occupational goals.
For example, officers confronted by a suspect
they label as “dangerous” will likely invoke the
normative order of safety to minimize the threat
of violence. Similarly, officers who enter into
a coercive encounter with a bellicose suspect –
an “asshole” in Van Maanan’s (1978) taxonomy –
may seek to demonstrate their competence to
other officers in their workgroup by using excessive force or brutality, which may be regarded as
situationally justified to reassert police authority
over a noncompliant suspect. Again, such an
application of coercion may conflict with rule of
law orders, but may be regarded as occupationally necessary if officers achieve the objectives of
their primary normative value.
Finally, to understand how police agencies
themselves may create a culture that fosters, if
not encourages, the excessive use of force by
their officers, it is important to consider the
research on police organizations. In Los Angeles,
the Christopher Commission assigned to investigate patterns of excessive violence in the LAPD
in the wake of the 1991 beating of motorist
3553
P
Rodney King identified a “siege mentality” that
was prevalent among officers of that department
(Independent Commission 1991, p. 95). This
siege mentality, the Commission argued, led officers to view their relationship with the community at large as adversarial (the “us vs. them”
perspective) and appeared to legitimize the
excessive use of force by police against suspects
in an environment that all but guaranteed immunity from accountability (Skolnick and Fyfe
1993). The confidence that officers placed in
their ability to freely “punish” suspects who committed affronts against police authority was perhaps best illustrated by the routine exchanges
among officers via their mobile data terminals
in which anecdotes describing violent activities –
often in racist terminology – were openly
exchanged (see: Independent Commission
(1991)).
Appealing to occupational templates to help
explain the processes by which excessive force
may be used and legitimized by police officers is
instructive because they offer a social ecological
context within which to interpret police brutality
that is often absent from the empirical literature
that examines excessive force. Templates, such
as Muir’s (1977) extortionate transaction model,
Van Maanen’s (1978) “asshole” perspective,
Herbert’s (1997) identification of workgroup
“normative orders,” and Skolnick and Fyfe’s
(1993) elaboration of the “siege” mentality,
show that police brutality may be influenced at
every level of a police officer’s contextual occupational life. Moreover, they offer insight into
how some officers who may be disinclined to
use brutality on suspects may operate in layered
contexts that favor, or even expect, them to use
excessive against certain suspects under certain
circumstances.
Some Consequences of Police Brutality
Perhaps the most obvious consequences of police
brutality are the direct effects endured by victims.
Although there are no national data on the subject, it seems clear that people who experience
police brutality can suffer physical harms ranging
P
P
3554
from minor aches and pains to serious injury, and
in some cases, death. It is also likely that people
who experience physical abuse at the hands of
police may suffer emotional trauma. In the
United States, however, police agencies are not
required to report use of force – let alone excessive force – incidents to any centralized authority,
leading researchers and policy-makers to make
guesses about the frequency and severity of
police brutality. This problem could be ameliorated by the implementation of a national
reporting system on the police use of force.
In addition to the direct effects of police brutality to those who experience such abuse, there
are often broader sociological consequences that
may result when police engage in the excessive
use of force, particularly when the force is systemic in nature and occurs in communities whose
residents perceive that they have been socially
and legally marginalized by their municipal governments. Among these, perhaps the gravest consequence is the loss of legitimacy police forces
can experience when officers are viewed as brutal
or abusive, and when community residents perceive little or no recourse to control them.
Tyler (1990) has argued that societies are
more likely than not to comply with the law
when they believe in the morality of the law and
in the legitimacy of the government making the
law. For a society to accept law as moral and the
government as legitimate, its members must
believe in the fairness of the law-making process
(Tyler 1990). This model of procedural justice
generally argues that the way a state achieves
and maintains legitimacy is by distributing its
authority through consensus-based processes
without regard for the social and economic statuses of societal subgroups. If the public becomes
polarized over issues related to the distribution of
legal authority, the government may lose legitimacy and find it difficult to fulfill its “regulatory
role” in society (Sunshine and Tyler 2003, p. 515)
without having to resort to the frequent use of
physical coercion. This issue is important for
municipal authorities, as some members of the
public, particularly those residing in communities characterized by racial segregation and/or
systemic economic resource deprivation, may
Police and the Excessive Use of Force
perceive inequities in the distribution of police
resources in violation of procedural justice
expectations (Jacob 1971).
Kubrin and Weitzer (2003) were among the
first researchers to find social scientific evidence
supporting the above arguments. In their study of
retaliatory homicide in St. Louis, Kubrin and
Weitzer found that certain types of homicides
predicted subsequent homicides in the form of
retaliation, suggesting that, had the initial homicide been stopped, the net reduction in many
cases would have been two, given that the retaliatory homicide also would have been negated. In
the qualitative portion of their study, Kubrin and
Weitzer found through unstructured interviews
that the initial homicides were allowed to occur
in large part because community residents
refused to share information related to criminal
activity with the police due to their distrust of law
enforcement authorities. This distrust stemmed
from perceived abuses – including brutality – of
community residents by police over a number
years. Kubrin and Weitzer’s (2003) findings
were largely consistent with Anderson’s
(1999) observations of the so-called Code of the
Street. Anderson argued that the Code – in which
respect was valued as a social good, earned and
maintained by the threat and/or use of violence –
prevailed in the north Philadelphia community he
studied in part because the police had virtually no
legitimacy due to their perceived abuses of
authority.
Kane (2005) made similar findings in New
York City in his examination of the consequences
of “over-policing” and police misconduct
(including violence) on violent crime. For present
purposes, Kane found that in economically stable
communities, police misconduct had no effect on
violent crime. In economically disadvantaged,
and “extremely” economically disadvantaged,
communities, police misconduct (including violence) led to subsequent increases in violent
crime. Kane attributed this relationship to the
same processes driving Kurbin and Weitzer’s
(2003) findings: As residents of socially/economically marginalized communities perceived
increases in police abuses, they likely ceased
sharing
crime-related
intelligence
with
Police and the Excessive Use of Force
authorities, allowing violent crimes to occur that
may have been otherwise prevented.
In his classic lecture, “Politics as a Vocation,”
Max Weber (see: Weber et al. 2004, pp. 78–79)
argued that because the State “claims the monopoly on the legitimate use of force” to enforce its
laws, politicians who hope to gain the “obedience” of those whom they govern need not simply
do the right thing; politicians must appear to do
the right thing. Weber’s arguments are made
salient by Kubrin and Weitzer (2003) and Kane
(2005): Failure of the police to “govern” effectively the territories they serve may lead to
noncompliance on the part of people who reside
in communities characterized by the perceptions
of police abuse of authority. In the “everyday”
sense, citizen’s noncompliance may manifest in
the form of people failing to readily yield to
police authority during traffic stops and stopand-frisk events – both of which may create
a context of hostility and distrust on the part of
the police and members of the public. In the
broader and more extreme sense, generations of
police abuse, including unchecked or redressed
brutality, may lead to urban violence. Indeed, two
of the most destructive urban riots in American
history have been attributed to long-standing patterns of police abuses, including unchecked
brutality.
On August 11, 1965, a white California Highway Patrol officer stopped an African-American
motorist in the Watts section of Los Angeles for
suspected Driving While Intoxicated. Even at
that time, Watts was an area of Los Angeles,
characterized by racially concentrated economic
resource deprivation, which had a difficult history with the police. The motorist – Marquette
Frye – resisted being taken into custody, which
led to an escalating confrontation with police
(Barnhill 2010). As more officers arrived on
scene, the crowd of bystanders also grew until
someone finally began throwing bottles and other
objects at the officers, who continued their struggle to subdue Frye (Barnhill 2010). In the process
of taking Frye into custody, officers also arrested
his mother and brother for unlawfully interfering
with the arrest, which led the crowd to switch
from throwing bottles to throwing rocks and
3555
P
other large items at police (Barnhill 2010). This
inciting incident led to 6 days of rioting in Watts,
leading to 34 deaths, 1,032 injuries, and over
3,000 arrests (Barnhill 2010). Although post-riot
analyses generally concluded that poor living
conditions and the social marginalization of
African-Americans in Los Angles represented
the root causes of the Watts riot (e.g., California
Governor’s Commission on the Los Angeles
Riots 1965), some researchers concluded that
years of police abuse and brutality in Los
Angeles – particularly during the tenure of Police
Chief William H. Parker, during which the LAPD
transformed into highly aggressive paramilitary
police force – was largely responsible for the riot
(Skolnick and Fyfe). Indeed, the McCone Commission, empanelled just a few months after the
riots and charged with identifying the causes,
noted from its many interviews of AfricanAmerican Watts residents:
“Police brutality” has been the recurring charge.
One witness after another has recounted instances
in which, in their opinion, the police have used
excessive force or have been disrespectful and
abusive in their language or manner (Governor’s
Commission 1965, p. 40)
The social and economic costs of the 1965 Los
Angeles riot were eclipsed only by those of the
1992 Los Angeles riot, which occurred in the
immediate aftermath of two white police officers
being acquitted for beating motorist Rodney
King. During the 1992 riot, the LAPD and the
California National Guard killed 53 people. More
than 2,000 people were injured, and over 1,000
buildings were destroyed by fire (Wood 2002).
The fact that both riots occurred in Los Angeles
was no surprise to Jerome Skolnick and James
J. Fyfe – two of the nation’s leading scholars of
police authority and accountability – who argued
that since William Parker was appointment
police chief of Los Angeles in 1950, the LAPD
became increasingly insular, aggressive, and
unaccountable to the mayor and city council
(Skolnick and Fyfe 1993). The unaccountability
was particularly pronounced under Chief Daryl
Gates, who enjoyed civil service protection and
indemnity from lawsuits (Skolnick and Fyfe
1993).
P
P
3556
It seems clear that, while people subjected to
direct forms of police brutality suffer the most
immediate harm from such abuse, those residing
in communities – even cities – in which the police
engage in the excessive use of force with little
accountability also experience the resulting
harms. Violent incidents in Los Angeles (1965
and 1992), Detroit (1967), Chicago (1968),
Liberty City-Miami (1980), Cincinnati (2001) –
even Tulsa in 1921 – were not so much riots as
they were urban rebellions. In each event, the
triggering incident was related to some form of
police abuse, and all post-riot analyses identified
abusive police practices, largely in communities
characterized by racially concentrated economic
resource deprivation and social marginalization,
as major causal factors (see: Independent
Commission 1991; Kerner Commission 1968;
Skolnick and Fyfe 1993).
Preventing Excessive Use of Force
As previously alluded to, people who experience
police brutality or the excessive use of force have
recourse through the conventional mechanisms
of police accountability: the citizen complaint
process, mediation, and the courts. Each of
these avenues of recourse has varying chances
of success, depending on the several factors
related to the event, including the social desirability of the accuser, the degree to which the
incident occurred in the public realm and/or in
view of credible witnesses, the extent to which
the accuser may have attempted to initially evade
or who otherwise resisted police officers, and in
many cases the quality of the accuser’s legal
counsel. While individual plaintiffs may prevail
in their attempts to hold police accountable for
the use of excessive force, most would probably
argue that best recourse would be the prevention
of police brutality before it occurs. The most
immediate methods by which to control and/or
prevent the excessive use of force by police are
probably rooted in the recruitment of police officers, training and supervision, and the establishment of organizational cultures within police
Police and the Excessive Use of Force
departments that value the protection of life over
the enforcement of law as a primary function.
The personnel processes police departments
use to recruit and hire officers play a crucial role
in shaping subsequent organizational behavior,
but in ways that are perhaps more subtle than
they are apparent. Police department hiring practices are generally designed to screen out undesirable applicants, which is substantively
different from trying to screen in desirable applicants. These so-called screening-out practices are
rarely evidence-based; and they are often not
followed consistently over time or across police
recruiters or background officers. Moreover, to
the extent that the excessive use of force is
a relatively rare event within police organizations, it is difficult to accurately predict who
might have become “problem” officers over the
course of their careers were they allowed to join
the force. Although emerging evidence supports
the practices of screening out applicants on the
basis of employment and criminal histories to
prevent general misconduct (see: Kane and
White 2012), there is virtually no evidence that
“violence-prone” (Toch 1996) police officers can
be identified and effectively screened out through
the normal recruiting process.
Screening methods that favor recruits whose
education, experience, cultural identities, and
world views would create police department cultures not dominated by a single gender or ethnic
group, and it would raise the possibility that
members of all subgroups would contribute to
the development of a sensitive workforce tolerant
of people across race and class. Empirical findings support this claim: In their study of careerending misconduct in the New York City Police
Department, Kane and White (2012) found that
from 1975 to 1998, as the NYPD became increasingly racially and ethnically diverse, it also
became better behaved. Rates of misconduct
decreased as diversity increased, independent of
external factors that often influence organizational rates of misconduct. Although the NYPD
over that time period may not have systematically
recruited persons with diverse characteristics or
personal histories, the results of bringing diverse
Police and the Excessive Use of Force
pools of recruits into policing are clear: Organizational misconduct declines.
In addition to recruitment, supervision of
police officers is also critical in preventing excessive use of force – particularly in communities
that historically have had difficult access to the
conventional mechanisms of police accountability. For example, in a study of police misconduct
in New York City police precincts, Kane (2002)
found that the traditional antecedents of social
disorganization – particularly, structural disadvantage – predicted increased rates of careerending police misconduct. Precincts characterized by the highest rates of racially concentrated
structural disadvantage also had the highest rates
of police misconduct. This finding led Kane
(2002, p. 891) to conclude that, “The very communities likely in need of the most protection by
the police. . .(are also) in need of the greatest protection from the police. . .” One way to offer that
protection is through the supervision of officers.
In many American police departments, the
average staffing ratios call for one sergeant for
every nine patrol officers (Walker and Katz
2010), and while scant research highlights what
supervisors actually do, at least one study suggests that low spans-of-control (i.e., high ratio of
supervisors to officers) decreased excessive force
(Terrill 2001). Although modern organizational
police theory generally argues that low spans-ofcontrol are inefficient to the administration of
police departments (making the assumption that
“flatter” organizations function more efficiently
than “tall” organizations) (e.g., Walker and
Katz), in communities where the risk of misconduct may be higher than average, it seems that
lower spans-of-control would reduce the risk of
brutality as supervisors have more contact
with line officers in the field. Indeed, under
such conditions, what may be sacrificed in terms
of organizational efficiency may be gained in
effectiveness.
Finally, recruitment and supervision to limit
police and the excessive use of force seem to have
their highest probabilities of success if the overall
police department culture supports the structural
efforts in place to guard against brutality. The
3557
P
tone must be set at the top through the implementation of progressive policies that strive for transparency in the disciplinary review process and
promote model tactics and strategies. In addition,
evidence suggests that when line officers are
involved in the development of use of force policies and field tactics designed to reduce police
officer violence (i.e., Fyfe 1997), then the use of
force within departments tends to decrease, partly
because police officers view themselves as contributors to the policy development process, rather
than simply as people who need to be controlled.
Once the chief administrator implements an
infrastructure of progressive policy and gains
the “buy-in” of line officers, the message of progressive policing must filter through all layers of
the police organization. One method by which to
accomplish the process of communicating the
department’s value system is via the field training
officers (FTOs). Toch (1996) argued that for the
purposes of controlling police violence, field
training officers were instrumental in bridging
the gap for new officers between the messages
they received in the training academy and the
messages they received from their peers once on
the street. Toch argued that FTOs serve a crucial
organizational role because, among other functions, they help probationary police officers practice their recently acquired skills in live settings
while under highly structured supervision. During this formative period in a police officer’s
career, FTOs can help new officers interpret
street encounters in ways that allow them to
learn and practice good policing through the metaphorical lenses of a protection of life mandate.
Once they complete their field training programs,
police officers may continue to practice good
policing and the protection of life in both patrol
and specialized settings.
Successfully preventing the use of excessive
force by police likely requires a multilayered
approach that considers the social ecology of
the total police organization. While the chief
administrator is an important part of the process
within the organization, it must be noted that the
chief also represents the will of the community.
Chief administrators for police departments are
P
P
3558
usually chosen by the elected bodies of local
governments (e.g., city councils, county boards
of supervisors, etc.), often for their stated positions on certain issues, such as violent crime,
fear of crime, urban disorder, etc. Excessive
use of force will be minimized only if governing
boards select chief administrators who are
actively interested in controlling and preventing
excessive use of force. Moreover, the governing
boards themselves need to share the same commitment to police accountability if excessive
force is to be controlled and prevented in
a police organization.
Conclusions
Evaluating the use of excessive force by the
police tends to be much more difficult than
other forms of misconduct, such as profitmotivated corruption. Unlike other forms of
police deviance, which tend to be wrong by
their very occurrence, the roots of police brutality
reside in the legitimate use of force – a tool society gives police to accomplish its mandate.
Therefore, when an officer is accused of brutality,
it is often difficult, if not impossible, to determine
where the justified use of force may have become
excessive. Even in the infamous, and unquestioningly brutal, Rodney King case, the police officers who contacted King had cause to use some
physical coercion in their initial attempts to bring
him into compliance. This entry identified classic
occupational templates of the police use of force
in an effort to identify context within which the
excessive use of force may be considered.
Although these templates have received little
empirical validation, they are meaningful
because in many cases, they offer a shorthand
for describing the conditions under which police
officers may resort to brutality. Although police
departments can use recruitment as a tool to prevent the excessive use of force by their officers,
the best prevention is likely based on a multilevel approach that involves local governing bodies selecting chief administrators with strong
commitments to transparency and the virtues of
“good” policing.
Police and the Excessive Use of Force
Related Entries
▶ Biased Policing
▶ Control of Police Misconduct
▶ Crime and the Racial Composition of
Communities
▶ Democratic Policing
▶ Drug Enforcement
▶ Hot Spots and Place-Based Policing
▶ Law of Police Use of Force
▶ Minorities Within the Police Workforce
▶ Police and the Military Nexus
▶ Police Corruption
▶ Police Culture
▶ Police Discretion and Its Control
▶ Police Integrity
▶ Police Leadership Styles
▶ Police Legitimacy and Police Encounters
▶ Police Line-Level and Agency Accountability
▶ Police Selection
▶ Police Use of Firearms
▶ Risk Management in Policing
▶ Sex Offenders and Criminal Policy
Recommended Reading and References
Anderson E (1999) Code of the street: decency, violence,
and the moral life of the inner city. New York:
W.W. Norton & Company
42 U.S.C. } 1983
Barnhill JH (2010) Watts riots (1965). In: Danver SL (ed)
Revolts, protests, demonstrations, and rebellions in
American history, vol 3. Santa Barbara, ABC-CLIO
Carter D (1985) Police brutality: a model for definition,
perspective, and control. In: Neiderhoffer A,
Blumberg A (eds) The ambivalent force. Holt,
Rinehart and Winston, New York
Fyfe J (1979) Administrative interventions on police
shooting discretion: an empirical examination. J Crim
Justice 7:309–323
Fyfe J (1986) The split-second syndrome and other determinants of police violence. In: Anne T. Campbell,
John J. Gibbs (eds) Violent Transactions. Oxford:
Basil Blackwell, pp 207–225
Fyfe J (1997) Good policing. In: Dunham R, Alpert GP
(eds) Critical issues in policing: contemporary readings,
5th edn. Waveland, Prospect Heights, pp 194–213
California Governor’s Commission on the Los Angeles
Riots (1965) Violence in the city: an end or a beginning?
Graham v. Connor, 490 U.S. at 388 (1989)
Herbert S (1997) Policing space: territoriality and the Los
Angeles Police Department. University of Minnesota
Press, Minneapolis
Police and the Military Nexus
Herbert S (1998) Police subculture revisited. Criminology
36:343–369
Hochstedler E (1981) Testing types: a review and test of
police types. J Crim Justice 19:451–466
Independent Commission (1991) Report of the Independent Commission on the Los Angeles Police Department. The Commission, Los Angeles
Jacob H (1971) Black and white perceptions of justice in
the city. Law Soc Rev 6:69–89
Kane R (2002) The social ecology of police misconduct.
Criminology 40:867–896
Kane R (2005) Linking compromised police legitimacy to
violent crime in structurally disadvantaged communities. Criminology 43:469–498
Kane R, Cronin S (2011) Maintaining order under the rule
of law: occupational templates and the police use of
force. J Crime Justice 34:163–177
Kane R, White M (2012) Jammed up: bad cops,
police misconduct, and the New York City
Police Department. New York University Press,
New York
Kerner Commission (1968) Report of the National Advisory Commission on civil disorders. Bantam Books,
New York
Klockars C (1985) The idea of police. Sage, Beverly Hills
Kubrin C, Weitzer R (2003) Retaliatory homicide: concentrated disadvantage and neighborhood culture. Soc
Probl 50:157–180
Muir WK (1977) Police: streetcorner politicians. University of Chicago Press, Chicago
Skolnick J, Fyfe JJ (1993) Above the law: police and the
excessive use of force. Free Press, New York
Sunshine J, Tyler T (2003) The role of procedural justice
and legitimacy in shaping public support for policing.
Law Soc Rev 37:513–547
Tennessee v. Garner, 471 U.S. 1 (1985)
Terrill W (2001) Police coercion. LBF Scholarly Publishing, New York
Toch H (1996) The violence-prone police officer. In:
Geller W, Toch H (eds) Police violence: understanding
and controlling police abuse of force. Yale University
Press, New Haven, pp 94–112
Tyler T (1990) Why people obey the law. Yale University
Press, New Haven
Van Maanen J (1978) The asshole. In: Manning PK, Van
Maanen J (eds) Policing: a view from the street. Goodyear, Santa Monica, pp 221–238
Walker S (2010) Sense and nonsense about crime, drugs,
and communities, 7th edn. Wadsworth, Belmont
Walker S, Katz C (2010) Police in America: an Introduction. McGraw-Hill, New York
Weber M, Owen D, Strong TB (2004) The vocation lectures: “Science as a Vocation” and “Politics as
a Vocation,” (trans: Livingstone R). Indianapolis:
Hackett Publishing Company
Westley W (1970) Violence and the police. MIT Press,
Cambridge
Wood DB (2002) L.A.’s darkest days. Christian science
monitor. Retrieved 2 Nov 2012
3559
P
Police and the Military Nexus
Otwin Marenin
Department of Criminal Justice and
Criminology, Washington State University,
Pullman, WA, USA
Synonyms
Constabularization; Militarization
Overview
The police and military (armed forces), the two
coercive agencies of the state, are restricted by
law, tradition, and policy to two domains in
which they exercise authority. The police protect
the domestic order against risk, threats, and crimes
by persuasive communications, the enforcement
of laws, and, ultimately, force (or its potential use),
while the military defends the nation and the state
against aggression from without. These analytical
and policy distinctions between two agencies (or
systems of agencies) which are authorized to
employ legitimate internal and external use of
force to deal with threats have become blurred in
practice (Easton et al. 2010). The argument and
worry that the police in the USA and other democratic countries are increasingly becoming militarized, hence will loose the civil and democratic
orientation to their work, the use of force, and the
collections of intelligence, has been raised by
a number of scholars who study the police and
by the police themselves (Kraska 2001, 2007;
Kraska and Kappeller 1997).
Much of the concern that the blurring of distinctions between the policing and the military
work will undermine civil and democratic policing rests on two limited understandings about the
police, namely, that policing has never been militarized to any degree and by comparing two
“ideal types” (yet legitimate conceptions) of
both the police and the military rather than their
historical configurations, relationships, and work.
P
P
3560
The blurring of policing and military work and
tasks is not new; it only has increased and taken
many new forms in recent times, specifically by
recent domestic and international developments
(transnational crime, fears of terrorism, civil violence in failed states, peacekeeping and peacebuilding interventions, the weakening of border
controls, technological advances in information
acquisition, processing, and distribution). The
distinctions drawn between military and police
forces have always been blurred in practice.
Police and military have worked in the same
occupational space – the protection of society
against threats; they have been bestowed by
their societies with the right to use force, and
the core values enshrined in their occupational
cultures (internal authority, discipline, closed
ranks, loyalty, courage, masculinity) are similar
(den Boer et al. 2010, 225–226).
Ideal Typing the Police and the Military
Ideal Typing Democratic, Civil Policing
The classic image of the democratic police officer
is the patrol officer, the constable walking the
beat, known by and knowing the community
and the village; approachable, protective, and
friendly when encountering citizens in need;
committed to service, the protection of human
rights, and the rule of law; doing her/his work
with integrity; and ensuring the effective protection of public order and the security of property
and lives of the populace against threats and
crimes. Democratic, civil police retain substantial domains of discretion in how to respond and
to adjust their decisions to situational exigencies,
guided by law, professional norms, public
demands, and democratic political oversight
(OSCE 2006). Since full service tasks are most
common work done by police in democratic
countries, this image is not a false one.
The essence of democratic policing is captured in six aspects of their work: authority to
control and suppress illegal conduct is limited by
and large to the domestic arena; the composition
of police personnel should approximate the distribution of salient identity groups in society;
Police and the Military Nexus
orientation to public service; the constrained use
of force; transparency in actions; and accountability to civic society by various mechanisms.
The occupational culture of democratic police
officers and the organizational policies which
shape their work and guide their discretion stress
that policing is a service to the public and not to
the state. Their use of force should be strictly
limited by law, professional norms, and organizational directives and used only as a last resort
when dealing with all people they encounter. The
police have a responsibility to explain when
asked, and the public has a right to know what
the police are doing (except for legitimate
secrecy in pending cases and investigative
aspects of their work). The police cannot hide
behind the shield of expertise and secrecy related
to their general work, nor deny the media, the
public, and the political leaders’ access to the
information in their possession, unless legitimately protected. Lastly, if allegations of misconduct, abuse of power, or corruption are made
against them, the police have an obligation to
submit to external oversight, and the public has
the right to require their submission to civic
reviews. They do not have the exclusive authority
and power to investigate allegations of misconduct in their ranks by themselves and take corrective actions. As government officials having
great power over the routines of living in society,
they cannot claim impunity for their conduct.
Ideal Typing the Professional Soldier
The military are an armed, professionally trained,
and hierarchically controlled organization in
which lower ranks are expected to respond
quickly and without questioning to commands
from the top. The military protects the nation,
its people and territory, and the “national interest” by using its ability to detect and suppress
external threats. It uses its capacity to exercise
force to capture, disable, or kill enemies of the
nation in an efficient, fast, and effective manner.
The military collects information in secret via
sophisticated and complex technologies and
human spies and evaluates and distributes information as “intelligence” which is typically kept
secret by national security justifications and
Police and the Military Nexus
tactical contingencies. Transparency is severely
limited by the need to keep the country safe from
aggression which, were the nature of the intelligence and the means by which it was collected be
known, would jeopardize the ability to plan and
execute the collection of information. Accountability of the military for its strategic and tactical
conduct is first to itself, with civilian oversight by
elected leaders the last overriding democratic
option. By law and tradition, the military in democratic societies only intervene in domestic security when national and local policing systems are
unable to maintain control.
Distinctions
In short, ideal-type descriptions of police and
military emphasize fundamental distinctions
between the two security agencies authorized to
use force in order to protect the nation-state, its
government, its economic and political ways of
life, and its people and properties against domestic and external threats.
Specifically, the police see the people they
have to deal with primarily as customers, innocents, and suspects, while the military sees real
and potential enemies. The police are committed
to maintaining public order, provide services, and
enforce domestic laws while the military seeks
the defeat of enemies and victory. The police use
force only as the last resort in encounters, while
the military resorts to force immediately to disable, capture, or kill enemies. The police tend to
work as individuals (or limited teams) while
teamwork is essential to the military. The police
at the lowest ranks of the organizational hierarchy have substantial discretion in specific actions
and encounters, a discretion that is circumscribed
by situational exigencies and the best judgment
of the officers on what is the appropriate, effective, and balanced action in the situation they are
faced with. The military limits discretion at the
lowest ranks through intensive training,
a commitment to authoritative interpretations of
what needs to be done in a situation (“that’s an
order, soldier”), and loyalty to group. The discretion of the police is influenced by personal morality and ideological beliefs, while the military
seeks to socialize new recruits and experienced
3561
P
soldiers to an ideology that eliminates personal
views as guides to actions.
The police are accountable to public oversight
and the courts; the military has its own legal
system, code of conduct, and “criminal justice”
institutions (military police, administrative tribunals and courts, investigative units). Civilian law
enters only at the appeals level, after military
institutions have reached a decision on alleged
criminal misconduct or violations of the code of
conduct. Control over actions by a professional
military in a democracy is exercised largely internally, supported by professional norms which
accept that the military is ultimately subject to
civilian leadership (by the “Commander in
Chief,” or some such term). Unlike the police,
which are always guided and judged by laws
made through an open, participatory political
process external to the police, the military judges
itself by its own laws and rules with external
domestic political oversight and international
conventions on the conduct of war as the ultimate
backstops.
The argument predicts that the “militarization” of the culture, policies, and actions of
police – that is, the adoption by civilian police
forces “of militaristic practices, hardware, technology, values, language, and ideology” (Kraska
1997: 299) – will be a danger to democratic
norms and professional ethics of the police.
Rather than follow the rule of law, professional
codes of conduct, or commitment to public service and the protection of human rights, the
police will drift toward a professional ideology
which stresses the need to protect national and
local security against threats and risks without
interference or oversight by legal institutions,
political oversight, or public demands. The necessary democratic balance between the rights of
people and the practical need for security will tilt
strongly toward security.
This argument rests on four main assertions.
First, the police will alter their orientation to the
use of force and the identification of categories of
people against whom force can and should be
used quickly and easily, justified by appeals to
an “enemy” and “threat” language. Second, the
police will shift from an orientation to public
P
P
3562
service toward a suppression of crime and disorder mentality and compliance with political definitions of threats to public order and safety.
Third, the police will be drawn into the secretive
collection of intelligence and categorical risk
assessments which invade the privacy of people
and assume threats, risks, and criminal intentions
based on an individual’s membership in categories of persons (e.g., young male members of
minority or ethnic groups, bikers, Roma,
Muslims, travelers from certain countries) as
these categories are defined by the police and
intelligence agencies. Perceived threats and
risks will have to be deterred and prevented by
proactive police and intelligence collection policies before they can be executed. Fourth, transparency will decline. The police will become less
accountable via civic oversight of their actions.
There is evidence to support these arguments
but less support for the conclusions that democratic policing is threatened and may be replaced
by new forms of aggressive, forceful policing
which undermine basic privacy rights and equal
treatment of people and communities.
Beyond Ideal Typing: Existing Overlaps
in History and Space
The reality is that the police have always done
a variety of work in nondemocratic ways and that
the military has always worked with the police on
specific security and order maintaining tasks. In
practice, historically and currently, the police and
the military have exhibited overlapping characteristics and work habits.
Varieties of Policing
Being visible, available, and responsive to the
needs of people, though crucial to any conception
of democratic policing, is not the only type of
work done by any police force. Tasks specified
for the police have always included multiple
goals and roles ranging from civil policing to
state security policing (Brodeur 1983). All police
forces anywhere include functional specializations which approach military-style tactics and
mentalities: riot control, specialized units
Police and the Military Nexus
confronting dangerous situations (armed
response teams, paramilitary police units), border
policing, state security police protecting the political regime against threats to its rule, and the
complete exception to democratic norms of colonial policing. Specialized police agencies and
units tend not to share the mentalities, priorities,
and practices which are claimed to be the defining
traits of civil, democratic, and accountable police
force.
Control of riots and demonstrations is not
targeted primarily against individual malfeasance
but is deployed as an organized, forceful police
response to public, frequently violent, disorders
(Della Porta and Reiter 1998). In most countries,
should the police fail to control disorders, the
military (e.g., the National Guard in the USA)
stands in the background ready to enter the fray
and restore order.
Distinct units trained in and employing military-style tactics (PPUs, Paramilitary Police
Units) have infiltrated normal policing,
a process which Kraska and Kappeller (1997)
have argued is expanding. PPUs are expensive;
not using them wastes public money; they have
high status within the police profession; hence,
their arms, group cohesion, and tactics are attractive to other police.
Persons entering a country at recognized border crossing points, whether legally or illegally,
have restricted democratic rights based on the
overarching need to keep the country secure
against threats and risks. Seeking to cross outside
of legal control points will bring swift and forceful reactions from border guards who can be
police, military, or integrated units (Andreas
2009; Caparini and Marenin 2006).
At the US land borders, Marine units are
deployed along the Mexico-US border, nominally for exercises but also to occupy and preempt territory through which threats and illegal
migrants might enter the USA (Dunn 1996). At
the Canada-US border, IBETs (Integrated Border
Enforcement Teams), which include Canadian
and US police, military, and intelligence agents,
are a more recent development dealing with security threats (IBETs, web). Also, at that border,
military drones have been called upon to assist
Police and the Military Nexus
local law enforcement in detecting fleeing felons
and have responded successfully. The Coast
Guard, an arm of the armed forces, has been
responsible for the protection of sea borders. In
most countries, including the European Union
member states, border control is shared with
police agencies or done exclusively by the
military.
State security police tasked with protecting the
state, the ruling regime, and the powerful against
challenges to their rule and status have few
qualms about abiding by legal and democratic
norms, nor do state security police take their
orders from or are responsive to public concerns.
Their job is to keep those in power safe. This is
true in democratic countries and even more so in
authoritarian, transitional, or military states. The
prototype of state security police agencies were
the KGB in the Soviet Union and other socialist
and authoritarian countries, but they can also be
found in democratic countries, as during the
Franco dictatorship in Spain when the Guardia
Civil (a constabulary force) was the tool used to
terminally eliminate political and intellectual
opponents of the regime, or in lesser and more
benign forms, such as transnational threat assessment and anti-terrorism units in police departments or specialized security agencies, to arrest
suspects before they can act (Brodeur 2000).
In the history of state development (Bayley
1985; Tilly 1990) and the creation of security
forces, there has always been a significant overlap between police and military work. That overlap has been strengthened in recent years by the
shift in organized violence from traditional forms
of aggression (requiring a military response)
toward diffused, asymmetrical, nonconventional,
substate forms – civil wars, guerilla insurgencies,
identity group-based riots, and killings; the emergence in many parts of the world of powerful
organized transnational crime groups who exert
influence on states and societies by corruption,
fear, and violence; and the growth in privatized
policing and international policing and security
providing companies. These developments have
been enabled and smoothed by new technologies
for sharing information and intelligence on
a (almost) real-time basis. This new world (dis)
3563
P
order (Kaldor 2001; Oakley et al. 1998) has led to
a transformation in how all state-based, and even
non-state, actors (e.g., private risk assessment
and security companies) can be engaged in an
integrated manner against new threat and risk
dynamics.
This convergence is most apparent in internationally sanctioned peacekeeping and peacebuilding operations. Police in UN and regional
peacekeeping interventions, where local security
forces have collapsed (e.g., Haiti, Timor Leste,
Kosovo), have had to assume semi-military suppression and control functions as well as “normal”
policing operations (Bayley and Perito 2010;
Goldsmith and Harris 2010; Greener 2009).
The biggest exception to democratic, nonmilitarized policing have been colonial policing
systems. When European colonial powers developed police organizations for their colonies, they
did not export policing systems deemed appropriate for their own, that is, civilized people, but
constabulary forces to protect their economic,
political, and “civilizing” interests in their colonies. As the majority of the world’s territories and
states are former colonies, at independence they
inherited nondemocratic and militarized policing
systems which since have been quite resistant to
changes, despite massive efforts to make them
more democratic in their performance.
Constabulary forces are semi-militarized
police units created in different political, organizational, and cultural forms. In France, the gendarmerie developed as an occupying force in
outlying areas of what is now France, in regions
which opposed the imposition of central rule
from Paris. In consequence, until very recently,
the gendarmerie was located within the Ministry
of Defense, even though it had become the police
force which protected small towns and rural areas
in France.
British policy on policing England versus
policing the colonies exemplifies the use of constabulary forces. The British practiced how to
control restless populations during their centuries-long and brutal rule of the Irish. After Irish
independence, policing in Northern Ireland, which
remained part of the United Kingdom but was
beset by ethnic, religious, and political violence
P
P
3564
and conflicts, continued the constabulary tradition
via the Royal Ulster Constabulary, which was used
to protect Irish loyalists in Northern Ireland often
in cooperation with the British military. The
lessons learned by the British in their colonial
rule of Ireland became the model for policing
their colonies (Brogden 1987).
The Spanish and Portuguese colonizers
exported their security systems to Latin and Central America which closely combined military
and policing functions under one overarching
organization. The police in most Latin American
countries remained part of the military until fairly
recently but still retain their titles (e.g., the
Policia Militar in Brazil) and their outlook on
citizens and criminals, captured nicely in the
title of Husain’s (2007) study of police reform
in Rio de Janeiro, “Those Who Die in War Are
Not Innocent.”
The USA established constabulary forces in
Central America and the Caribbean to help protect
American investments and personnel against “revolutionary” agitations. The constabularies became
converted into (para)military forces once local
control took over. In Nicaragua, the armed forces
of the Somoza government, which engaged in
a protracted war with the Sandinista Revolutionary
Front to try to keep the government in power, were
drawn mainly from that constabulary and were
supported by illegally funded paramilitaries (the
contras) by the USA. US police assistance under
the aegis of OPS (Office of Public Safety), in the
late 1960s and early 1970s, stressed counterinsurgency over democratic norms and roles which
were easily picked up by local police forces in
Southeast Asia and Latin America, the recipients
of the majority of OPS funding and mentoring
(Huggins 1987).
Colonial policing had few attributes that one
would characterize as civil or democratic policing. The police in colonies were used mainly by
the colonial administrators to “pacify” the restless native, engage in military-style campaigns to
punish local groups which resisted the imposition
and demands (pay taxes, conscripted work) of
colonial rule, protect tax collectors and the persons and property of commercial enterprises and
missionaries, and suppress local riots and
Police and the Military Nexus
sabotage. There was no pretense that the police
were to serve the population in any meaningful
way; their job was to enable and protect colonial
rule and all its activities, privileges, and laws.
The modification of imposed policing system
after decolonization has been slow and incremental, and much of the occupational cultures and
priorities of the police have not been responsive
to demands for reforms. For example, the most
populous democracy in the world, India, still uses
the colonial Police Act, passed in 1861, as the
basis for defining the jobs of the police which,
despite numerous constitutional, ideological, and
legal statements, still shape the way in which
state and federal police understand their job
(Verma 2011). In Morocco, military officers
still occupy the commanding positions in the
gendarmerie, the police force which patrols the
roads and services the rural areas of the country.
In short, democratic policing, operationally
defined as being available and willing to be of
service to people in need, has been a rare policing
model in history or is currently across the globe.
Still, democratic policing as practiced by developed countries, currently in the guise of some
form of community policing, has become the
goal of police reforms promoted through international assistance programs. Militarization is seen
as threatening progress toward that goal.
Reforms of policing systems in countries
where policing and justice systems have collapsed increasingly include the institutional separation of the police from the military and argue
for the demilitarization of the occupational culture of the police. For example, the Chapultepec
Peace Accords, signed in 1992, which ended the
civil war in El Salvador, dissolved the National
Police, then under the control of the military,
and reconstituted a National Civil Police with
detailed instructions on the organization, tasks,
recruitment, training, and democratic responsibilities. Reform and democratic governance
were deemed not possible unless the police were
reconstituted and removed from the organizational authority of the military.
Yet one must be careful when drawing
a distinction between the organizational home
of “policing agencies,” such as the gendarmerie,
Police and the Military Nexus
which are housed within the armed forces, and
organizational and occupational cultures. Being
under the administration of the military does
not prevent democratic policing, nor does removing them from military administrative oversight
create more democratic policing. The police
can be quite forceful and abusive whether they
work under military or civilian organizational
umbrella. What matters are the organizational
and occupational norms and cultures which are
the guiding doctrines and “recipe rules” for the
police.
Constabularization of the Military
The military, in its various institutional divisions,
has been increasingly drawn into what used to be
considered “policing” work in a variety of ways:
assisting in border controls, intelligence sharing
with domestic law enforcement agencies, counterinsurgency strategies to reach the hearts and minds
of local populations, United Nations or regionally
authorized peacekeeping operations which include
military and police contingents, and “nation-building” after multilateral and regional interventions
in failed and conflicted-ridden states. These trends
undermine the traditional restrictions of the armed
forces to war-fighting and war-ending work by
involving them in tasks for which they have little
training, few skills, and limited willingness
(Friesendorf 2010; Friesendorf and Kempel
2011; Perito 2011).
Training for military and police being
deployed in international interventions, and
increasingly for public order policing domestically, is conducted in a language which blurs
traditional lines, which has led to both the
constabularization of military goals and cultures
and the paramilitarization of policing ideologies
(Moelker 2010), and which will continue to
undermine the bright line between police and
military organizational cultures and practices
and the distinction between domestic and external authority to exercise coercion.
This convergence is not by design but forced
by the nature of current international and domestic security climates. For example, the explosion
of transnational crimes which affect domestic
security (drugs, human trafficking, arms trade)
3565
P
requires that police now work in other countries,
in cooperation with local police, militaries,
and intelligence agencies to control crime.
Preventing convergence, or alternatively ensuring a continued separation and distinction
between police and military work, will be difficult to maintain.
Beyond Dichotomies: Theorizing
the Police-Military Nexus
How to theorize the new security structures? The
most promising theoretical approach is the notion
of a security sector or system (SS) and
corresponding conceptualizations of reform
(SSR) and governance (SSG). The security sector, in its leanest definition, includes the armed
forces, police, border guards/police, and intelligence agencies as the core state security providers. In more expansive conceptions, the SS
includes the criminal justice system, legal aspects
of security, and non-state providers (Bryden and
H€anggi 2004; OECD Organization for Economic
Co-operation and Development 2007).
Security sector theorizing stresses the interconnectedness of the four core state security
agencies. Reforms of the security sector, or specifically the police, to establish more democratic
professional and accountable norms and actions
cannot be simply focused on the police but must
take their connections to other state-based security and private security providers into account.
Security is provided by many actors, and that
requires that they cooperate and work together
domestically and internationally.
The political nature of policing, and other
security policies, is central to the notion of
a security sector and its governance. In addition
to demilitarizing the police, reformers argue that
the police be de-linked from centralized, political
control of operational policies and tactics.
Concluding Comments
Concerns about militarizing the police are not
illusory. Civil policing would change
P
P
3566
significantly if military cultures, norms, and practices began to infiltrate democratic cultures and
styles of policing.
At the same time, militarization is not the
greatest threat to democratic policing. More
serious is the involvement of the police in secretive intelligence gathering, often along stereotypical categorizations of terrorist, transnational
crime groups, or local “troublemakers,” such as
“suggestions” from the FBI to local police to
map the distribution of mosques in their communities or efforts by the antiterrorist unit in the
New York Police Department to infiltrate and
keep tabs on Islamic leaders. The police who
conduct such secret invasions and information
gathering of people’s normal, and legal, activities by the justification of local and national
security will loose, once their police activities
become known, political support and
legitimacy.
The secret gathering of intelligence is just as
likely to change the norms and cultures of policing as does militarization. The loss of transparency in collecting intelligence about normal
activities disconnected from actual criminal
threats and the convergence of domestic and
international work from both the police and the
military perspectives, both driven by political and
national security justifications and the changing
global world, are the historical and current reality
which will not disappear. For example, President
Reagan declared, by his authority, international
drug trafficking a national security threat, not just
a crime, thereby authorizing crime control and
intelligence collecting actions which were legally
prohibited to domestic police.
The militarization of the police happens for
diverse reasons, be it by the promotion of that
style of defining their work by police themselves;
by the acceptance of a frustrated public fed up
with disorder, crime, and fears which beset their
lives and willing to give up some rights for
greater protection; by depictions in the mass and
entertainment media which portray armed police
tactics as the normal response to increasingly
dangerous situations in normal police work; and
by the exploitation of risks and threats by political polemicists.
Police and the Military Nexus
The issues are whether the militarization of
policing and the constabularization of the military will become dominant norms and ideologies
and whether the in-creeping of military styles
into democratic policing will move that style
permanently way from its service and order protection goals. It seems an unlikely possibility, as
democratic policing is promoted worldwide, as
part of larger political changes, by reformers,
police officials, and progressive political leaders
and also demanded by populations, as long as
effectiveness levels are maintained sufficient to
keep populations willing to grant legitimacy to
democratic styles which balance crime control
with justice. In the end, the greater threat to
democratic policing is not militarization but the
politicization of policing even in democratic
countries, of which militarization is only one
piece of evidence, and the increasing closure of
the police to public knowledge and oversight
linked to an increasing disregard for the rights
of people and due process.
Related Entries
▶ Conceptualizing of Police
▶ Democratic Policing
▶ Policing of Peacekeeping
Recommended Reading and References
Andreas P (2009) Border games: Policing the
U.S.-Mexico divide, 2nd edn. Cornell University
Press, Cornell
Bayley DH (1985) Patterns of policing. A comparative
international analysis. Rutgers University Press,
New Brunswick
Bayley DH, Perito R (2010) The police in war. Fighting
insurgency, terrorism, and violent crime. Lynne
Rienner, Boulder
Brodeur JP (1983) High policing and low policing:
Remarks on the policing of political activities. Soc
Probl 30:507–520
Brodeur JP (2000) Cops and spooks: the uneasy partnership. Police Pr Res 1:299–321
Brogden M (1987) The emergence of the police – the
colonial dimension. Brit J Criminol 27:4–14
Bryden A, H€anggi H (eds) (2004) Reform and reconstruction of the security sector, Baden-Baden. Nomos
Verlagsgesellschaft, Germany
Police Corruption
Caparini M, Marenin O (eds) (2006) Borders and security
governance. Managing borders in a globalised world.
LIT Verlag, Vienna
Della Porta D, Reiter H (eds) (1998) Policing protest.
The control of mass demonstrations in western
democracies. The University of Minnesota Press,
Minneapolis
den Boer M, Janssens J, Vander Beken T, Easton M,
Moelker R (2010) Epilogue. Concluding notes on the
convergence between military and police roles. In:
Easton M, den Boer M, Janssens J, Moelker R, Vander
Beken T (eds) Blurring military and police roles.
Eleven International, The Hague, pp 223–228
Dunn TJ (1996) The militarization of the U.S.-Mexico
border, 1972–1992. Texas University Press, Austin
Easton M, den Boer M, Janssens J, Moelker R, Vander
Beken T (eds) (2010) Blurring military and police
roles. Eleven International, The Hague
Friesendorf C (2010) The military and law enforcement in
peace operations. Lessons from Bosnia-Herzegovina
and Kosovo. Geneva Centre for the Democratic Control of Armed Forces, Geneva
Friesendorf C, Kempel J (2011) Militarized versus civilian
policing: problems of reforming the Afghan national
police. Peace Research Institute, Frankfurt (PRIF
Report No. 102)
Goldsmith A, Harris V (2010) Police-military cooperation
in foreign interventions: Timor-Leste and the Solomon
islands. In: Lemieux F (ed) International police cooperation. Emerging issues, theory and practice. Willan,
Cullompton, pp 221–237
Greener B (2009) The new international policing.
Palgrave Macmillan, New York
Huggins MK (1987) U.S.-supported state terror: a history
of police training in Latin America. Crime Soc Justice
2728:149–171
Husain S (2007) Those who die in war are not innocent
(‘Na Guerra, quem morre Não É innocente): human
rights implementation, policing, and public security in
Rio de Janeiro. Rozenberg, Brazil/Amsterdam
Integrated Border Enforcement Teams (IBETs). http://
www.rcmp-grc.gc.ca/ibet-eipf/index-eng.htm
Kaldor M (2001) New & old wars. Organized violence in
a global era. Stanford University Press, Stanford
Kraska PB (1997) The military as drug police. Exercising
the ideology of war. In: Gaines LK, Kraska PB (eds)
Drugs, crime, and justice. Contemporary perspectives.
Waveland, Prospect Heights, pp 297–320
Kraska PB (ed) (2001) The militarizing of the American
justice system. The changing roles of the armed forces
and the police. Northeastern University Press, Boston
Kraska P (2007) Militarization and policing – its relevance
to 21st century police. Policing 1:501–513
Kraska P, Kappeller VE (1997) Militarizing American
police: the rise and normalization of paramilitary
units. Soc Probl 44:1–18
Moelker R (2010) Cultures converging upon constabularization. In: Easton M, den Boer M, Janssens J,
Moelker R, Vander Beken T (eds) Blurring military
3567
P
and police roles. Eleven International, The Hague,
pp 151–169
Oakley RB, Dziedciz MJ, Goldberg EM (eds)
(1998) Policing the new world disorder: Peace operations and public security. National Defense University
Press, Washington, DC
OECD (Organization for Economic Co-operation and
Development) (2007) OECD handbook on security
system reform. Supporting security and justice, 2007
Edition. OECD Publishing, Paris. www.oecd.org/
dataoecd/43/25/38406485.PDF
OSCE (Organization for Security and Co-operation in
Europe) (2006) Guidebook on democratic policing.
Author: Vienna
Perito RM (2011) The Iraq federal police. U.S. police
building under fire. Washington, DC, United States
Institute of Peace, Special Report 29x
Tilly C (1990) Coercion, capital and European states, AD
990–1990. B. Blackwell, Cambridge, MA
Verma A (2011) The new khaki. The evolving nature of
policing in India. CRC Press, New York
Police Corruption
Sanja Kutnjak Ivković
School of Criminal Justice, Michigan State
University, East Lansing, MI, USA
Overview
Police corruption is a form of police misconduct
or police deviance typically defined through
the motivation to achieve personal gain.
Police corruption includes many heterogeneous
forms of behavior that could be classified on
the basis of several criteria, including the
motivation for corruption (i.e., economic corruption v. noble-cause corruption), regularity of
payments (i.e., pads v. scores), consequences
(i.e., distortive v. non-distortive corruption), and
the level of aggressiveness (i.e., grass-eaters v.
meat-eaters). Barker and Roebuck developed
a typology of police corruption which recognizes
corruption of authority, kickbacks, opportunistic
theft, shakedowns, protection of illegal activities,
the fix, direct criminal activities, and internal
payoffs. Punch added flaking (i.e., planting of
evidence) or padding (i.e., supplementing of
evidence) as the ninth type of corruption.
P
P
3568
Theories explaining the causes of police corruption could be classified into four larger
groups: (1) theories oriented toward individual
police officers and their characteristics
(i.e., “individualistic theory” or “rotten apple
theory”), (2) theories examining the relation
between opportunities for corruption and police
corruption (“occupational theories”), (3) theories
focusing on the relation between the characteristics of a specific police agency and police corruption (“organizational theories”), and (4) theories
concerning the relation between the society
at large and police corruption (“social structure
theories”). Despite the existence of extensive
literature on the theories of police corruption,
research that tests the theories is very limited.
Police corruption could be measured using
several different methodologies. A relatively
simple, yet very inaccurate way of measuring
corruption is through the official data such as
the arrest rates, complaint rates, and conviction
rates. Somewhat more promising attempts to
assess the actual level of corruption are public
opinion surveys, police officer surveys, field
studies, and investigations by independent
commissions.
Police Corruption
Definition
Police corruption is a form of police misconduct
or police deviance typically defined through the
motivation to achieve personal gain (e.g., Barker
and Carter 1986; Goldstein 1975; Klockars et al.
2000; Kutnjak Ivković 2005; Sherman 1974).
The definition contains several essential
elements.
First, corrupt behavior could be defined as
a violation of the penal codes, administrative
agency rules, or the codes of ethics. Like other
citizens, police officers could violate norms of
federal and state criminal codes. In addition,
federal and state codes establish certain crimes
which only public officials can commit
(e.g., bribery of public officials and witnesses,
Title 18 of the U.S. Code, Chapter 11,
Section 201, 1999; extortion by public officials,
Police Corruption
Title 18 of the U.S. Code, Section 872, 1999;
deprivation of civil rights, Title 18 of the U.S.
Code, Section 242, 1999). Police agencies,
particularly large municipal agencies, have
administrative rules and policies compiled in the
standard operating procedure manuals. These
administrative rules channel the use of discretion,
describe appropriate conduct of police officers,
prohibit inappropriate conduct, instruct officers
to complete written reports after critical
incidents, and require supervisory oversight
(see, e.g., National Research Council 2004).
Even in the agencies with extensive rules, the
rules could be ambiguous or fail to regulate
certain aspects of corruption. Lastly, the International Association of the Chiefs of Police has
developed the code of ethics for the US police
officers, which explicitly asks of police officers to
promise that they would not engage in corruption,
bribery, or gratuities themselves nor condone
such conduct by their fellow police officers.
Second, while corruption is typically viewed
as corrupt behavior (thus suggesting that something is being actively done), corruption
can include both active (i.e., act) and passive
(e.g., omission) forms. A police officer who
fixes a felony ticket in exchange for a bribe
provides an example of an action, while a police
officer who does not write a ticket in exchange for
a bribe from a motorist caught speeding provides
an example of omission.
Third, various definitions discuss the gain
resulting from corrupt activities in terms of
explicit personal gain (e.g., Goldstein 1975;
Kutnjak Ivković 2005; Moore 1997) or implied
personal gain (i.e., the definition rests on the idea
that the money/goods from the corrupt act are to
be consumed by the individual officers; e.g.,
Barker and Wells 1981; Roebuck and Barker
1974). Several authors (e.g., Bracey 1995; Carter
1990) included organizational gain in their definitions as well.
The goal typically has a monetary value; it
typically involves “some tangible object, either
cash, services, or goods that have cash value”
(Barker 1996, p. 25). However, it is also possible
that the gain is nonmonetary. While monetary
value could usually be attached to, or calculated
Police Corruption
for, most of the items, in some instances it is
difficult, or borderline impossible, to determine
it. For example, fixing a speeding ticket for
a sergeant’s niece could have different price
tags attached to it, depending on the extent of
sergeant’s gratitude and the ways in which he is
ready and willing to express them. It can potentially lead to preferential treatment by the
sergeant and the early promotion, thus yielding
thousands of dollars in gain.
The size of the gain may be an important
factor. The discussion in the literature about the
minimum amount, that is, the amount below
which the gifts of this value would not be considered as corruption, is intense. The Code of Ethics
is explicit in prohibiting the acceptance of gratuities (“. . .never accepting gratuities”; 2011).
Although the value of these gratuities may be
small in each individual case, if such gifts are
given on a regular basis (e.g., Ruiz and
Bono 2004) and/or to a large number of police
officers (e.g., Pennsylvania Crime Commission
1974), their total value vastly increases their
individual value. Furthermore, the purpose of
the gratuities may be to influence the police officers’ decisions, as was illustrated on the examples uncovered by independent commissions
(e.g., Knapp Commission 1972; Mollen Commission 1994; Pennsylvania Crime Commission
1974). Barker and Wells (1981, p. 11–12) argue:
Police corruption for many officers often begins
with the shared belief among the police peer
group that “policemen have a right to a break,”
and the progression along the continuum of corruption is often so gradual that an officer is deeply
involved before he realizes it.
The acceptance of gratuities could start “the
slippery slope of corruption,” leading from
the less serious forms of corruption toward the
more serious ones; “. . . [i]t is claimed that
the acceptance of small gratuities such as free
cups of coffee by police officers will increase
the likelihood of, or lead by degrees to, or is not
significantly different from, corruption of the
worst kind (Kleinig 1996, p. 174). There are
several steps in the process of “becoming bent,”
as Sherman describes it (1985), from the acceptance of minor “perks” and free drinks to regular
3569
P
payoffs and the involvement in the distribution
and use of narcotics. As Michael Dowd (a former
NYPD police officer who stole money and drugs,
participated in larger drug rings, became a drug
dealer, and was eventually caught, tried, and
sentenced to 14 years in prison) emphasized in
his testimony before the Mollen Commission, he
would test the new officers with initial temptations which were but trivial violations of the
NYPD’s policy (Mollen Commission 1994).
Once the officers crossed the line between the
allowed and the forbidden, it was easier for
them to continue and justify to themselves their
further involvement in various violations of the
Department’s policy.
Proponents of the acceptance of gratuities
(e.g., Kania 1988, 2004) argue that the acceptance of gratuities contributes toward the development of a friendly bond between the police
and the citizens. Kania (1994, p. 2) continues
that, although the “cheerful waves, warm greetings and welcome smiles,” and letters of appreciation may result in better service just as likely
as may the free drinks, police administrators
obviously do not want to discourage such
friendly gestures.
Typology of Police Corruption
Police corruption is a group of heterogeneous
activities, from the acceptance of a bribe for not
issuing a speeding ticket to leaking the information about the upcoming raid to the drug dealers.
Literature has used several different ways of classifying these diverse activities.
Economic v. Noble-Cause Corruption
Corruption described in this entry is economic
corruption or material-reward corruption. Its
primary motivation is the material reward, be it
in the form of money or other goods and/or services. Noble-cause corruption is different; it is
defined as “corruption in the name of the moral
rightness of good ends” (Crank and Caldero
2000, p. 5). When police officers behave in
a manner consistent with noble-cause corruption,
their actions are no longer regulated by the law;
instead, they “act if they are the law” (Crank and
Caldero 2000, p. 75). Police officers attempt to
P
P
3570
rationalize this “noble-cause corruption” as
follows (Moore 1997, p. 63):
I did something wrong, but justice demanded it, not
tolerated it but demanded it, because I could put the
guy away who otherwise wouldn’t be successfully
prosecuted. I, the police officer, wouldn’t gain personally from it; I didn’t get anything from it. I only
acted for the community in the community sense of
justice to accomplish this goal.
The relation between the police and the law
changes in the case of noble-cause corruption
(e.g., Crank and Caldero 2000). Police officers
violate the law to achieve higher purposes or ends
of policing. The causes of economic corruption
and noble-cause corruption are different, as are
the efforts used to control them.
Pads v. Scores
In terms of the regularity of the payments, corrupt
activities could be divided into “pads” and
“scores.” If the payment takes place just once,
as the opportunity presents itself, it is called
a “score.” The Knapp Commission (1972, p. 66)
described the “score” as “a one-time payment
that an officer might solicit from, for example,
a motorist or a narcotics violator. The term is also
used as a verb, as in ‘I scored him for $1,500.’”
On the other hand, if there is an established
arrangement between a citizen and a police officer
and the payments occurs on a regular basis, they
are called “pads.” The Knapp Commission (1972,
p. 66) described “pads” as “regular weekly,
biweekly, or monthly payments, usually picked
up by a police bagman and divided among fellow
officers. Those who make such payments as well
as policemen who receive them are referred to as
being ‘on the pad.’”
Distortive v. Non-distortive Corruption
This classification focuses on the outcome of the
police officer-citizen interaction. If, compared to
the interaction without corruption, the outcome
of the interaction with corruption is different,
then such corruption is called “distortive corruption.” For example, a police officer stops
a motorist for speeding and wants to write
a speeding ticket. If the citizen does not offer
a bribe, the police officer will write the ticket.
Police Corruption
If, on the other hand, the citizen offers a bribe and
the police officer accepts it, the police officer will
not write the ticket. Thus, the outcome of the
interaction (i.e., ticket v. no ticket) is different,
depending on whether the bribe was offered and
accepted. In cases of distortive corruption, the
police officer does something that he was not
supposed to do (e.g., revealed an undercover
operation to a drug dealer, issued a permit to
a nonqualified applicant) and does not do something he was supposed to do (e.g., did not arrest
a citizen caught violating the law, did not issue
a license to a qualified applicant).
If the outcome of the interaction is the same,
regardless of whether the bribe was given or not,
then such corruption is called “non-distortive
corruption.” For example, a police officer issues
a passport or a liquor license to a qualified applicant. The police officer does what he is supposed
to be doing or does not do something that he is not
supposed to be doing. The reason for the bribe is
that the citizen wants to secure something that he
has a legal right to but may need to speed up the
bureaucratic machine (hence the “grease money”
term used for this type of corruption).
Grass-Eaters v. Meat-Eaters
According to the Knapp Commission (1972,
p. 4, 65), dishonest, corrupt police officers
can be classified as either “grass-eaters” or
“meat-eaters.” “Meat-eaters” are police officers
who aggressively misuse their police power for
personal gains, while “grass-eaters” simply
accept the payoffs the circumstances of police
work throw their way (Knapp Commission
1972, p. 4). A police officer like Michael Dowd
who beats up drug dealers and steals their money
and drugs (e.g., Mollen Commission 1994) would
be a clear example of a “meat-eater.” A police
officer who is on the pad and who accepts the
money that the bagman is giving him every 1st of
the month (see, e.g., Knapp Commission 1972)
would be an example of a “grass-eater.”
A strong support of the code of silence and
reluctance of honest police officers to report their
corrupt colleagues, on the one hand, and
an aggressive misuse of police powers by
“meat-eaters,” on the other hand, provide an
Police Corruption
atmosphere in which it is very easy and natural
for “grass-eaters” to accept payoffs. The
Knapp Commission (1972, p. 65) suggested that
“grass-eaters” continued to accept the gain from
corrupt activities out of their feeling of loyalty to
their fellow officers and that the “grass-eaters”
are “the heart of the problem” since “[t]heir great
numbers tend to make corruption ‘respectable’”
(Knapp Commission 1972, p. 4). According to
the Knapp Commission (1972, p. 65), unlike
“meat-eaters,” “grass-eaters” are not willing to
take considerable risks in order to obtain illegal
gains, and their behavior is considerably more
likely to be affected by a change of atmosphere/
attitudes in the department.
Barker and Roebuck’s Typology
The most frequently used typology of police
corruption has been developed by Barker and
Roebuck (1973; Roebuck and Barker 1974). In
their joint work, Barker and Roebuck (1973;
Roebuck and Barker 1974) point out that corruption takes many forms. Based on several dimensions (e.g., acts and actors involved, norms
violated, support from peer group, organizational
degree, police department’s reaction), they
specify eight types of corruption: corruption of
authority, kickbacks, opportunistic thefts, shakedowns, protection of illegal activity, the fix,
illegal criminal activity, and internal payoffs.
The first type involves cases of corruption of
authority. Barker and Roebuck (1973, p. 21)
describe it along the following lines: “the
officer’s authority is corrupted when he receives
officially unauthorized, unearned material gain
by virtue of his position as a police officer without violating the law per se.” Thus, the actions are
not violations of criminal law, but they do violate
departmental policies. Examples of such activity
would include acceptance of free drinks, free
food, discounts on merchandise, or providing
property protection for a fee. According to the
findings of the Knapp Commission (1972) and
Barker and Roebuck’s study (1973, pp. 22–23),
the practice seemed to be widespread (31 % of the
businessman in a study openly acknowledged
providing favors to the police), and the corruptors, typically respectable citizens, according to
3571
P
Barker and Roebuck, give these gifts for a reason.
Police officers perceive these gratuities as informal rewards and approve of and support them
within the peer group; police officers who refuse
such gifts are perceived by their fellow officers as
deviant.
The second type of police corruption involves
kickbacks. They are defined as the acceptance
of goods, services, or money for referring
business to various businesses and service providers, including towing companies, ambulances,
garages, lawyers, and doctors (Barker and
Roebuck 1973, p. 24). These activities do not
violate criminal laws, but they do violate departmental policies. According to Barker and
Roebuck (1973, pp. 24–25), the corruptors are
legitimate businessmen and professionals whose
purpose is to develop and maintain a good working relationship with the police. These payments
are perceived by police officers as clean and are,
therefore, supported by the peer group (Barker
and Roebuck 1973, p. 25), whereas the departments typically either condone or overlook these
kickbacks, provided that they are made by legitimate businesses discreetly. In the case of goods
and services, the department’s reaction ranges
from acceptance to mild sanctions, with cash
rewards resulting in a more severe punishment.
The third type of police corruption involves
opportunistic theft from arrestees, victims,
crime scenes, and unprotected property (Barker
and Roebuck 1973, p. 26). These behaviors
violate both the departmental rules and the criminal rules. The disciplinary measures vary from
mild to very serious (including criminal charges
and dismissal).
The fourth type of police corruption involves
shakedowns. Shakedowns occur upon the police
officer’s discovery of both the criminal violation
and the violator, and they result in the police
officer’s acceptance of a bribe in exchange for
not making an arrest (Barker and Roebuck 1973,
p. 27). One of the characteristics of shakedowns
is that the victim is rather unlikely to complain,
because the victim and the police officer could be
found guilty of another crime – bribery. These
actions are, of course, violations of both
the departmental rules and the criminal codes.
P
P
3572
The peer culture and the department itself may
react differently to the “clean” money than to the
“dirty” money, which are distinguished by the
source of the money. Officers who are publicly
exposed receive severe punishment, including
potentially dismissal, and may face a criminal
prosecution (Barker and Roebuck 1973,
pp. 28–29).
The fifth type of police corruption involves
protection of illegal activities. Individuals
involved in illegal activities reward police
officers in order to be able to operate without
police harassment (Barker and Roebuck 1973,
p. 29). The corruptors may be citizens with long
criminal records, as well as legitimate businesses
operating illegally. Some of the activities protect
the illegal services and goods which citizens
desire or perceive as necessary. This form of
police corruption involves a high degree of
organization; police members must be coordinated and know which places enjoy police protection. Departmental reaction depends on the
“degree of its own involvement with criminal
organizations or legitimate businesses that
operate illegally, informal definition of clean
money, identity of the corruptor, and whether or
not there is public disclosure of flagrant
violations” (Barker and Roebuck 1973, p. 33).
The sixth type of police corruption involves
the fix. The fix may include “the quashing of
prosecution proceedings following the offender’s
arrest and . . .the taking up (disposal of record) of
traffic tickets” (Barker and Roebuck 1973, p. 34).
The corruptors are arrestees who want to avoid
the court action in their case, and the police
officer involved typically “fails to request prosecution, tampers with the existing evidence, or
gives perjured testimony” (Barker and Roebuck
1973, p. 34). The peer approval depends on
whether the case to be fixed is a felony, misdemeanor, or a traffic case. Departmental reaction
to the fixing of criminal cases is generally severe
(Barker and Roebuck 1973, p. 35).
The seventh type of police corruption is the
police officers’ involvement in direct criminal
activities. Barker and Roebuck (1973, pp. 35–36)
argued that “policemen directly commit crimes
against the person or property of another for
Police Corruption
material gain, acts which are clear violation of
both departmental and criminal norms.” Because
the profit coming from this transaction is perceived to be “dirty” money, the fellow police
officers would typically provide very little
support for this type of police corruption and
even the departments that may tolerate other
forms of police corruption, will react severely
by firing the police officers and pressing criminal
charges (Barker and Roebuck 1973, p. 36). Some
organization is typically required for this type of
police corruption to be carried out, and it involves
small groups of police officers.
The last, eighth type of police corruption
involves internal payoffs. In the case of internal
payoffs, both the corruptors and the corrupted are
police officers who sell or buy assignments, offdays, holidays, promotions, etc. (Barker and Roebuck 1973, p. 36). Peer groups either do not support this type of police corruption or perceive it as
necessary and inevitable if they are engaged in
other forms of police corruption (Barker and Roebuck 1973, p. 37). In the departments in which
other forms of police corruption flourish, this form
exists as well and is highly organized. Possible
reactions by the departments may range from
informal approval to dismissal and pressing criminal charges (Barker and Roebuck 1973, p. 38).
Punch (1985) added flaking (i.e., planting of
evidence) or padding (i.e., supplementing of evidence) as the ninth type of corruption. According
to Punch (1985), this type of corruption is particularly evident in drug-related cases.
In a nationwide study of more than 3,000
police officers from 30 diverse US police agencies, Klockars and colleagues (2000) found that
shakedowns and opportunistic thefts were evaluated to be the most serious types of corruption in
all 30 agencies, while the acceptance of gratuities – be it on a regular basis or only for the
holidays – was viewed as the least serious form
of corruption in all 30 agencies, with the cases of
internal corruption and kickbacks lying somewhere between these two extremes. These results
are very consistent with results of surveys of
police officers from 13 other countries as diverse
as Croatia, Finland, Japan, Pakistan, and South
Africa (Klockars et al. 2004).
Police Corruption
Causes of Police Corruption
The National Research Council (2004, p. 271)
points out that “[t]he research literature
[on causes of police corruption] is long on theory and short on evidence about what causes
police corruption.” The existing literature proposes several different approaches toward
understanding why police officers engage in
corrupt behavior, ranging from police officer
individual characteristics to characteristics of
the police agency itself and its larger social
environment.
The first group of studies focuses on individual police officers and their characteristics
(e.g., Muir 1977). The literature in this area
tries to ascertain the features which make police
officers prone to corruption, such as their
prior criminal record and weak moral values.
The Knapp Commission (1972) called this the
“rotten apple approach.” Over time, scholars
discovered that psychological screening tests,
traditionally used to prevent future “rotten
apples” from entering the police organization,
are not accurate predictors of future behavior.
The second group of studies examines the
relation between opportunities for corruption
and police corruption. By its nature, policing is
viewed as an occupation rife with opportunities
for corruption (Klockars et al. 2000). Crank and
Caldero (2000, p. 63) provide an example:
When we think of police corruption, graft typically
comes to mind. The police, in their day-to-day
pursuits, are exposed to great temptations, and
there are few observers to watch what they do.
Imagine this. You’re a street officer. You’ve just
made an incredible drug bust. Your reputation in
the department is assured, and you’re feeling
charged! There’s a pile of money on the floor. All
you have to do is reach down, scoop up a handful,
and put it in your pocket, and you can put your kid
through college.
The opportunities for corruption vary assignments, ranks, units, and police agencies. Detectives, particularly those assigned to narcotic
units, have especially extensive opportunities
for corruption (General Accounting Office
1998), as do police officers in charge of laws
without moral consensus and vague-defined
laws (Knapp Commission 1972).
3573
P
The third group of studies analyzes the relation between the characteristics of a specific
police agency (“rotten barrels” or “rotten
orchards,” Punch 2009) and police corruption.
This approach argues that police agencies have
the dominant role in addressing police misconduct by creating systems that establish rules,
enforce rules, detect corruption, and control the
code of silence (e.g., Klockars et al. 2000;
Kutnjak Ivković 2005; Sherman 1974; 1978).
The police chief and his top administrators have
critical
roles
(e.g.,
Goldstein
1975;
Knapp Commission 1972; Kutnjak Ivković
2005; Pennsylvania Crime Commission 1974),
while the roles of first-line supervisors
(e.g., Knapp Commission 1972; Mollen Commission 1994) and peers (e.g., Chen 2003; Klitgaard
1988; Kutnjak Ivković 2005; Stoddard 1974) are
not negligible either.
The fourth group of studies examines the
relation between the society at large and police
corruption. The police agency is part of the larger
environment and is influenced by the legal norms
(e.g., Knapp Commission 1972) and public
expectations (e.g., Goldstein 1975; Sherman
1977). Sherman (1977) proposed that communities differ greatly in their expectations, from
“communities with a more public-regarding
ethos” (like Charlotte, North Carolina, Kansas
City, Missouri, and Portland, Oregon) to communities with more “private-regarding” (like
New York City and New Orleans, Louisiana).
The same argument could be used to explain
differentiation in corruption rates across the
world; in the study of International Crime Victim
Surveys, Kutnjak Ivković (2003, p. 612) reveals
that “the countries with the reputation in the
international business community of being
more corrupt, as indicated by a low score on the
1999 Corruption Perception Index (CPI) . . .
appear also to have a higher percentage of the
respondents who said that they had been asked to
pay a bribe to a police officer last year.”
Measuring the Extent of Police Corruption
The measurement of the extent of police
corruption could be attempted at different levels
(Fig. 1).
P
P
3574
Police Corruption,
Fig. 1 The funnel of police
corruption and the data
collection methods. source:
Kutnjak Ivković 2003
Police Corruption
Offenders
sent to
prison
• Prison records
• Court records
• Prosecutor’s records
Offenders
sentenced
• Court records
• Prosecutor’s records
Offenders prosecuted
• Prosecutor’s records
Offenders referred for
prosecution
• Prosecutor’s records
Offenders known to internal
formal system of control in
the police agency
Actual extent and nature of corruption
The data sources seeking to assess the actual
level of police corruption include citizen and
police officer surveys. The results of citizen
surveys portray a heterogeneous picture of the
extent of police corruption across the country;
in the 1960s, fewer than 2 % of Caucasian respondents nationwide perceived that most of the
police were corrupt (President’s Commission on
Law Enforcement and Administration of Justice
1067b), in contrast to 93 % of New Yorkers in the
1990s who perceived corruption to be widespread
(Kraus 1994).
The public opinion surveys indicate that, on
the one hand, the public tends to have a relatively
positive opinion about police honesty, while, on
the other hand, the public perceives that police
officers frequently engage in corruption. A 1987
survey of Philadelphia citizens (Moore 1997,
p. 62) revealed that the public provided very
positive ratings of the police service, although
one-third of the respondents thought that police
officers often took bribes. Similarly, a 1994
• Complaints
• Disciplinary records
• Surveys: police officers,
payers, citizens, experts
• Observations
• Interviews
• Case studies
• Investigations
survey of New York citizens showed that, while
93 % of the surveyed citizens perceived that
corruption is either “widespread” or “limited,”
about one-half of these same citizens estimated
that the police are doing a “good” or an
“excellent” job (Kraus 1994).
On the comparative front, the International
Crime Victimization Survey asked the respondents whether they have paid a bribe and, if so,
who the recipient of the bribe was. About one
percent of the respondents or fewer from Western
democracies, including the USA, reported paying
a bribe to the police, while the corresponding
percentages were dramatically higher (between
10 % and 20 %) in some East European, Asian,
and Latin American countries (Kutnjak Ivković
2003). In addition, the Gallup International 50th
Anniversary Survey (1996) reports that approximately one-third of the respondents in the West
European countries and Israel, and more than
two-thirds of the respondents in the East European countries, the Far Eastern countries, and the
Police Corruption
Central and South American countries assessed
that police corruption was widespread in their
countries.
According to the results of the police officer
surveys, it seems that the perceived frequency of
occurrence of corruption was negatively related
to perceived severity of corruption. In the 1994
Illinois study (Martin 1994, p. 33), fewer than
0.4 % of the respondents said that they saw
a police officer accepting a bribe, stealing property, or purchasing stolen merchandise in the past
year, while 81 % of the police officers said that
they saw a police officer accepting free coffee or
food from a restaurant. Similarly, Ohio police
officers participating in the 1996 survey
(Knowles 1996) reported observing serious
types of police corruption (acceptance of
a payment to overlook illegal activity, purchase
of stolen merchandise for personal use or gain)
very infrequently (less than 0.6 % observed it last
year; less than 4.7 % observed it during their
careers), while they reported observing police
officers accepting free coffee or food from
restaurants quite frequently (71 % observed it
over the course of the last 12 months; 87.3 %
observed it during their careers).
Surveys of citizens and police officers alike
are burdened with the methodological problems
(see, e.g., Kutnjak Ivković 2003; National
Research Council 2004, p. 269). For various reasons, neither police officers nor citizens (participants or potential witnesses) have motives to
report a corrupt transaction; in fact, they have
motives not to report it (e.g., Klockars et al.
2000; Kutnjak Ivković 2003; Kutnjak Ivković
2005; Stoddard 1974). In addition, most of the
existing surveys focus on one country and are
conducted locally, usually with the purpose of
surveying the population of a particular city.
Another source includes independent commission report. While providing the results of an indepth investigations, these independent commissions are limited because they refer to a specific
police agency, time period, and are bound by the
resources and powers granted to the commission.
The Knapp Commission (1972) and the Pennsylvania Crime Commission (1972) find widespread
corruption in the NYPD and Philadelphia Police
3575
P
Department, respectively, and the presence of
a strong code of silence. In contrast, the Mollen
Commission (1994) reports that most police officers in the NYPD are honest but still found
pockets of police officers aggressively seeking
opportunities to obtain money and drugs.
A potential source of information about corruption includes field studies of the police.
Typically, they focus on a small number of police
agencies and involve a combination of methods
(e.g., observation, interviews, analyses of
documents). The President’s Commission on
Law Enforcement and the Administration of
Justice made a push for field studies of the police
in the 1960s. Reiss’s study of the police (1971),
originating as one of the studies submitted to the
President’s Commission on Law Enforcement
and the Administration of Justice, also contained
the data about the frequency with which police
officers engage in police misconduct. The rates of
corrupt behavior per 100 police officers, based on
the observations by the researchers, self-reports
by police officers, and allegations of misconduct
by others (calculated from the data provided in
Reiss’s study 1971) were 22.8 in City X, 20.5
in City Y, and 15.6 in City Z (see Kutnjak
Ivković 2003).
Another level of measuring corruption is
through the official data, be it arrest rates or the
complaint rates. At the federal level, there were
between 83 and 150 officers convicted annually
in the period from 1993 to 1997 (General
Accounting Office 1998, p. 11). A comparison
of the official data with the reports by independent commission implies that official data may be
just the tip of the iceberg. When the
Knapp Commission (1972) reported widespread
corruption n the NYPD, the prosecutors filed
charges in only about 30 cases of corruption
annually (Kutnjak Ivković 2003). Similar problems could be expected for the police agency’s
internal records of corruption complaints. At the
same time when the Knapp Commission (1972)
discovered widespread corruption in the NYPD,
the complaint rate in the NYPD was less than
1 per 100 officers.
The most novel approach is to measure the
extent of police integrity instead. Developed by
P
P
3576
Klockars and Kutnjak Ivković (Klockars et al.
2000), it measures the level of police integrity
and avoids the methodological problems associated with the direct measurement of police
corruption. A study of 30 police agencies
(Klockars et al. 2000) revealed that police
agencies varied considerably in the contours of
their police integrity.
Related Entries
▶ Control of Police Misconduct
▶ Law of Police Searches
▶ Law of Police Seizures and the Exercise of
Discretion
▶ Law of Police Use of Force
▶ Police and the Excessive Use of Force
▶ Police Culture
▶ Police Discretion and Its Control
▶ Police Integrity
▶ Police Lying and Deception
Recommended Reading and References
[Mollen Commission] New York City Commission to
Investigate Allegations of Police Corruption and the
Anti-Corruption Procedures of the Police Department
(1994) Commission report. Mollen Commission,
New York
Barker T, Carter DL (1986) A typology of police deviance. In: Barker T, Carter DL (eds) Police deviance,
3rd edn. Anderson Publishing, Cincinnati
Barker T, Roebuck J (1973) An empirical typology of police
corruption. Charles C. Thomas Publisher, Springfield
Barker T, Wells RO (1981) Police administrators’
attitudes toward the definition and control of police
deviance. FBI Law Enforc Bull 51(3):8–16
Bracey DH (1995) Police corruption. In: Bailey WG (ed)
The encyclopedia of police science, 2nd edn. Garland,
New York, pp 545–549
Carter DL (1990) Drug-related corruption of police officers:
a contemporary typology. J Crim Justice 18(2):85–98
Commission to Investigate Allegations of Police Corruption and the City’s Anti-Corruption Procedures
(1972) Knapp Commission report on police
corruption. G. Braziller, New York
Crank JP, Caldero MA (2000) Police ethics: the corruption
of noble cause. Anderson Publishing, Cincinnati
Gallup International (1996) Gallup international 50th
anniversary survey. Unpublished manuscript. On file
with the author
Police Corruption
General Accounting Office (1998) Law enforcement:
information on drug-related police corruption. General
Accounting Office, Washington, D.C
Goldstein H (1975) Police corruption: a perspective on its
nature and control. The Police Foundation,
Washington, D.C
Ivković SK (2003) To serve and collect: measuring
police corruption. J Crim Law Criminol 93(2–3):
593–649
Ivković SK (2005) Fallen blue knights: controlling police
corruption. Oxford University Press, New York
Kania RR (1988) Should we tell the police to say ‘Yes’ to
gratuities? Crim Justice Ethics 7(2):37–49
Kania RR (1994) Invalidating the ‘Dowd Test’. Ethics
Roll Call 2(1):2–4
Kania RR (2004) The ethical acceptability of gratuities:
still saying “Yes” after all these years. Crim Justice
Ethics 1:54–63
Klitgaard R (1988) Controlling corruption. University of
California Press, Berkeley
Klockars CB, Ivković SK, Harver WE, Haberfeld MR
(2000) The measurement of police integrity. Research
in Brief. U.S. Department of Justice, Office of Justice
Programs, National Institute of Justice. Government
Printing Office, Washington, DC
Klockars CB, Ivković SK, Haberfeld MR (eds) (2004) The
contours of police integrity. Sage, Newbury Park
Knowles JJ (1996) The Ohio police behavior study. Office
of Criminal Justice Services, Columbia
Martin C (1994) Illinois municipal officers’ perceptions of
police ethics. Illinois Criminal Justice Information
Authority, Statistical Analysis Center, Chicago
Moore M (1997) Epilogue, Police integrity: public service
with honor. U.S. Department of Justice. Available at
https://www.ncjrs.gov/pdffiles/163811.pdf
Muir W (1977) Police streetcorner politicians. University
of Chicago Press, Chicago
National Research Council Committee to Review
Research on Police Policy and Practices (2004) Committee on Law and Justice, Division of Behavioral and
Social Sciences and Education. In: Skogan W, Frydl
K (eds) Fairness and effectiveness in policing: the
evidence. The National Academies Press: Washington
Pennsylvania Crime Commission (1974) Report on Police
corruption and the Quality of Law Enforcement in Philadelphia. Pennsylvania Crime Commission, Saint Davids
President’s Commission on Law Enforcement and
Administration of Justice (1967) Task force report:
the police. GPO, Washington, D.C
Punch M (1985) Conduct unbecoming. Tavistock, London
Punch M (2009) Police corruption: deviance, accountability and reform in policing. Willan Publishing, Portland
Reiss AJ Jr (1971) The police and the public. Yale University Press, New Haven
Roebuck JB, Barker T (1974) A typology of police corruption. Soc Probl 21:423–437
Sherman LL (1974) Becoming bent: moral careers of
corrupt policemen. In: Sherman L (ed) Police corruption. Anchor Press, Garden City
Police Culture
Sherman LL (1977) Police corruption control. In: Bayley
D (ed) Police and society. Sage, Thousand Oaks
Sherman LL (1978) Scandal and reform. University of
California Press, Berkeley
Stoddard ER (1974) A group approach to Blue-Coat
crime. In: Sherman LW (ed) Police corruption:
a sociological perspective. Anchor Press, Garden City
Police Culture
Eugene A. Paoline III
Department of Criminal Justice, University of
Central Florida, Orlando, FL, USA
Synonyms
Occupational culture; Organizational culture;
Police styles; Police subculture
Overview
The job of a police officer can be both mentally
and physically exhausting. As such, officers rely
on one another for emotional and physical
support. Terms like brotherhood, thin blue line,
and blue code of silence are common illustrations
of the cultural bond that officers share. Most often
connotations of police culture (or police subculture to some) are negative. For example, when
officers misuse their coercive power, it is often
police culture that gets blamed (for endorsing
a no nonsense aggressive approach to dealing
with citizens). When organizational reforms,
like community policing, are met with officer
resistance, the culture is cited as the primary
barrier to soft policing innovations brought on
by “out of touch” administrators. At the same
time, police culture can operate as a powerful
positive mechanism in helping officers mutually
buffer the various tensions of their job.
Unfortunately, a great deal of vagueness
comes with the concept of police culture. That
is, if you asked people to define and explain it,
you are liable to receive 100 different responses
from 100 different people. The most popular
3577
P
depiction of police culture views it as an
occupational phenomenon shared by all police
no matter where (or for whom) they work. As
such, there is more of a focus on the commonalities across police personnel over the differences.
Even among those who endorse the occupational
view, there is a great deal of variation in describing the facets of police culture. Cumulatively,
this is problematic because it limits our ability
to fully understand, and comprehensively
research, this important aspect of policing.
In illustrating the ways of conceptualizing
police culture, consider the occupation of criminal justice professor. Undoubtedly, across the
various universities that house professors, there
are commonly shared cultural ways of handling
the work-related strains created by students as
primary clientele (e.g., class size and preparation,
advising, grade appeals, plagiarism) and administrative supervisors (e.g., expectations for peerreviewed publications, procuring funded
research, service). At the same time, it is reasonable to expect that such demands, and subsequent
ways of coping, vary depending on the type of
institution where one works (e.g., large researchoriented university versus small private teaching
college). Moreover, these strains might be met
differently based on the working style of the
professor (e.g., pro-research versus proinstructor). Even among criminal justice professors housed within the same institution, concerns
and demands might be interpreted differently
across the ranks of assistant, associate, and full
professor. Relatedly, for those in the upper tiers,
cultural variation might also be a function of
changing assignments (e.g., research coordinator,
departmental chair, associate dean, dean). One’s
background (e.g., gender, race, educational
experiences) might also work to produce
differences in the ways in which criminal justice
professors deal with the pressures of their work. If
one could effectively tease out the various
influences, across these divergent conceptualizations, it would undoubtedly produce a deeper
comprehension of criminal justice professor
culture(s). A similar approach is taken in this
chapter in understanding police culture.
P
P
3578
The Occupational Culture of Policing:
The Monolithic Model
The foundation for understanding police
occupational culture dates back to the seminal
work of William Westley (1970) who, in the
1950s, conducted one of the first sociological
studies of policing. In doing so, Westley
highlighted important themes of the police
occupation. More specifically, he found that the
police he studied were a secretive tight-knit
group that often faced violence on the streets in
performing their duties. In detailing these informal aspects of policing, Westley was painting the
first picture of its kind. In essence, the message
delivered was that cops believed they have to
stick together in their interactions against
potentially dangerous and hostile citizens.
The other significant contribution to our initial
impressions of police culture was the research of
Jerome Skolnick (1966), who explained how the
occupation actually worked to shape a distinct
police personality. Similar to Westley (1970),
Skolnick detailed life on the streets for the police
officer as a dangerous endeavor, one where
officers wield a tremendous amount of authority
over citizens. Skolnick also tapped into a third
primary feature that helped form the police personality, the need to appear efficient to superiors,
which introduced an equally dangerous and hostile work environment – the police organization.
Both of these ethnographic accounts of police
work highlighted many of the informal aspects of
the job while also providing a loosely defined
template for scholars in comprehending the
occupational culture of police. More specifically,
these two studies started the process of thinking
about the ways in which police officers collectively cope, via their attitudes, values, and norms,
with the strains created by their work – which is
precisely what defines an occupational culture
(Paoline 2001). Unfortunately, unlike foundational explanations of crime causation that tended
to clearly (and parsimoniously) posit sets of interrelated propositions, full accounts of police occupational culture have been as vague as the
concept of love. That is, you tend to know it
when you see it but can mean different things to
Police Culture
different people. From a theoretical and empirical
standpoint, this can be both confusing and
frustrating.
The works that followed Westley (1970) and
Skolnick (1966) described and analyzed several
different components of culture that tapped
various dimensions of the street and organizational environments. What was lacking though
was a clear comprehension of how these individual elements, that research identified, contributed
to an overall understanding of the police culture
process. That is, what exact features are police
dealing with in their primary work environments
that are in need of occupational collectiveness? In
addition, what are the so-called collective
responses that officers utilize to manage the
strains of these environments? Finally, what are
the consequences of the strains of the work environments and the coping mechanisms that are
used to deal with such concerns? While the elements for such an understanding were present in
the volumes of research (e.g., an explanation of
police loyalty, the crime fighting mandate, citizen
danger), a concise roadmap of these correlates, in
explaining exactly what a police culture entails,
was not readily existent. The monolithic model of
police culture, presented by Paoline (2003),
helped clarify these concerns.
Paoline’s (2003, p. 201) description of the
monolithic occupational culture of policing
presents, based on extant research, a way to
conceptualize the stressful factors of the work
environments that operate to produce collective
coping mechanisms and outcomes. In a path-like
presentation, he identifies interactions on the
street with citizens (i.e., occupational) and those
with supervisors in the department (i.e., organizational) as the two primary work environments
of the police. As a starting point for understanding what produces the coping mechanisms of the
monolithic culture, the occupational street
environment is described as dangerous, with the
defining mandate to display one’s coercive
authority over citizens. The model also details
the organizational environment as being characterized by uncertain supervisor scrutiny of police
decisions (i.e., watchful and punitive superiors)
and role ambiguity whereby officers are expected
Police Culture
to perform all police functions equally yet really
only recognized for crime fighting duties. Both
the occupational and organizational environments produce equally intense stress and anxiety
that is relieved through the collective coping
mechanisms found within the occupational police
culture.
In illustrating the second principal feature of
the monolithic police culture model, Paoline
(2003) identifies suspiciousness (in dealing with
danger) and maintaining the edge (in properly
displaying coercive power) as the primary coping
mechanisms that officers employ in handling the
strains created on the street with citizens. In terms
of dealing with supervisors within the equally
hostile organizational environment, the culture
prescribes that officers lay low/c.y.a. (from
watchful and punitive superiors) and embrace
the crime-fighter orientation (in minimizing role
ambiguity). These prescriptive coping mechanisms of the police occupational culture are transmitted across officers via a socialization process
that begins in the training academy and continues
throughout one’s career (Van Maanen 1974). The
final stage of the model highlights the consequences of the strains of the work environments
and the coping mechanisms prescribed by the
culture, which include a socially isolated
occupational group that is extremely loyal to
one another.
Collectively, the work that helped produce this
model paints a caricature sketch of police as
a socially isolated group that are distrustful and
suspicious of their primary clientele, as they
continually attempt to maintain the upper hand in
utilizing their coercive authority. Moreover,
officers approach police work solely in crime fighting terms while laying low from supervisors,
choosing to only trust their immediate peers.
While this is the dominant portrayal of police culture, other research points to important ways in
which the occupational group may be fragmented.
Sources of Variation in Police Culture
The preceding discussion regarding a police
culture rests on the assumption that the strains
3579
P
that police face on the streets and within the
department are the same across the United States
and therefore officers’ responses should be
similar as well (Crank 1998). By contrast, there
are other lines of research that cast doubt on the
notion of homogeneity among police by pointing
out important sources of cultural variation and
segmentation.
Organizations
Although often used synonymously, occupational and organization cultures are not the same
phenomenon. As opposed to a universally shared
set of responses to the internal and external
strains of the job (for all police) that make up
occupational accounts of culture, organizational
cultures represent ways that police deal with
specific concerns from the various places where
they are housed. A principal difference between
occupational and organizational accounts of
police culture lies within the locus of influence.
Occupational cultures are usually formed and
maintained
by lower
level
personnel
(Van Maanen and Barley 1984), while organizational cultures are created by upper level management and imposed downward through the
ranks (Schein 1992). Irrespective of such
nuances, this suggests that police organizations,
embedded within the overall occupation, exert
cultural influence over police officers. James
Q. Wilson’s (1968) seminal study accentuates
such points.
Wilson’s (1968) examination of organizational culture, which he believed was indicative
of a departmental style, was based on research
conducted in eight communities. Wilson
identified three organizational styles that differed
in terms of their priorities toward core role orientations of enforcing the law and maintaining
order. The author asserted that it was the top
police administrator that defined the given style
of the department based on their interpretation
of the primary needs of the community being
served.
According to Wilson (1968), some departments are situated within urban environments
characterized by greater crime concerns, and
thus the agency will embrace more of a formal
P
P
3580
crime fighting detached approach to police work.
In such legalistic style departments, arrests and
tickets will be more frequent compared to other
organizational styles. By contrast, Wilson’s
watchman style departments are more common
in low crime rural areas, focusing primarily on
maintaining public order. Watchman style
departments are less likely to formally respond
to citizen transgressions unless they are serious in
nature. Wilson asserted that the relational
distance between the police and the public in
watchman areas would be much smaller than
that found in legalistic style departments. Finally,
service style departments are described by
Wilson as those where law enforcement and
order maintenance is not an overall priority, but
would be handled as needed. These departments
focus on providing assistance and are more likely
to be situated in suburban areas with less crime
and disorder. Wilson explained that service style
departments are closely connected with citizens,
choosing to intervene frequently (when needed)
but not formally (i.e., arrests and tickets).
Wilson’s (1968) research suggests that organizational environments vary, and as such, it highlights the fact that individual police responses to
different work conditions might also vary. Interestingly, the legalistic style department he identified comports closely to the organizational
environment found in explanations of the monolithic occupational culture. By contrast, the watchman
and
service
style
organizational
environments represent vastly different internal
work arenas where intense supervisor scrutiny
and role ambiguity (noted in the occupational
account) would be much less common. Likewise,
Wilson’s work also points out rather stark variation in the occupational/street environments, as
watchman and service style areas are not characterized as overly hostile and crime ridden. As such,
cumulative coping mechanisms where officers are
suspicious of (and maintain the edge over) citizens, strictly endorsing the crime-fighter image,
while covering their ass from supervisors might
be functional in a legalistic style department but
certainly would be out of place (and probably not
tolerated) in both watchman and service style
agencies.
Police Culture
Recent empirical inquiries have provided
additional support for the ways in which policing
functions and philosophies differ across organizations embedded in urban, rural, and suburban
contexts (Crank 1990; Liederbach 2005). This
line of research casts doubt on occupational
accounts of culture that suggest that all organizations are the same while also acknowledging that
management can impact police culture.
Rank
Occupational accounts of police culture focus
heavily on the homogeneity of officers’ attitudes,
values, and norms. The socialization process
transmits the culture across occupational
members and is often described as intense, especially for new personnel (Van Maanen 1974).
Because all police at some point in their career
are assigned to patrol functions, the supposition is
that culture originates at the lower ranks. This is
understandable given that patrol officers are those
most likely to deal with citizens on the streets and
with supervisors in the department. What is much
less clear is the role that culture plays in officers’
lives once they move beyond entry level positions. Does the occupational culture still buffer
the strains of the work environments for sergeants, lieutenants, captains, majors, deputy
chiefs, and chiefs the same way(s) that it did
when these officers were assigned to patrol, or
does culture change by rank? The works of
Reuss-Ianni (1983) and Manning (1994) help
answer such questions.
Reuss-Ianni (1983), based on research
conducted in the NYPD, asserts that there are
two distinct cultures in policing. The street cop
culture, which Reuss-Ianni delineates in a series
of codes, concentrates on the “here and now” in
policing and embodies many of the values
outlined previously as part of the police occupational culture. Street cops are found at the patrol
level and tend to rely on their developed craft, as
well as their loyal peers, in surviving on the street
and controlling crime at the local level (i.e., beat
and precinct). By contrast, management cop
culture focuses on city-wide long-term concerns
(e.g., crime control, citizen responsiveness,
organizational efficiency), taking into account
Police Culture
political, social, and economic factors. Importantly, Reuss-Ianni believes that temporal
changes in the dynamics of policing (e.g., officer
composition, resource competition, accountability concerns) have contributed to a weakening of
the street cop culture of the “good old days” when
organizational leaders could still utilize the codes
of the street in running police departments.
The importance of Reuss-Ianni’s (1983) work
lies in her recognition that rank contributes to
differences in police culture. This makes sense
given that duties and concerns do change as one
moves upward from patrol with the organization.
This might also help explain some of the tension
noted between officers and their supervisors in the
occupational account of police culture. While it is
clear that street cop culture originates and protects
occupational members at the patrol level, the
author fails to inform the reader exactly when
police enter management cop culture. That is, it
is easy to deduce that top commanders are embedded in the management cop culture, but what about
ranks like sergeants and lieutenants (i.e., middle
managers), who are in organizational limbo
between the very bottom and the very top. Manning’s (1994) work helps alleviate such concerns.
Manning (1994) agrees that police culture is
hierarchically segmented, although he differentiates rank in a three-tiered manner based on
a series of themes and meta-themes. At the first
tier are lower participants (i.e., patrol and street
sergeants) whose culture, like Reuss-Ianni’s
(1983) street cops, focuses on the immediate
aspects of “real” police work. The author’s
second tier of culture focuses on middle management (i.e., some sergeants up to department brass)
who emphasizes supervisory control but also
concentrates heavily on buffering concerns of
line members of the street (i.e., the first tier) and
top police officials (i.e., the third tier). Finally,
top command culture (i.e., commanders, superintendents, deputy chiefs, chiefs) represents the
third tier. Here, there is a concentration on
dealing with the politics of running a police
department internally while also buffering the
organization from external audiences.
Interestingly, in terms of middle managers,
Manning (1994) points out that sergeants
3581
P
assigned to street functions can rest within the
first tier (i.e., lower participant culture), while
middle managers with other responsibilities
beyond patrol may be part of the second tier.
Here, the author allows for behavioral freedom
for sergeants who, as supervisors, might still
align more with the line level versus management. Although not addressed by Manning, the
same argument can be made for first tier lower
participants, who depending on their assignment
(or aspirations) might differentially align with
police culture(s). For example, those who have
jobs with duties that are based on more traditional
aggressive crime fighting mandates (e.g., criminal investigations, tactical, K-9, anti-street
crimes, SWAT, emergency response units)
might be more apt to orient themselves with the
lower participant culture. By contrast, those in
positions that focus more on long-term solutions
to controlling crime with increased officer
participation (e.g., crime prevention, community
outreach, school resource, community policing,
victim advocate, homeland security, tourism
units) might orient themselves more like upper
ranking officials.
Both Reuss-Ianni (1983) and Manning (1994)
highlight the dynamic nature of police culture as
changing and adapting to insulate group
members’ issues and concerns unique to their
position in the organizational hierarchy. As
such, their work suggests that police culture is
not monolithic, but instead comprised of multiple
cultures based on rank.
Style
Yet another way that police culture can be
segmented is by the style of the officer. In
a series of studies published in the 1970s,
researchers constructed typologies of the police
based on their orientations toward various
features of their occupational/street and
organizational environments (Broderick 1977;
Brown 1988; Muir 1977; White 1972). Of interest
is the fact that these researchers were utilizing
different dimensions to construct their typologies
and were working independent of one
another across various departments, areas, and
time – yet they concluded with almost identical
P
P
3582
types/styles of officers. This was evidenced by
Worden’s (1995) synthesis of American police
typology research into five distinct officer styles.
Reiner (1985) also summarized typology
research, utilizing some of the studies that
Worden (1995) included (i.e., Broderick 1977;
Muir 1977), although he incorporated work
done on British and Canadian police. Interestingly, both authors concluded with very similar
syntheses of officer types.
Worden (1995) provides a thumbnail sketch of
the five primary styles of police officers.
Worden’s tough-cop approaches police work in
a cynical fashion, often conflicting with citizens
and supervisors in their aggressive approach to
selectively fighting serious crime. These officers
do not want to be bothered with trivial police
matters (e.g., maintaining public order or providing service), and they place a premium on their
street experience over formal college education.
This style of officer would be your stereotypical
gruff “cop’s cop” that is often portrayed in
television and movies, as one that is not afraid
to push (or perhaps even exceed) the limits of
police power if deemed necessary.
A second style, clean-beat crime-fighter,
resembles the tough cop in their cynical
approach, strong orientations toward crime
fighting, and conflicting relationships with supervisors. What separates this style of officers from
the former is their undying pursuit of enforcing
all laws, not just the serious ones (in keeping
a clean beat), and their belief that police should
follow the procedural rules (i.e., not violate citizen rights).
Avoiders are described as a cynical group that,
as their name implies, avoid as much work as
possible. This style of officer has a very narrow
approach to police work, as they take “laying
low” to extremes. Their orientations toward all
aspects of the job are rather detached, as they
attempt to just do their time.
Problem solvers are those officers with very
favorable attitudes toward citizens and the
service style of policing. Problem solvers are
the group that is least oriented toward crime
fighting and aggressive policing tactics. As
opposed to traditional ways of handling
Police Culture
situations, problem solvers focus on outcomes,
desiring to see the problems they are called
upon to deal with through to their resolution. To
many traditional police officers, the approaches
endorsed by the problem solver would appear as
“soft.”
Professionals embody the values of the
professional reform movement and are characterized as the most positive of the five policing
styles. Officers in this group hold favorable orientations toward all aspects of their street and
organizational environments. Professionals are
portrayed as the most well-rounded group
in terms of performing multiple functions
(i.e., broad role orientation) in a manner in
which citizens are not treated aggressively and
where supervisors are pleased.
Police typology research illustrates that there
are a variety of ways in which officers cope, via
their style, with the demands of their job.
Interestingly, Worden’s (1995) tough-cop
orientation comports closely to the attitudes and
values associated with occupational accounts of
culture, as well as Reuss-Ianni’s (1983) street cop
culture and Manning’s (1994) lower participant
culture. The other four styles summarized by
Worden suggest the type of ideological differentiation that forms the basis for occupational
subcultures (Van Maanen and Barley 1985).
While research has yet to empirically validate
these working styles, one thing is clear – there is
more than just a single way for officers to deal
with the strains of policing.
Officer style research illustrated cultural
fragmentation among police during a time when
officers were relatively demographically homogenous (i.e., White males with a high school
education). Changes in the overall composition
of police (i.e., more females, non-Whites, and
college educated), as well as policing philosophies (i.e., community policing), in recent
decades prompted reexaminations of police
culture(s) (Paoline et al. 2000). Moreover, recent
empirical approaches have utilized advanced
statistical classification techniques, such as
cluster analysis and discriminant function
analysis, to examine more than the two or three
attitudinal dimensions that were used for
Police Culture
previous police typology inquiries (Cochran and
Bromley 2003; Jermier et al. 1991; Paoline
2001). Similar to the research that produced evidence of policing styles, studies that utilized
quantitative classification schemes also found
multiple groups of officers with varying
alignment to a single occupational culture.
Because much of this (and typology) research is
based primarily on officers’ attitudinal orientations, it begs the question as to the applicability of
such approaches on explanations of officer
behavior. The age old attitude-behavior link has
stymied social scientists for decades, as empirical
connections have failed to verify intuitive
expectations that one’s attitudes affect one’s
behavior (Frank and Brandl 1991). Paoline’s
(2001) classification scheme provides one of the
few exceptions.
Paoline (2001) utilized survey data collected
as part of the Project on Policing Neighborhoods
(POPN) to analyze several prominent attitudinal
features of prior police culture research. Paoline
identified seven groups of officers with varying
cultural orientations. Five of the groups very
closely resembled those Worden described in
the framework he presented in 1995: tough cops
(i.e., traditionalists), clean-beat crime-fighters
(i.e., law enforcers), avoiders (i.e., lay lows),
problem solvers (i.e., peacekeepers), and
professionals (i.e., old pros). The other two
officer groups were anti-organizational street
cops, who were distinguished in terms of their
strong negative views of supervisors and their
strong favorable views of citizens, and dirty
harry enforcers, who were distinguished in
terms of their strong beliefs in aggressively
fighting crime and disorder, even if it meant violating the rights of citizens.
Although Paoline’s (2001) seven groups were
attitudinally and empirically distinguishable
from one another, there was no genuine way to
rank order them (for comparison purposes) in
terms of a continuum of police culture commitment. The groups did provide the author with
a conservative manner in which to categorize
them in terms of positive orientations toward
culture (i.e., pro culture), negative outlooks or
antithesis of cultural expectations (i.e., con
3583
P
culture), and those that fell somewhere in the
middle of the two extremes (i.e., mid culture).
Paoline’s trichotomized classification scheme of
police has been utilized to empirically examine
the connection between cultural attitudes and
theoretically related behaviors of citizen searches
(Paoline and Terrill 2005) and the use of force
(Terrill et al. 2003). In both instances, statistically
related differences were noted between officers’
cultural alignment and their behavior. More
specifically, those who held attitudes that were
strongly congruent with the traditional police
culture (i.e., pro culture) or mixed (i.e., mid
culture) searched citizens more often during
traffic stops and used force more frequently (and
at higher levels) than those who ardently resisted
cultural attitudes (i.e., con culture).
Background Characteristics
Besides the organization, rank, and policing
style, officer background represents another
potential source of cultural variation. As
previously mentioned, the foundation for understanding the monolithic occupational account of
police culture was built during a time when
policing was largely a demographically homogenous occupation (i.e., made up almost exclusively of high school-educated White males).
Researchers have detailed the integration struggles and resistance that female (e.g., Martin and
Jurik 1996), non-White (e.g., Bolton 2003), and
gay/lesbian (e.g., Miller et al. 2003) officers have
faced in a male-dominated masculine occupation.
To a lesser extent, college-educated officers have
been chided for their “book smarts” over the
preferred “street smarts,” in learning the craft of
policing (Fielding 1988). The idea here is that
those with different backgrounds from the
modal officer are excluded from the broader
occupational culture.
As departments continue to diversify their
personnel in terms of their individual characteristics (Hassell and Brandl 2009), it is certainly
reasonable to expect that these previously
excluded policing members will, in some
manner, contribute to a less cohesive occupational group. To date, we lack a concrete understanding of how such advancements in officer
P
P
3584
diversity affect police culture(s). That is, despite
qualitative (and detailed case study) accounts of
cultural dissociation, background characteristics
of police provide limited power in explaining
cultural perceptions among police (Paoline et al.
2000). In fact, the impetus for some of the recent
officer classification schemes, in revisiting
working styles, was built on the notion that
diversification of personnel should produce
differences in cultural alignment (Paoline 2001).
Even among these studies that utilize advanced
statistical techniques with larger sample sizes,
there is little (if any) systematic connection
between officer background characteristics and
their group membership. In the end, such work
suggests that cultural variation may have little to
do with one’s background.
The lack of a strong empirical connection
between officer characteristics and alignment
with police culture(s) could be a function of
a couple of factors. It could be that demographic
changes that are occurring are still too recent, and
thus, we have yet to see the full cultural effects of
such transitions in membership. A second
explanation could be that the forces of cultural
socialization as so strong, as some traditionalists
would assert (Van Maanen 1974), that individual
differences (via one’s sex, race, education, sexual
orientation, etc.) are washed away as officers
collectively deal with the strains of the
occupation. As such, the primary environments
of policing are what shape the culture or
subcultural styles, irrespective of what officers
bring to the job.
Conclusion and Future Research
Police culture has been a topic of study for over
half a century. Although the monolithic depiction
still tends to dominate popular conceptions,
this chapter illustrates that police culture is
a multidimensional concept that is molded and
shaped at various levels (i.e., organization, rank,
style, assignment, background). Police scholars,
in attempting to understand this police phenomenon, have wrestled with such complexities. At the
same time, police culture is not just an academic
Police Culture
concept, as practitioners recognize its power,
vagueness, and malleability. As noted by the
Community Relations Service of the United
States Department of Justice’s (2002, p. 9)
handbook on “Police Use of Excessive Force,”
in order to change police culture, “one must
analyze and understand the currently existing
culture.” While this chapter summarizes many
of the advancements in detailing the pathways
and contours of police culture, much more work
is needed. As Fielding (1988, p. 185) accurately
contends “if occupational culture is to serve as an
empirically satisfactory concept as well as
theoretically necessary one, the sense of its internal variations and textures must be brought out in
the same fashion as have conceptions of culture
in relation to delinquency.” What follows are
a few suggestions for future research inquiries.
Ideally one would want a totally comprehensive study, which would require starting at the
occupational level and working through differences between and among organizations, ranks,
styles, assignments, and backgrounds. Simply
put, this may be asking way too much at this
point. A more reasonable request starts by
considering the various sources of influence on
police culture.
The monolithic account of the police culture
was created based on early ethnographic studies
interested in capturing themes across officers. It
is not surprising then that cultural similarities
were the focus over cultural segmentation. Even
if we acknowledge the sources of cultural
variation, typology (and recent officer classification scheme) research did find one style of officer
that deals with the strains of the primary work
environments in the same manner presented by
monolithic characterizations. Paoline (2003)
presented a model for understanding this
monolithic version of culture, and while parts of
it have been researched, the entire model has yet
to be empirically tested. As a logical starting
point, it would be interesting to gather baseline
data regarding the extent to which this traditional
culture model is endorsed among contemporary
police. This is especially salient given that
policing philosophies are revisiting more bottom
line crime fighting goals (e.g., COMPSTAT) at
Police Culture
the cost of many community policing initiatives
that have permeated agencies for the last 20
years. At the occupational level, we still do not
have a firm empirical grip on the things that
officers might share versus the exact points
where they might be divided.
Regarding the influence of the organization,
researchers should continue efforts aimed at
examining departmental style as it impacts police
culture(s). In doing so, empirical work can be
done to deduce whether organizational environments, via agency style, are similar or dissimilar
from that noted by Wilson (1968). For example,
the rural and suburban areas that made up the
smaller watchman and service departmental
styles might be organized and operating in different ways today as criminogenic forces (e.g.,
gangs, unincorporated areas for producing and
distributing drugs) are not just reserved for legalistic urban areas. As such, the directives and goals
espoused as part of today’s watchman and service
organizational styles may look more similar to
the legalistic style than that noted in the 1960s.
With respect to officer styles, efforts should be
made to examine the connected (or disjointed)
way(s) in which individual level approaches to
dealing with the strains of policing operate within
organizational level, top command-driven, styles.
Jermier et al. (1991) found differences between
subcultural groups of officers and the legalistic
style crime fighting official organizational
culture. This begs the question – what about
other departmental styles and potential organizational survival of officer styles? For example, the
crime fighting approaches of Worden’s (1995)
“tough cop” and “clean-beat crime-fighter”
would certainly seem out of place compared to
a “problem solver” or “professional” in a service
style suburban department, as opposed to
a legalistic style organization where the former
would be more of a fit than the latter.
Regarding rank-related sources of variation,
disentangling officer assignment and formal
designation within the police hierarchy would
be useful in understanding how culture(s) differ
across Manning’s (1994) three primary tiers. For
example, can lower level participants, with
assignments that call for increased officer
3585
P
participation and long-term orientations, align
more like middle managers and their culture
than their own? Conversely, can captains that
are assigned to aggressive crime fighting street
duties orient themselves more like lower level
participants? In addition, do officer styles differ
across the three tiers or is the culture monolithic
by rank? If policing styles, as proxies for
subcultures, develop at the patrol level, how and
when do they change as one advances through the
organizational hierarchy? In what way(s) do
supervisory styles impact the development and
maintenance of subordinate styles? These are but
a few questions whose answers would add greatly
to our police culture knowledge base.
Finally, despite the lack of consistent statistical associations, the demographic changes that
have occurred in policing should continue to be
part of police culture research. In doing so, racial
(and ethnic) groups other than dichotomous
White and non-White (predominantly comprised
of African Americans) classifications should be
part of such inquiries (e.g., Hispanic, Asian,
Pacific Islander, American Indian, Alaskan
Native). As such, previously excluded groups
continue to populate police organizations, especially above the patrol level, cultural (and sub
cultural) differences might be more visible. The
same holds for those with divergent educational
backgrounds and sexual orientations that differ
from the traditional blue-collar aggressive crime
fighting officer depicted in characterizations of
the police culture “brotherhood.” Moving beyond
descriptive accounts of the struggles of those with
divergent backgrounds and focusing on quantitative multivariate modeling will enhance our
understanding of the potential independent
effects of such factors on police culture(s).
The individual pieces of the multidimensional
puzzle have been presented here in an attempt to
detail the various intricacies of police culture.
If scholars, practitioners, and funding agencies
are serious about the overall importance of
understanding (and possibly changing) police
culture, it is high time to devote the empirical
effort required to fully comprehend the way(s)
in which officers deal with their work
environments.
P
P
3586
Related Entries
▶ Communities and the Police
▶ Community Policing
▶ COMPSTAT
▶ Control of Police Misconduct
▶ History of the Police Profession
▶ Managing Innovation of Policing
▶ Minorities Within the Police Workforce
▶ Police and the Excessive Use of Force
▶ Police Corruption
▶ Police Discretion in Law Enforcement
▶ Police Integrity
▶ Police Legitimacy and Police Encounters
▶ Policing the Police
▶ Role and Function of the Police
▶ Women in Policing
Recommended Reading and References
Bolton K Jr (2003) Shared perceptions: black officers
discuss continuing barriers in policing. Pol Int
J Police Strateg Manag 26:386–389
Broderick JJ (1977) Police in a time of change. General
Learning Press, Morristown
Brown MK (1988) Working the street: police discretion
and the dilemmas of reform, 2nd edn. Russell Sage,
New York
Cochran JK, Bromley ML (2003) The myth(?) of the
police sub-culture. Pol Int J Police Strateg Manag
26:88–117
Crank JP (1990) The influence of environmental and
organizational factors on police style in urban and
rural environments. J Res Crime Del 27:166–189
Crank JP (1998) Understanding police culture. Anderson
Publishing, Cincinnati
Fielding NG (1988) Joining forces: police training, socialization, and occupation competence. Routledge,
London
Frank J, Brandl SS (1991) The police attitude-behavior
relationship: methodological and conceptual considerations. Am J Police 58:83–103
Hassell KD, Brandl SS (2009) An examination of the
workplace experiences of police patrol officers: the
role of race, sex, and sexual orientation. Police
Q 12:408–430
Jermier JM, Slocum JW Jr, Gaines J (1991) Organizational subcultures in a soft bureaucracy: resistance
behind the myth and facade of an official culture.
Organ Sci 2:170–194
Liederbach J (2005) Addressing the elephant in the living
room: an observational study of the work of suburban
police. Pol Int J Police Strateg Manag 28:415–434
Police Culture
Manning PK (1994) Police occupational culture: segmentation, politics, and sentiments. Unpublished
manuscript, Michigan State University
Martin SE, Jurik NC (1996) Doing justice, doing gender.
Sage, Thousand Oaks
Miller SL, Forest KB, Jurik NC (2003) Diversity in blue:
lesbian and gay police officers in a masculine occupation. Men Masc 5:355–385
Muir WK Jr (1977) Police: streetcorner politicians.
University of Chicago Press, Chicago
Paoline EA III (2001) Rethinking police culture: officers’
occupational attitudes. LFB Publishing, New York
Paoline EA III (2003) Taking stock: toward a richer
understanding of police culture. J Crim Justice
31:199–214
Paoline EA III, Myers SM, Worden RE (2000) Police
culture, individualism, and community policing:
evidence from two police departments. Justice
Q 17:575–605
Paoline EA III, Terrill W (2005) The impact of police
culture on traffic stop searches: an analysis of attitudes
and behavior. Pol Int J Police Strateg Manag
28:455–472
Reiner R (1985) The politics of the police. St. Martin’s
Press, New York
Reuss-Ianni E (1983) Two cultures of policing.
Transaction, New Brunswick
Schein EH (1992) Organizational culture and leadership.
Josey Bass Publishers, San Francisco
Skolnick JH (1966) Justice without trial: law enforcement
in democratic society. John Wiley, New York
Terrill W, Paoline EA III, Manning PK (2003) Police
culture and coercion. Criminology 41:1003–1034
U.S. Department of Justice (2002) Police use of excessive
force: a conciliation handbook for the police and the
community.
Community
Relations
Service,
Washington, DC
Van Maanen J (1974) Working the street: a developmental
view of police behavior. In: Jacob H (ed) The potential
for reform of criminal justice. Sage, Beverly Hills,
pp 83–130
Van Maanen J, Barley SR (1984) Occupational communities: culture and control in organizations. In: Staw BM,
Cummings CC (eds) Research in organizational behavior, vol 6. JAI Press, Greenwich, pp 287–365
Van Maanen J, Barley SR (1985) Cultural organization:
fragments of a theory. In: Frost PJ, Moore LF, Louis
MR, Lundberg CC, Martin J (eds) Organizational
culture. Sage, Beverly Hills, pp 31–53
Westley WA (1970) Violence and the police:
a sociological study of law, custom, and morality.
MIT Press, Cambridge, MA
White SO (1972) A perspective on police professionalization. Law Soc Rev 7:61–85
Wilson JQ (1968) Varieties of police behavior: the
management of law and order in eight communities.
Harvard University Press, Cambridge, MA
Worden RE (1995) Police officers’ belief systems:
a framework for analysis. Am J Police XIV:49–81
Police Discretion and Its Control
Police Discretion and Its Control
Gary Cordner1 and Michael S. Scott2
1
Department of Criminal Justice, Kutztown
University, Kutztown, PA, USA
2
Law School, University of Wisconsin,
Madison, WI, USA
Overview
Police officers exercise a tremendous amount of
discretion in carrying out their functions. That is,
they make many choices from a range of possible
actions or inactions available to them, and their
choices are not specifically prescribed by law,
policy, procedure, or rule. This simple notion,
which seems self-evident to some and controversial to others, lies at the heart of many issues of
policing in democratic societies.
That the police do exercise discretion was only
recognized and openly acknowledged beginning
in the 1960s (J. Goldstein 1960; H. Goldstein
1963; LaFave 1965; Davis 1969). The conventional view prior to that time, and persisting
among some long thereafter, was that the police
function was entirely a ministerial one: that the
police only took actions that were specifically
authorized or mandated by legislative bodies.
Under this view, policing was understood to be
simply a matter of enforcing the laws on the
books. But a number of pioneering research studies of policing in action found that the law was
silent on many important matters involving
police action and ambiguous on others, and
police officers did not always adhere to what the
law prescribed, even where the law was clear and
specific (Banton 1965; Skolnick 1966; Wilson
1968; Westley 1970; Reiss 1971; Brown 1981).
The exercise of discretion is nearly inevitable
in policing (H. Goldstein 1977). Some laws are
practically unenforceable because they are outdated, or widely unpopular, or unconstitutional,
or lack enforceable sanctions. Legislatures pass
laws, and fail to abolish others, for a variety of
purposes, only one of which is to establish clear
expectations and guidelines for enforcement
3587
P
through police action. Moreover, the police seek
to achieve various objectives in carrying out their
duties, and at times, those objectives conflict with
one another. In such instances, the police must
decide which objectives take precedence over
others. For example, during a public demonstration held in the streets, the police may find that
the objective of keeping traffic avenues clear
conflicts with the objective of safeguarding citizens’ rights to peaceful protest.
Even absent conflicting objectives, police and
other criminal justice resources are far too limited
to allow the police to enforce all laws exhaustively. Most communities would not tolerate full
enforcement of the law, even if resources would
allow it, preferring a degree of police tolerance,
especially for minor legal transgressions. The
very capacity of the criminal justice system to
continue functioning in many communities
depends to a great extent on the police not fully
enforcing the law. Sudden increases in police
arrest activity can seriously challenge the capacity of the legal system to process the resultant
cases.
Discretion is exercised in policing at all levels
of the police hierarchy. In contrast with other
occupations and professions, the greatest amount
of discretion in policing is exercised at the line
level by patrol officers and detectives, but supervisors and policy makers also exercise large
amounts of discretion.
Fundamentals of Police Discretion
Perhaps the most profound types of discretionary
decisions made in policing are the decisions to use
force and to arrest. In regard to the use of force,
Bittner (1970: 46) famously described the core of
the police role as “the distribution of nonnegotiably coercive force employed in accordance
with the dictates of an intuitive grasp of situational
exigencies.” What Bittner meant is that the police
are regularly called to the scenes of trouble and
sometimes find it on their own. These situations
are often confusing, if not chaotic. In these situations, police officers are authorized to use reasonable force, if necessary, in order to quell the
P
P
3588
trouble. The decision about whether the use of
force is necessary, and if so, how much force to
use, is discretionary. The decision is made by the
officer or officers on the scene, based on their
“intuitive grasp” of the circumstances (the “situational exigencies”) as they unfold.
Police decisions about whether to arrest are
discretionary too. Generally, the police are authorized to arrest (1) with a warrant issued by a court,
(2) whenever a person commits any crime in their
presence, and (3) whenever they have probable
cause to believe that a felony (serious) crime has
been committed and that a particular person has
committed it, even if the probable cause is based
on secondhand or thirdhand information. In
regard to warrantless arrests (types 2 and 3
above), which are the most commonplace arrest
situations, law and official policy rarely say that
the police must arrest. Most often the law says, or
is interpreted as meaning, that the police may
arrest, leaving discretion in the hands of ordinary
police officers.
Perhaps the most ubiquitous discretionary situations encountered by the police are the decisions whether to stop drivers who violate traffic
laws and, when traffic stops are made, whether to
cite the drivers and, further, to search the vehicle
for evidence or contraband. Many observed traffic violations are minor in nature (such as driving
1 mile over the speed limit) and therefore frequently ignored by the police or, at most,
addressed with a warning. By the same token,
poor driving causes many deaths and injuries
every year and a huge amount of property loss,
so traffic policing, and its associated discretionary decision making, should not be dismissed as
a trivial matter. As well, when officers decide to
issue citations to drivers, fines are routinely in the
hundreds of dollars today and can result in the
loss of driving privileges, which in turn can have
serious consequences for employment and
income. Moreover, the police commonly engage
in traffic enforcement as a method of more serious crime control: to locate illegal guns and
drugs, to locate persons with outstanding arrest
warrants, to disrupt illegal drug and prostitution
markets, to deter burglars and thieves from
prowling neighborhoods, and so forth.
Police Discretion and Its Control
There exists some research data about the frequency of these three major discretionary situations – use of force, arrests, and traffic
enforcement. Clearly, the police frequently
choose not to use force, or have no justification
to do so – estimates are that the police use physical force (beyond handcuffing) in only about
20 % of all arrests (Terrill et al. 2008), in only
2–3 % of all encounters with the public (Friedrich
1980), and in less than 1 % of all traffic stops
(Eith and Durose 2011). Similarly, the police
often choose not to arrest – one study estimated
that the police make arrests in about one-half of
situations when it would be lawful to do so (Black
1980), and another study found that arrests are
made in only 5 % of total police encounters with
the public (Whitaker 1982). This still results in
a large number of arrests though – 13 million in
2010, of which 2.2 million were for violent or
property offenses, 1.6 million for drug offenses,
and 1.4 million for drunken driving (FBI 2011).
The picture is similar regarding traffic enforcement, with just over half (55 %) of 18 million
drivers stopped in 2008 reporting that they were
given a citation (Eith and Durose 2011).
Beyond use of force, custodial arrests, and
traffic citations, the police make many other
types of discretionary decisions, including decisions about which laws to invoke when an arrest
or other form of detention is made (e.g., whether
to add a charge of resisting arrest when an
offender does not come along peacefully);
whether to refer matters to other agencies; what
tactics to adopt in mounting proactive operations;
what conduct to investigate, what investigative
techniques to apply, and how intrusive those
techniques are; what level of resources to commit
to various activities, places, and problems;
whether to secure prior authorization for certain
actions (e.g., whether to apply for search warrants
or other court orders); what level of urgency to
give to various duties; whether to grant permission for certain activities to take place (e.g.,
where the police are responsible for issuing
parade permits); and so forth. Many of these
discretionary decisions are made by patrol officers and detectives, but some are also made by
supervisors and commanders.
Police Discretion and Its Control
Variations in Discretion
The range of choices open to police officers
varies in different types of situations, as does
the degree of latitude that they have in making
those choices. So, for example, if an officer is
dealing with a person experiencing a mental
health crisis, it matters if there is a specialized
unit that can respond and take over, or if there is
a mental health crisis center to which the person
can be taken, or a psychiatric unit within the local
hospital. If none of those options is available,
then the number of choices open to the officer is
more limited – to perhaps not much more than
curbside counseling or an arrest for disorderly
conduct.
In some jurisdictions, while the law may
authorize the police to arrest for certain offenses,
that option may not really be available. For example, the county jail may be completely full, causing the jail administrator or sheriff to advise local
agencies not to make any arrests except in
extreme cases. In this situation, police officers
may still have the option to issue a citation to an
offender, but the option of physical or custodial
arrest would be severely curtailed. In another
scenario, it is common for prosecutors to inform
police departments that, because of heavy caseloads facing the court, arrests should not be made
for particular categories of minor offenses. This
situation would discourage even the issuance of
a citation, unless the accused had the option of
paying a fine in lieu of trial, thus preventing any
additional workload burden on the court.
Another real-world constraint on police discretion is the number of police officers available
to handle other police matters should an officer
take some action that removes him or her from
the streets for an extended period. Whether in
small agencies where there is only one police
officer on duty or in larger agencies where police
reinforcements are a long distance away, an officer might have to weigh the benefits of making an
arrest against the costs of being tied up for several
hours processing the arrestee, leaving no or few
other officers to police the area. Both the public’s
and any remaining officers’ safety must be taken
into account. An added dimension could be that
3589
P
the county jail, hospital, or other facility is many
miles away, building even more processing time
into the detention and literally taking the officer
out of the jurisdiction. These kinds of practical
constraints may narrow the real range of choices
available to officers when handling various situations, thereby further constraining their
discretion.
In regard to the use of force, not all police
officers and police departments have the same
equipment available to them, which can affect
their range of options in some situations. For
example, conducted energy devices (e.g., Tasers)
have become commonplace, but some officers do
not have them and thus do not have one of the
more effective less-lethal weapon options in confrontations with combative subjects. Similarly,
when a high-speed pursuit is underway, the
range of options is affected by whether the officers out ahead of the pursued vehicle have stop
sticks or similar technology. Likewise, in
a hostage or barricaded-person scenario, some
police officers might have sophisticated communications, eavesdropping, or breaching technology available, while others might not, affecting
the tactical choices open to them.
From a comparative perspective, the 18,000
chief police executives in the United States probably have broader discretion than their counterparts in most other countries, because of the
highly decentralized structure of American policing. A US police chief has the authority to choose
the type of pursuit-driving policy to implement,
whether to place a higher priority on drug
enforcement or traffic enforcement, whether to
require applicants to have a college degree, how
to address various types of policing incidents and
problems, and so forth. In a country with a unitary
(national) police system, these and other decisions are usually made at the national level, leaving much less discretion in the hands of local
commanders. Sheriffs in the United States may
have the broadest executive-level discretion of
all, since they are locally elected and are primarily accountable to the electorate rather than to
other government executives.
Differing legal systems may also affect police
discretion. Within the United States, local and
P
P
3590
state police officers generally exercise wider discretion than do federal law enforcement officers,
whose investigations, in particular, are guided to
a larger degree by prosecutors (US attorneys)
than is common at the state and local levels.
Similarly, in countries that have civil law or
inquisitorial legal systems, judges and/or prosecutors often direct investigations, reducing the
degree of discretion available to police. In some
countries, frontline police personnel (the equivalent of patrol officers in the United States or
constables in the UK) do not have the same
broad authority or discretion in the use of force,
arrest, and investigation described in the previous
section. The actual legal authority of ordinary
police officers is more restricted, or organizational rules and procedures require them to get
permission from superior officers before taking
significant actions. Of course, it is also true that in
some countries the authority of the police to
detain, search, and interrogate is even broader
(under the law or simply in practice) than
discussed here, effectively increasing the amount
of discretion available to those permitted to exercise the authority, whether it is frontline police
personnel or superior officers.
Factors That Affect Police Use of
Discretion
An important theme in police research has been the
study of police decision making. Besides
documenting the existence of discretion, as
described above, many studies have looked for
patterns and correlations in police decisions to
use force, stop vehicles and pedestrians, make
arrests, issue citations, and conduct searches. Factors affecting police decision making and the use of
discretion by police officers fall into four general
categories: situational characteristics, police officer
characteristics, police organization characteristics,
and neighborhood/jurisdiction characteristics.
Consistent with the finding that the police
exercise broad discretion and also consistent
with Bittner’s (1970) argument that police behavior is guided by “situational exigencies,” characteristics of the situations in which the police
Police Discretion and Its Control
encounter the public have been found to have
the greatest influence on police decisions to
arrest, cite, and use force (Skogan and Frydl
2004). Among these, studies have found that the
most influential are legal factors – the seriousness
of the offense committed by the suspect, the
suspect’s previous offending history, the amount
of evidence available, and the degree of threat
posed by the suspect toward the officer. This
finding – that police use of discretion is affected
more by legal factors than any others – is
reassuring. A related quasi-legal factor that also
has some influence is the stated preference of the
victim, especially when the victim asks the officer not to arrest the suspect.
Compared to these legal factors, a variety of
extralegal situational factors have been found to
have mixed and inconsistent effects on police use
of discretion. These extralegal factors include the
demeanor of the suspect and the suspect’s race,
sex, and social class. Each of these factors has
been found to influence police decision making in
some studies, but not always in the same direction, plus many other studies have found no
effects. Of these factors, the one that has gotten
the most empirical support, especially in more
recent studies, is race (Kochel et al. 2011). Studies of so-called racial profiling consistently find
that people of color are disproportionately
stopped by the police, whether as drivers or as
pedestrians, and once stopped are more likely to
be searched. The key question, of course, is
whether these police decisions to stop and search
people of color are made based on the extralegal
factor of the race of the person, or whether they
are based on legal factors such as the seriousness
of the person’s behavior and the amount of evidence that is available. Unfortunately, research
on racial profiling has not generally been able to
answer that key question.
Police decision making and use of discretion
might also be influenced by the personal characteristics of the police officer, such as his or her
race, sex, and level of education. However, the
cumulative findings from many studies indicate
that these factors do not have large or consistent
effects on police officer behavior. Similarly, neither police officers’ general orientations toward
Police Discretion and Its Control
their role nor their levels of authoritarianism,
cynicism, or job satisfaction have been found to
have any consistent influence on their use of
discretion.
A third set of factors that might affect police
officer decision making is organization-level
characteristics. For example, officers in small
police departments might use their discretion differently than officers in larger departments, officer decision making might vary between different
types of agencies (e.g., municipal police departments versus sheriff’s offices), or decision making might be influenced by the degree to which
a police organization is centralized, specialized,
or formalized. The internal climate or culture of
a police organization might also lead to variations
in police behavior, including decision making.
While a good case can be made that organizational characteristics such as these ought to affect
the use of discretion by organization members,
the kinds of multiorganization studies that would
be needed to verify such effects have simply not
been conducted.
It has also been argued that police use of
discretion varies according to characteristics
external to the police, such as between rural,
suburban, and urban communities, or between
poor neighborhoods and affluent ones. Once
again, a theoretical case can be made that community characteristics should influence police
behavior, including decision making in discretionary situations, but relatively few studies
have been able to incorporate multiple neighborhoods and/or multiple jurisdictions, and the
results from those few studies have not been
consistent (Varano et al. 2009). As was true for
officer characteristics and organizational characteristics, community characteristics have not
been found to have as much impact on police
officers’ use of discretion as situational characteristics, and within the latter, legal factors seem
to have the greatest influence.
Consequences of Police Discretion
As discussed above, police discretion seems inevitable given the broad mandate of the police, the
3591
P
incredible variety of situations that the police
encounter, the tendency of legislatures to pass
laws that no one expects will be fully enforced,
the limited capacity of jails and courts, and
numerous other philosophical and practical considerations. So the police have broad discretion,
though more in some situations than others. Also
as noted, studies indicate that situational legal
factors affect police officers’ decisions more
than any other factors, although future studies
using better methodologies may reveal that extralegal factors (such as the race of the suspect) and
the characteristics of police organizations and
communities have more influence than can presently be demonstrated.
Given the existence of police discretion, what
are its consequences? The most obvious immediate result is that the police often choose not to
enforce the law, thus softening the impact of the
criminal law on people’s lives. With discretion,
officers get to choose how to handle various situations, and at least half the time when an arrest
could be made or a citation issued, they choose
under-enforcement or some other version of
leniency. Because of discretion, officers’ actions
can be affected by sympathy and empathy, and
officers have the opportunity to choose what they
consider to be the most fair or just resolution of
the situation, even if that entails not enforcing the
law. Discretion enables the police to treat people
as people, taking into consideration a person’s
circumstances, motives, intent, remorse, and
promises – many would say that discretion allows
the police to humanize the application of the
criminal law.
This phenomenon of under-enforcement
enabled by discretion might be a cause for more
serious concern if it reduced the deterrent effect
of arrest and punishment, emboldening potential
offenders and creating a sense of impunity within
society. Such a concern is less warranted though,
since the police exercise their discretion mainly
in response to minor crimes and disorders. It is
generally the person caught drinking in public,
driving 10 miles over the speed limit, or
shoplifting a small item who benefits from an
officer’s decision to handle a matter informally,
not the burglar, robber, or murderer. With that
P
P
3592
said, however, it is also true that detectives and
prosecutors have discretion in making “deals”
even with serious crime suspects, as when they
offer to reduce or drop charges in return for
confessions, incriminating evidence or testimony
against codefendants, and other types of cooperation. The exercise of discretion by detectives has
not been studied as closely as the discretionary
arrest and use of force decisions made by patrol
officers.
Concern about the consequences of underenforcement and leniency cannot be ignored,
despite the evidence that the police have been
making lots and lots of arrests. “Zero tolerance”
campaigns remain very popular, at least in the
United States. It is not usually necessary to declare
zero tolerance for murder or robbery – rather, zero
tolerance campaigns usually target such lesser
offenses as street prostitution and panhandling, or
else they define a specific geographic area within
which all criminal laws will be enforced, no matter
how minor. In the public forum and in police
strategy circles, zero tolerance is most commonly
presented as a strict crime-control policy, leading
to debates about whether it is fair, sustainable, and
effective in reducing crime. However, zero tolerance is also a direct assault on police discretion.
A zero tolerance policy takes discretion away from
police officers since it eliminates every option
other than enforcement.
In light of the concern that police discretion
might result in too much leniency, it is ironic to
note that, at least in the United States over the last
20 years, jails and prisons have been full, if not
overflowing. In other words, the police have had
broad discretion, and the evidence indicates that
they have regularly used it to avoid making
arrests – and yet, court dockets are crowded and
jails are full. One might wonder what the situation would be if not for police discretion.
There is one aspect and consequence of this
discretion/leniency phenomenon that deserves
special consideration. Police discretion sometimes extends to deciding whether a reported
event counts as a crime, whether to record it as
a serious crime or a lesser crime, whether to
thoroughly investigate it, and whether to recommend it for prosecution (Lum 2011). If the victim
Police Discretion and Its Control
of a reported crime is perceived by an officer as
unworthy of serious concern, the crime might be
discounted or ignored. For example, this might
occur to a woman domestic violence victim or
sexual assault victim if the police officer identifies more with the other party involved, or if the
officer believes that the victim somehow contributed to her own misfortune. Similarly, an officer
might fail to act in response to a reported crime if
the victim is someone whose status is not fully
respected, such as a recent immigrant or a person
whose sexual orientation is considered deviant by
the officer. In these examples, discretion allows
the police to act leniently toward offenders in
situations in which their victims are held in low
regard. It can be argued that these are simply
more examples of discretion enabling the police
to pursue justice and fairness rather than mandatory enforcement, but it is at least equally possible that, in these types of cases, discretion allows
officers to act on the basis of their own stereotypes and biases to the disadvantage of deserving
victims.
Another potential problem with discretion and
under-enforcement/leniency arises if the police are
more lenient toward majority groups and higherstatus individuals than others. The police might
seem lenient if they only arrest 30 % of the people
they stop for drinking in public, or if they only
conduct searches in conjunction with 10 % of their
traffic stops. Analyses often indicate, however,
that members of minority groups are disproportionately arrested for minor offenses and disproportionately subject to searches following traffic
stops. In other words, the overall picture is one of
under-enforcement, but often the enforcement that
is undertaken is differentially applied to persons of
color. As explained above, this pattern of police
discretionary decision making, often called racial
profiling, might still be accounted for by situational legal factors, but it might also be
a reflection of conscious or unconscious bias.
The inherent and most fundamental challenge
associated with discretion is simply inconsistency. Discretion gives police officers leeway in
deciding how to handle encounters and confrontations based on an “intuitive grasp” of each
situation. Ultimately, some people caught with
Police Discretion and Its Control
a marijuana cigarette get arrested and some do
not; some drivers who fail to signal their turns get
citations and some do not. This brings into focus
two competing conceptions of justice and fairness – one perspective says that fairness is
treating every rule breaker the same, while the
other perspective says that fairness is treating
each person as a person, according to their
unique circumstances. Many of us subscribe to
the first viewpoint as long as it is applied to
others, but want the second approach taken
when we are the ones against whom the rules
and laws might be applied.
Guiding and Controlling Police
Discretion
While police discretion may be inevitable, it is
widely accepted that it should not be unfettered:
that it needs to be guided and controlled in order
to avoid the kinds of negative consequences
discussed above, including inconsistency, undue
leniency, and discrimination. State statutes and
local laws can and do guide police discretion to
a certain degree. For example, some legislatures
have sought to restrict police discretion in specific types of cases, notably in the realm of
domestic violence through laws mandating arrest
of offenders (Phillips and Sobol 2010). Such
efforts sometimes merely shift the locus of discretion, however. If the law requires the police to
arrest whenever they have probable cause to
believe that a domestic assault has occurred, but
officers in some situations do not think that arrest
is the best response, they may become more
likely to decide that the evidence falls short of
probable cause, or that no crime has occurred at
all. In some cases this kind of decision making
may also be driven by a victim’s stated preference for no arrest or by knowledge that the prosecutor is only willing to take certain kinds of
cases into court. This is not to say that mandatory
laws have no impact on police discretion and
decision making, but it is important to recognize
that the police and other criminal justice decision
makers tend to react in ways that preserve their
discretion (Green and Kelso 2010).
3593
P
The supervising authority of the courts,
including specific court rulings and the risk of
civil liability, also has the potential to influence
the exercise of police discretion, at least at the
margins. The US Supreme Court’s Tennessee v.
Garner (1985) decision is a case in point. Prior to
Garner, the police had the discretion to use
deadly force in situations in which a fleeing or
escaping felon could be captured only by shooting them as they fled. The Supreme Court ruled
that this authority was too broad and effectively
revoked the “fleeing felon rule,” thus narrowing
police discretion in the use of deadly force.
By themselves, however, laws and court decisions are generally inadequate for guiding and
controlling police discretion. As noted, often
they merely shift the locus of discretion to
another point in the process. Also, court decisions
typically resolve one specific issue but leave
others unaddressed. The Garner case, for example, eliminated one ancient common law justification for police use of deadly force, but did not
posit a new rule in its place or seek to clarify other
justifications for police use of deadly force.
Moreover, one of the key legal principles underlying police use of force, including deadly force,
is “reasonableness” – the police can use any level
of force that is reasonable to defend themselves
or others from serious threats of bodily harm or to
carry out their other lawful responsibilities, such
as enforcing the law. While the Garner case ruled
that deadly force was unreasonable in one particular set of circumstances, the reasonableness
standard remains and it is inherently vague and
subjective. An inevitable consequence is that the
police have significant discretion in deciding
whether to use force and how much force to use.
In large part because of these limitations of the
law, written policies and procedures are increasingly used by police agencies to set out the
parameters of officers’ discretion in certain
types of cases and to provide contextual guidance
for the proper exercise of it. In the case of
domestic violence, for example, police department policies often reinforce and tighten legal
requirements by (1) articulating limited exceptions to mandatory enforcement, (2) requiring
written justification whenever nonenforcement
P
P
3594
is chosen in an applicable situation, (3) identifying decisions that require supervisory approval,
and (4) specifying penalties for failing to adhere
to legal and policy mandates. Similarly, in the
case of deadly force, police department policies
frequently define an even narrower range of justifications than anything found in the law, they
usually caution officers about the sanctity of
human life, and they invariably specify elaborate
processes that must be followed in the aftermath
of any incident in which the police exercise force
that could lead to death, regardless of whether
anyone dies or is even injured. In neither case,
domestic violence nor use of force, do written
policies completely eliminate discretion, but
they constrain and guide it more thoroughly
than is accomplished by law alone.
Somewhat less common, but no less important, are written policies that provide guidance for
the exercise of discretion in general. One example policy identifies factors that officers should
and should not take into consideration when
deciding whether to make an arrest (Scott 1995):
In general, police officers, using sound professional judgment, may take the following factors
into consideration when deciding whether or not
to arrest a citizen: (a) the seriousness and nature of
the offense (generally, the more serious the
offense, the more likely arrest is the preferred
course of action); (b) the potential that arrest will
effectively resolve a conflict; (c) the availability of
legal alternatives to arrest that would adequately
resolve the conflict or problem; (d) the likelihood
that the citizen will be deterred from future violations by warning and education; (e) the officer’s
belief that the citizen made an honest mistake in
violation of the law; (f) the victim-witness’s interest in prosecution; (g) the likelihood of formal
prosecution of the offense; (h) the potential that
arrest will create more serious breaches of the
peace or other problems (e.g., inciting riot); (i)
legitimate competing priorities for police
resources.
The following factors are among those that are
improper for a police officer to consider in deciding
whether or not to make an arrest: (a) the citizen’s
economic status, race, ethnicity, gender, or other
status for which the law prohibits legal discrimination; (b) the revenue likely to be generated by fines
or penalties imposed upon conviction; (c) the personal or professional relationship that the citizen
has to the police officer or to other influential
citizens; (d) the personal advantage to the officer
Police Discretion and Its Control
for processing or avoiding processing of the arrest
(e.g., overtime compensation, desire to finish tour
of duty, avoidance of paperwork, etc.).
Police organizations utilize several other
methods of guiding officers in the exercise of
discretion. Police agency accreditation standards, while not typically dictating the substance of most discretionary judgments, can
help create an organizational structure that supports administrative rule making. Police training
programs are essential for improving officers’
decision-making skills in the application of policies and procedures. The development of
a body of professional and scientific knowledge,
grounded in research and practice, about how
the police can effectively and fairly address
public safety problems also holds promise for
shaping important discretionary decisions
(Weisburd and Neyroud 2011; Center for Problem-Oriented Policing 2012). In addition, many
police agencies are increasingly looking to citizens to provide them with guidance on a range of
discretionary matters, from what public safety
problems to focus on to what means to use in
addressing them. Citizen input can be provided
at the policy-making level as well as through
systematic input to line officers (Livingston
1997).
Years ago, Jerome Skolnick (1966) observed
that the police, like members of most organizations, are affected by the need to appear productive. Thus, officers’ discretionary decisions about
whether to arrest or issue citations may be
influenced by supervisory or command-level
expectations about how much “activity”
a productive officer should generate. While not
new, these pressures seem to be greater today in
the era of “metrics” and heightened accountability. Many police departments now use some form
of Compstat, a system designed to help top-level
executives use crime analysis and statistics to
hold mid-level commanders accountable for
targeting and reducing crime in their areas of
responsibility. These mid-level commanders
often press their subordinates to increase enforcement in order to show that they are doing everything possible to reduce crime. The net effect is to
reduce discretion at the lowest levels in order to
Police Discretion and Its Control
satisfy pressures from above. Another effect, perverse and unintended, may be to encourage officers to use their discretion to downgrade or ignore
some reported crimes, thus making it appear that
crime reduction is occurring even if it is not
(Eterno and Silverman 2012).
Modern technology is also affecting police
discretion. Thanks to mobile radios, cell phones,
in-car computers, and automatic vehicle locator
systems, today’s police supervisors and commanders can keep track of their subordinates
more closely than was the case a decade or two
ago. Even more intrusive are in-car cameras and
body cameras (not to mention the ubiquitous cell
phone cameras in the hands of the public) that
record everything that police officers do, thus
magnifying the potential for after-the-fact
review. While these surveillance technologies
may not directly narrow the discretion that officers have in making their decisions, they probably have a substantial indirect effect, since
officers must give more consideration to how
their decisions will look and be judged in retrospect. In the near future, this technological
impact is likely to be magnified further when
the feeds from in-car and body cameras go
“live” back at police headquarters. When that
time comes, supervisors and commanders will
have a greatly expanded opportunity to direct
officer decision making in the field. This is likely
to shift the locus of some discretionary decision
making away from police officers and onto the
managers back in the electronic control room.
This kind of technological breakthrough could
produce a fundamental change to our traditional
understanding of police work and the centrality of
street-level police discretion. It will be quite
interesting to see whether it actually does have
this profound effect, or if, once again, it merely
causes some shifting in the forms and locus of
police discretion.
Related Entries
▶ Biased Policing
▶ Order Maintenance Policing
▶ Police Discretion in Law Enforcement
3595
P
▶ Police Discretion in Providing Services and
Assistance
▶ Police Legitimacy and Police Encounters
▶ Role and Function of the Police
Recommended Reading and References
Banton M (1965) The policeman in the community. Basic
Books, New York
Bittner E (1970) The functions of the police in modern
society. GPO, Washington, DC
Black D (1980) The manners and customs of the police.
Academic, New York
Brown MK (1981) Working the street: police discretion
and the dilemmas of reform. Russell Sage, New York
Center for problem-oriented policing (2012) POP Guides.
Online at http://www.popcenter.org/guides/
Davis KC (1969) Discretionary justice. Greenwood Press,
Westport
Eith C, Durose MR (2011) Contacts between police and
the public, 2008. Bureau of Justice Statistics,
Washington, DC, Online at http://www.bjs.gov/
content/pub/pdf/cpp08.pdf
Eterno JA, Silverman EB (2012) The crime numbers
game: management by manipulation. CRC Press,
Boca Raton
FBI (2011) Crime in the United States: 2010. Federal
Bureau of Investigation, Washington, DC, Online at
http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/
2010/crime-in-the-u.s.-2010/persons-arrested
Friedrich RJ (1980) Police use of force: individuals, situations, and organizations. Ann Am Acad Polit Soc Sci
452:82–97
Goldstein J (1960) Police discretion not to invoke the
criminal process: low-visibility decisions in the
administration of justice. Yale Law J 69:543–589
Goldstein H (1963) Police discretion: the ideal versus the
real. Public Admin Rev 23:140–148
Goldstein H (1977) Policing a free society. Ballinger,
Cambridge, MA
Green T, Kelso C (2010) Police discretion in domestic
violence incidents. Am Behav Social Sci J 14:24–44
Kochel TR, Wilson DB, Mastrofski SM (2011) Effect of
suspect race on officers’ arrest decisions. Criminology
49:473–512
LaFave W (1965) Arrest: the decision to take a suspect
into custody. Little Brown, Boston
Livingston D (1997) Police discretion and the quality of
life in public places: courts, communities, and the new
policing. Columbia Law Rev 97(3):551–672
Lum C (2011) The influence of places on police decision
pathways: from call for service to arrest. Justice Quart
28:631–665
Phillips SW, Sobol JJ (2010) Twenty years of
mandatory arrest: police decision making in the face
of legal requirements. Criminal Just Policy Rev
21:98–118
P
P
3596
Reiss AJ Jr (1971) The police and the public. Yale
University, New Haven, CT
Scott MS (1995) Lauderhill (Florida) police department
policy manual. Personal files of the author
Skogan W, Frydl K (2004) Fairness and effectiveness in
policing: the evidence. National Academies Press,
Washington, DC
Skolnick J (1966) Justice without trial. Wiley, New York
Tennessee v. Garner. (1985). 471 U.S. 1
Terrill W, Leinfelt FH, Kwak D-H (2008) Examining
police use of force: a smaller agency perspective.
Policing Int J Police Strat Manag 31:57–76
Varano SP, Schafer JA, Cancino JM, Swatt ML
(2009) Constructing crime: neighborhood characteristics and police recording behavior. J Criminal Just
37:553–563
Weisburd D, Neyroud P (2011) Police science: toward
a new paradigm, New perspectives in policing.
National Institute of Justice, Washington, DC
Westley WA (1970) Violence and the police. M.I.T. Press,
Cambridge, MA
Whitaker GP (1982) What is patrol work. Police Stud
4:13–22
Wilson JQ (1968) Varieties of police behavior. Harvard
University, Cambridge, MA
Police Discretion in Law Enforcement
Robert E. Worden1,2 and Sarah J. McLean2
1
School of Criminal Justice, University at
Albany, State University of New York, Albany,
NY, USA
2
John F. Finn Institute for Public Safety, Inc.,
Albany, NY, USA
Overview
Coercive authority is central to the police
role, and the discretionary use of police authority is a decision-making process, as officers
evaluate the situations in which they intervene
and choose a course of action from among
a set of alternatives. This entry approaches
the study of police discretion from this decision-making perspective. It is concerned with
several forms of authority: stops, frisks and
searches, arrests, tickets, and the use of physical
force.
Police authority is infrequently invoked, as
a proportion of all police-citizen contacts.
Police Discretion in Law Enforcement
Officers consider their authority as a resource in
“handling” situations, and often they are able to
handle situations without resorting to the use of
formal authority. Research has illuminated to
a degree the circumstances under which authority
is used, from which inferences are drawn
about the influences on police decision-making,
but research has not for the most part formed the
basis for normative judgments about the discretionary use of authority – when it has been used
well or poorly.
Explaining the police decision-making
process can be approached like any decisionmaking process. Behavior can be predicted or
explained when the premises for decision are
known. Police officers’ decisions turn on factual
premises, which concern the consequences of
alternative police actions and the nature of the
situation from which the consequences can be
projected, and value premises, about the goals
of police intervention and the desirability of alternative sets of consequences. Explaining police
decision-making is, however, complicated by
the ambiguity and uncertainty of police officers’
task environments.
Many different elements of police-citizen
encounters have been analyzed as hypothesized
influences on police behavior, but only a few of
them have been consistently found to affect the
use of police authority. Several of those are legal
factors: the seriousness of the offense, the
strength of evidence, and the preferences of
complainants for disposition. Among extralegal
factors, only the demeanor of suspects has
emerged as a fairly consistent predictor of police
action.
The discretionary exercise of police authority
varies across individual officers. It is fairly clear
that this variation is patterned by officers’ length
of service and also by officers’ outlooks and
personality traits. The influence of other individual factors, such as educational background, is
undetermined.
The use of police authority is subject to
regulation by departmental policies, but the circumstances under which policy regulation can
have maximum impact are seldom realized in
policing. The use of deadly force can be
Police Discretion in Law Enforcement
successfully restricted. Other uses of police
authority are probably less susceptible to policy
guidance.
Among the many open questions about police
use of their authority are three addressed in this
entry: the influence of citizens’ race, modeling
police decision-making, and how research might
contribute to improvements in the quality of
police decision-making.
3597
P
next section, the entry discusses the forms of
authority that police exercise and how they have
been analyzed; then it explains how a decisionmaking framework can be applied to police
discretion in their use of authority. In the following
section, the entry summarizes what is known about
the explanations of officers’ discretionary choices.
Then several open questions are addressed.
Introduction
Police Authority and Discretionary
Decision-making
Coercive authority is central to the police
officer’s role; it is a unique occupational prerogative that enables police to “handle” urgent situations. The situations may be crimes, with clear
and serious violations of criminal law, the
response to which is the apprehension of perpetrators, or the situations could be disorders, with
ambiguities stemming from whether the law
applies, who bears how much responsibility for
the situation, and what should be done about it.
But all of these situations share an element of
urgency, in that prompt action is needed, whether
it is a burglary in progress, a drunk driver, a loud
party, a domestic conflict, or a blocked driveway
(see especially Bittner 1974; Muir 1977).
In each instance, police exercise discretion,
which is the capacity to choose among courses
of action based on one’s judgment. Police officers
typically work outside of direct supervision
and often in private settings in which only the
officer(s) and the parties directly involved have
knowledge of the circumstances and the actions,
if any, that police take. So while thick books of
policies and procedures exist in many agencies,
in practice, officers are left to define the situation
and fashion a response, thus determining whether
and, if so, which policies and procedures apply.
The exercise of police discretion is a decisionmaking process, as officers evaluate the situations
in which they intervene and choose a course of
action from among a set of alternatives; if their
involvement is not at the request of a citizen, they
also make decisions about whether to intervene.
This entry approaches the study of police discretion from this decision-making perspective. In the
Police Authority
Police exercise several forms of authority. An
officer may stop and detain someone when he/she
has reasonable, articulable suspicion that the
person has committed, is committing, or is about
to commit a crime. An officer may frisk, or pat
down, an individual if she/he has reasonable suspicion that the person might be armed. An officer
may conduct a more extensive search of a person
incident to that person’s arrest and may conduct
a warrantless search of a person, a vehicle, or even
a building under some other circumstances. An
officer may take someone into custody for the
purpose of charging them with a crime; if the
suspected offense is a felony, an officer needs
only probable cause to believe that a crime was
committed, while for most offenses of lesser seriousness, the officer may act only if she/he witnesses the offense or a citizen signs a complaint.
For some types of offenses, officers may issue
a summons that directs the alleged offender to
appear in court, rather than make a custodial arrest;
this is of course common for traffic violations and
also for some other less serious offenses (such as
public drinking). Finally, police may use the physical force that is reasonably necessary to perform
their duties; the US Supreme Court ruled in Tennessee v. Garner that police may even use deadly
force if an “officer has probable cause to believe
that the suspect poses a threat of some physical
harm, either to the officer or others.” The forms of
physical force that police may use has multiplied
with the development and adoption of less-lethal
weaponry, such as pepper spray and conducted
energy devices, in the last 10–20 years.
P
P
3598
Social science has focused mainly on the exercise of authority by patrol officers, and it has
normally conceived, measured, and analyzed
these forms of police authority one by one – that
is, some studies examine arrest, while others
analyze the use of physical force, and still others
analyze stops, etc. In most instances, officers’ use
of police authority consists of a discrete act –
making an arrest, conducting a search; the main
exception is the use of physical force, which is
commonly broken down into forms of force that
vary in their seriousness, such as physical
restraint, impact methods (punches or kicks),
and the use of weapons. These are not mutually
exclusive forms of behavior, of course; to the
contrary, an officer might well stop and detain,
frisk, arrest, and physically restrain an individual
in a single encounter. One effort to form a single
scale of authority weighted the individual
components, but empirically it was not much
different from an arrest/no arrest dichotomy.
In general, police authority is infrequently
invoked, as a proportion of all police-citizen
contacts. Many years ago, James Q. Wilson
(1968) observed that the tendency of the police
is to underenforce the law, declining to take legal
action even when the circumstances authorize
them to take legal action. Thus it was not uncommon for police to release in the field offenders
about whose culpability police had evidence
sufficient to warrant arrest. Wilson pointed out
that police see their authority as a resource in
accomplishing their objectives, such that the
application of their authority is not an end but
a means, and in many instances their objective is
merely to “handle the situation” – that is, to
restore order and to prevent immediate violence.
When an arrest was made, the “formal charge
justifies the arrest but is not the reason for it”
(Bittner 1974, p. 27); the reason for such an arrest
is that the situation could not be handled in
some other, informal fashion. Similarly, studies
of the use of force have found that police seldom
use physical force, and some research has found
that police frequently refrain from using physical
force when they could legally do so or use less
physical force than the citizen’s resistance would
justify. Sometimes police presence by itself is
Police Discretion in Law Enforcement
sufficient to restore order. And even if presence
alone is not enough, police may be able to
resolve matters by talking: persuading, cajoling,
exhorting, negotiating, or mediating.
Since Wilson made his observation, with
changes in the (semiprofessional) status of
police, criminal procedure, and civil liability,
the view of the police may be somewhat different,
but not dramatically different. Even far into (and
perhaps beyond) the “community era” of
policing, authority is central to the police role.
Police may now supplement their use (or threatened use) of authority with other approaches, but
even when they perform problem-oriented policing, authority remains a resource on which they
can and often do draw. In addition, the difficulty
of reorienting police to a more community- and
problem-oriented role – getting officers to think
in terms of groups of related incidents rather than
individual incidents and to consider unconventional solutions to those problems – suggests that
the use of coercive authority has not changed
dramatically in the last 40–50 years. Arrest and
physical force remain infrequent acts. Police
proactivity – contacts with citizens initiated by
the police – is almost certainly at a higher level
now, however, and it has occasioned some
controversy about its benefits and the social distribution of its costs.
Another change over time might lie in the
public’s expectations for police effects on social
conditions. In the early 1970s, Egon Bittner could
say that police are “empowered and required to
impose or . . . coerce a provisional solution upon
emergent problems . . .” (1974: 18; emphasis
added), suggesting that the solution extended no
further than the immediate situation. Yet we evaluate the police in terms of more – for example,
research on policing spouse assault assessed
police interventions in terms of recidivism by
assailants over a 6-month post-intervention
period, not in terms of more “provisional”
solutions.
Aside from hypothesized biases in the use of
authority, to be discussed below, research has
tended not to address the normative dimensions
of police authority, that is, whether an arrest or
the use of physical force was a good choice or
Police Discretion in Law Enforcement
a bad choice. The situations that police handle are
complex, and so a normative judgment about the
use of police authority is contingent on many
factors. Moreover, the criteria against which
such a judgment should be made are not at all
clear. Officers’ choices could be judged in terms
of immediate outcomes, such as whether agitated
citizens are calmed, or in terms of longer-term
outcomes, such as whether a domestic assailant
recidivates. None of these have been much considered in research on police use of authority.
Another potential criterion is a professional
one of “workmanship.” Bittner (1983) distinguishes legality from workmanship as standards
against which to judge police performance. The
former involves the application of “explicitly
formulated schemes of regulation,” while the latter is concerned with officers’ “ability to call
upon the resources of knowledge, skill, and judgment to meet and master the unexpected within
one’s sphere of competence.” Bittner goes on to
explain that the standard of workmanship would
subject officers’ performance to the scrutiny and
critique of others authorized to assess their work
after the fact: peers, who are subject to the same
review, and the community, on whose behalf
the work is done. Similarly, Klockars (1996)
proposes a “craft standard” against which the
use of force could be compared, that is, what
a skilled police officer would have done under
the circumstances. A skilled police officer would
take all possible steps to structure the situation
into which she/he intervenes in such a way that
either resistance will not be offered or the resistance can be overcome with a minimum of force.
Few efforts have been made to develop such
standards conceptually or empirically, however.
As a general proposition, social scientists
seem agreed that authority is a resource that
police should use sparingly. Muir (1977)
describes a “good” officer as one who is morally
comfortable with the use of coercive power as
needed but who uses it only as needed. Moore
(2002) allows as how police are more efficient
when they use a minimum of authority. The use
of police authority, he points out, depletes the
“stock of private liberty” (2002: 26) that the
public enjoys, and the “net value” of policing
3599
P
must take account of not only the crime and
disorder that police prevent but also the costs of
policing, one of which is the use of authority.
In the last 10 years or so, research has reflected
a better appreciation for how the quality of police
performance might have as much (or more) to do
with how authority is exercised as it does with
what form of authority is exercised. Surveys of
citizens have repeatedly shown that they are more
satisfied with their contacts with the police when
they believe that police treated them with dignity
and respect, gave them an opportunity to explain
their situations and listened to what they had to
say, and explained what police have done and/or
will do, so that it is clear that officers are taking
account of people’s needs and concerns and basing their decisions on facts. Moreover, these
patterns hold even for people who experience
unfavorable outcomes, such as being ticketed.
Unfortunately, however, research on police
behavior has only scratched the surface in its
examination of the behavioral manifestations of
procedural justice.
Decision-Making
The discretionary use of police authority is
a decision-making task, as officers must make
choices – numerous choices – in their encounters
with citizens. Among their choices, the uses of
their authority are prominent: to stop and question, to frisk or search, to arrest, to cite, and to use
any of a variety of forms of physical force. As an
encounter with a citizen unfolds over time, the
choices may recur.
Explaining this decision-making process, and
the behavior that ensues therefrom, can in principle be approached like any decision-making process. Behavior can be predicted or explained
when the premises for decision are known. Police
officers’ decisions turn on factual premises,
which concern the consequences of alternative
police actions and the nature of the situation
from which the consequences can be projected,
and value premises, about the goals of police
intervention and the desirability of alternative
sets of consequences (see Worden and Brandl
1990, pp. 302–303). Most research on police
behavior has, in effect, treated as hypothesized
P
P
3600
decision premises the characteristics of policecitizen encounters, such as the nature of the
offense, the demeanor of suspects, and the preferences of complainants for a disposition. Several
studies have also examined as potential influences the characteristics of police organizations;
organizational policies, procedures, and informal
norms may serve as decision premises. Some
research has, in effect, allowed for variation in
decision-making among classes of decisionmakers, for example, male and female officers,
novice and experienced officers, and collegeeducated and less-educated officers.
Explaining police decision-making in terms of
situational and organizational cues is complicated by the ambiguity and uncertainty of police
officers’ task environments. Situational cues are
ambiguous in that the meaning or significance of
any cue, or set of cues, is subject to different
interpretations. The credibility of citizens’ testimony (which often is contradicted by other
citizens’ testimony), the severity of injuries, the
degree of citizens’ sobriety, and many other
features of police-citizen encounters are all matters of officers’ judgment, and these judgments
can vary from situation to situation and from
officer to officer. The same cue can be interpreted
differently across multiple events or even ignored
as irrelevant in some of them. The meaning and
significance even of readily recognizable cues is
ambiguous, subject to different interpretations by
different officers. Similarly, organizational policies and group norms are ambiguous, requiring
that officers interpret them as they are applied in
individual cases.
Explanation is further complicated by uncertainty about cause-and-effect relationships. Once
an officer interprets the meaning of situational cues
and the applicability of organizational policies,
she/he must choose a course of action based partly
on the projected consequences of those alternatives, but no body of scientific or other technical
knowledge guides officers in making such projections, and even if there were, they would be in any
case only probabilistic. Officers may be socialized
into some common working rules, but they also
learn through their own idiosyncratic experiences,
drawing from them some equally idiosyncratic
Police Discretion in Law Enforcement
lessons for practice. Thus we can posit that the
meaning and also the implications of situational
and organizational cues are heterogeneous across
officers, which would attenuate the explanatory
power of situational and organizational variables.
Prior research on police decision-making, for
the most part, takes one of two approaches: ethnographic inquiry and quantitative analyses of
decision outcomes (e.g., police decisions to arrest
or to use force) in terms of situational factors that
are the hypothesized stimuli to which decisionmakers respond. The former have been very
illuminating, though they are limited in their generalizability. The latter are scientifically rigorous,
exploiting numerical data on the characteristics
of the immediate decision task to estimate (often
through regression analysis) their independent
effects on decision outcomes. Researchers draw
inferences about the forces that shape decisionmaking from the regression coefficients, treating
the process by which informational inputs are
interpreted and judgments are made to reach
decisions as a “black box.” But these analyses
are limited in their explanatory scope to factors
that are of a priori significance, and we know that
these factors fall far short of explaining police
decisions.
Moreover, much of the research has been
oriented not to explaining discretionary decision-making as such, but rather to determining
whether and to what extent the application of
police authority is influenced by “extralegal”
factors, such as race, sex, and age. Thus, the
analyses seek to control for the legal factors that
legitimately shape police behavior, whereupon
a determination can be made about whether residual variation is related to extralegal factors.
Influences on Discretionary
Decision-Making
Most of the factors that have been analyzed as
parts of an explanation of discretionary decisions
by police have been classified into one of several
sets, each a different level of explanation. As
noted above, some of these factors, and those at
the lowest level of explanation, are features of
Police Discretion in Law Enforcement
the police-citizen encounters in which the discretionary decisions are made; these are often
referred to as situational factors. A second set,
and the next higher level of explanation, are comprised of the characteristics and outlooks of
individual officers. A third set, and a still higher
level of explanation, are characteristics of the
neighborhoods that form the immediate social
context of police-citizen encounters. The fourth
set include characteristics of police organizations, including both the formal, structural
features of police agencies – rules, regulations,
and specialization – and the informal organization: the peer culture(s) and norms. Finally, some
research has attended to another, still higher level
of organization: the requirements and constraints
of the law, especially that of criminal procedure.
This entry concentrates on three of these sets:
situational, officers, and organizational.
Situational Factors
Situational factors are external to the
decision-maker and form the immediate decision
environment. The most succinct statement of this
explanatory perspective is Bittner’s, who asserts
that “the role of the police is best understood
as a mechanism for the distribution of
non-negotiably coercive force employed in
accordance with the dictates of an intuitive
grasp of situational exigencies” (1970: 46).
Situational factors represent these exigencies.
A more conventional sociological formulation is
that of Black and Reiss (1967), who posit that
police action turns on the social status, situational
status (e.g., as complainant, suspect), and
“subversive capability” of citizens. Most
research of this genre, however, has been less
theoretically driven in its assessments of situational exigencies.
Many different elements of police-citizen
encounters have been analyzed as hypothesized
influences on police behavior, but only a few of
them have been consistently found to affect the
use of police authority. Several of those are legal
factors: the seriousness of the offense, the
strength of evidence, and the preferences of complainants for disposition. Among extralegal
factors, only the demeanor of suspects has
3601
P
emerged as a fairly consistent predictor of police
action; the estimated effects of sex and age have
been mixed, as have those of race, to which the
entry returns below.
It should come as no surprise that the use of
authority by American police is affected by legal
factors, but research confirms that police are
more likely to use their authority when they
have evidence of criminal offending – and the
stronger the evidence, the more likely is the use
of authority – and when the offense is more
serious. It is perhaps remarkable, however, that
the effects are not stronger still. Certainly, police
use of authority is not determined by legal
factors; as noted above, police often do not take
legal action even when they have grounds for
legal action.
Furthermore, when complainants articulate
a clear preference for or against legal action,
police tend to comply, especially when the preference is for lenience. This tendency “gives
police work a radically democratic character,”
as Black (1971) points out, and as a result, police
apply a standard of justice that varies with the
moral standards of complainants, not a uniform
standard of justice. However, as Mastrofski et al.
(1995) observe, the cooperation of complainants
is likely to affect prosecution, and so when
complainants prefer that police not take legal
action, police might reasonably treat that as an
“instrumental” legal consideration. The frequency with which victims of domestic violence
are reluctant to cooperate has, given concerns that
these complainants are not exercising choice
freely, prompted changes in laws and policies
that are intended to attenuate the effect of complainant preference in such cases.
Both qualitative and quantitative studies have
found that police are more likely to respond punitively to suspects who display disrespect toward
the police. Empirical support for this relationship
is nearly uniform in research on police behavior,
even in the absence of a single, widely accepted
definition and operationalization of demeanor,
and notwithstanding whatever random measurement error has been contained in indicators of
demeanor. This relationship is open to one or
both of two interpretations: police penalize
P
P
3602
people who flunk the “attitude test” with arrest,
ticketing, or a “thumping” (Van Maanen 1978),
or police often give “breaks” to deferential suspects but seldom to those who are hostile or
disrespectful. Some support can be found for
both of these causal mechanisms.
Klinger (1994) questioned this “criminological axiom”, arguing that demeanor had been
improperly conceived and measured in previous
research and also that previous research had
failed to control (adequately or at all) for crime
committed by suspects during their encounters
with police, including especially assaults on
police officers. However, reanalyses of thenextant data, using measures that took account of
Klinger’s critique, affirmed the previous conclusions, and other elements of Klinger’s argument
have also been disputed (Worden et al. 1996). But
Klinger’s critique surely revealed shortcomings
and inconsistencies in the conceptualization and
measurement of demeanor, and subsequent studies have exhibited the greater care that Klinger’s
critique demanded.
Even so, theoretical ambiguity remains.
Research now differentiates “resistance” from
“demeanor,” yet many forms of (especially
passive or only verbal) resistance are entirely
legal, and we have reason to believe that many
forms of (both legal and illegal) resistance are
interpreted by police as failures of the attitude
test, such that we should interpret their effects as
the effects of demeanor (see Worden et al. 1996).
So long as research is interested in explaining the
discretionary use of police authority, and not
merely in forming post hoc judgments about its
propriety, conclusions about the effects of
demeanor must turn on how officers interpret
the various manifestations of disrespect.
More generally, the theory (such as it is) and
much of the empirical analysis of situational
effects has been based on a sociological framework, and one that makes an implicit assumption
that the effects are not contingent on the backgrounds or outlooks of individual officers. This
assumption is not entirely without merit,
inasmuch as early research on police emphasized
the shared working conditions, experiences,
values, and norms among police; hence assuming
Police Discretion in Law Enforcement
(albeit implicitly) that officers attend to the same
cues, interpret them similarly, consider the same
set of alternative courses of action, and choose
based on the same criteria is not unreasonable.
Furthermore, some results seem to suggest that,
on some dimensions of decision-making, officers
approach their choices similarly: the arrest decision is shaped by the seriousness of the offense,
the strength of evidence, the preferences of complainants, and the demeanor of suspects, and
while there may be individual differences on
these scores, the differences are not stark, and
across the board it appears that these variables
have a substantial degree of explanatory power.
In other choices, especially choices among
informal responses, however, the effects of situational factors have not been so robust (Worden
1989), and the use of authority other than arrest
may be more susceptible to individual differences. Several qualitative studies in the 1970s
(e.g., Brown 1981; Muir 1977) showed that officers do not all see their occupational environment
the same way and that they develop distinct
“operational styles”; different studies using different frameworks nevertheless described a set of
officer types that were broadly congruent
Officer Characteristics and Outlooks
While it is clear that officers’ discretionary
choices are structured by the characteristics of
the incidents that they handle, it is also clear
that these situational elements do not completely
determine officers’ responses, as much of the
variation in their behavior is unexplained by situational variables. Because officers’ decisions
are of “low visibility” to their superiors and to
the legal system, scholars have sometimes
speculatively attributed the unexplained variation
in behavior to officers’ attitudes and values. This
perspective directs attention to factors internal to
decision-maker, which might affect how situational cues are interpreted and evaluated and
how options are assessed. For example, Toch
(1980) found that violence-prone officers are
especially sensitive to citizens’ challenges to
their authority. Such an account has intuitive
appeal, and it is of more than theoretical significance, inasmuch as it underlies initiatives to
Police Discretion in Law Enforcement
change the composition of or otherwise
“upgrade” police personnel, since it is supposed
that minority officers, female officers, or collegeeducated officers have different outlooks and
therefore perform the police role differently.
It also underlies efforts to inculcate different
outlooks through training.
Systematic research has not always confirmed
these intuitive propositions. One of the earliest
and most striking findings from observational
research on police was that, while many officers
professed to be prejudiced against African-Americans, officers did not act on that prejudice in their
encounters with citizens (Reiss 1971). A later and
more general examination of attitude-behavior
consistency (Worden 1989) found only weak
relationships between officers’ occupational attitudes (for example, their conceptions of the
police role, attitudes toward citizens, and
attitudes toward legal restrictions) and their
behaviors (for example, proactivity in field interrogation and dispute resolution). Such results
perhaps should not be surprising, in view of the
ambiguity and uncertainty of police work.
The ambiguous situations into which police intervene are subject to different interpretations, and
they are not governed by well-established causeand-effect relationships between police actions
and desired outcomes, so officers’ choices
among alternative courses of action may be only
loosely structured by their attitudes and values.
Officers with the same occupational outlooks
might judge a situation differently, reading the
informational cues in a situation in different
ways, or perceive different ways of reaching the
same objectives.
Be that as it may, analyses in several agencies
have shown that small numbers of officers
account for disproportionately large fractions of
citizen complaints and use-of-force reports. Furthermore, Brandl et al. (2001) found that less
experienced officers are disproportionately
represented among officers with multiple complaints about the use of excessive force. Scrivner
(1994) discovered five groups of officers among
those referred to police psychologists due to their
use of excessive force, including officers with
personality disorders; officers whose job-related
3603
P
experiences – for example, traumatic incidents
such as police shootings – put the officers at risk
for abusing force; young and inexperienced
officers who were also “highly impressionable
and impulsive”; officers who develop inappropriate patrol styles; and officers with personal problems. Harris (2010) showed that officers differ in
their career “trajectories” of misconduct.
Further, more recent findings have provided
more support for the hypothesized effects of officers’ outlooks, mainly because it has analyzed
attitudes as bundles in terms of which subsets of
officers can be differentiated, rather than estimating the additive effects of attitudes one by one.
For example, Terrill et al. (2003) found that
officers whose occupational attitudes conform
more closely to the tenets of the traditional police
culture are more prone to the use of their coercive
authority. Similarly, Paoline and Terrill
(2005) found that such officers are more likely
to conduct searches during traffic stops. Cuttler
and Muchinsky (2006) found that personality
traits and work history predict “dysfunctional
job performance.” Other characteristics of
officers – their race, sex, and educational background – have all been hypothesized to affect
how officers do their jobs, though the evidence
on these hypotheses is mixed and inconclusive
(National Research Council 2004).
Thus it is fairly clear that the discretionary
exercise of police authority varies across individual officers. It is also fairly clear that this
variation is patterned by officers’ length of service and also by officers’ outlooks and personality traits. The influence of other individual
factors, such as educational background, is
undetermined.
Organizational Factors
The hypothesized effects of organizational
factors, and particularly the formal, structural
characteristics of rules, regulations, standard
operating procedures, and the hierarchical chain
of command through which they are enforced, are
central to common ideas about police accountability. Police departments are widely known as
quasi-military organizations, with many of the
trappings of such organizations (such as ranks
P
P
3604
and insignia), as well as the aura of a tight
chain of command. The nature of police tasks,
however, which makes discretion an inherent element of police work, also circumscribes the
effects of these organizational factors on the use
of police authority.
One mechanism by which police departments
might influence the use of force by their officers
is, of course, through rules and regulations that
set limits on when police may use force and how
much force they may use. Research on the use of
deadly force suggests that this mechanism can be
effective, under some circumstances. In 1972,
well in advance of the Garner ruling by the
Supreme Court (in 1984), the New York City
Police Department (NYPD) modified its policy
about the use of deadly force, imposing much
tighter restrictions, providing that “every other
reasonable means will be utilized for arresting,
preventing, or terminating a felony or for the
defense of oneself or another before a police officer resorts to the use of his firearm” (quoted in
Fyfe 1979). A study of the effects of this policy
change showed that it had a substantial effect on
officers’ use of their firearms (Fyfe 1979). The
policy had an especially pronounced effect
on what Fyfe (1988, p. 185) characterized as
“elective” shootings, that is, situations in which
officers could have chosen not to shoot at no risk
to themselves or other parties. Furthermore, the
reduction in shootings was achieved with no
increase in officer injuries or deaths. The effectiveness of this policy almost certainly turned to
a large extent on its enforcement. The NYPD
established a Firearms Discharge Review Board
to investigate and adjudicate all firearm discharges, holding officers accountable for their
use of deadly force. Other research also found
that restrictive deadly force policies are effective
only if they are vigorously enforced.
No one should suppose that departmental
policies that regulate the use of nonlethal force
are or can be equally effective, however. The
circumstances under which nonlethal force is justified are more heterogeneous and also more
ambiguous; thus policy prescriptions concerning
nonlethal force cannot achieve equivalent clarity.
Police Discretion in Law Enforcement
Use-of-force continua and matrices specify
different kinds and degree of force, and (sometimes) the degrees of citizen resistance to which
forms of force correspond, but the resistance can
take widely varied forms and the behavioral lines
that conceptually demarcate different levels of
resistance are far from clear in practice; ambiguity is inescapable. Moreover, the use of nonlethal
force is probably less reliably reported by
officers, and even if such force is infrequent as
a proportion of police-citizen contacts, it happens
with sufficient frequency that police superiors
may not be able to routinely investigate every
use of nonlethal force in order to determine officers’ compliance with policy. Furthermore, even
when investigations can be conducted, one can
anticipate that many will be inconclusive, with
evidence limited to the contrasting accounts of
the officer and the citizen against whom force
was used. Thus the enforcement of such policies
would almost inevitably lack the teeth associated
with NYPD’s reviews of firearm discharges
(where, as Fyfe points out, the population density
is so high that one might reasonably suppose that
few firearms discharges could be safely
concealed). This is not to imply that policies
would be altogether ineffective in limiting and
guiding the use of nonlethal force, however; one
recent study suggests that some types of policies
are more effective than others – and presumably,
better than no policy at all – in promoting the use
of force that is proportional to citizen resistance
(Terrill et al. 2011).
Similar expectations might be formed with
respect to policies that are intended to promote
or even mandate arrest in cases of domestic violence. The situations are ambiguous, and so the
applicability of policy is subject to officers’ interpretations in the field, and enforcement of policy
requirements would be hindered by the cost and
difficulty of determining whether officers
complied with the policy in individual cases,
which are numerous. Even so, some research
indicates that pro-arrest policies yield increases
in arrests in domestic assault cases, though examples of resistance and noncompliance can be
found.
Police Discretion in Law Enforcement
Open Questions
The open questions about police use of their
authority are far too numerous to even list here,
much less consider them. Three questions are
addressed: the influence of citizens’ race, modeling police decision-making, and how research
might contribute to improvements in the quality
of police decision-making.
Research on the police consistently finds
that African-Americans are disproportionately
represented among the people against whom
police use their authority. It has inconsistently
found that, other things being equal, race affects
at the margin the application of police authority.
The question of racial (and ethnic) bias in policing dates to at least the 1960s, when the
grievances of urban minority populations against
the police were cited as contributing factors to
civil unrest. It has been especially salient in the
aftermath of incidents in which police have been
suspected of abusing their authority in their treatment of minority citizens, of which the Rodney
King incident in 1991 may be the most widely
known example. But since the mid-1990s,
concerns about racial profiling by police – that
is, the use of a drug courier profile that included
race as a feature, and more generally stopping and
searching African-Americans and other people of
color based on their race or ethnicity – have been
prevalent and sustained, even prompting many
states and localities to mandate that police departments track the race of the people who are
stopped. Unfortunately, the volume of analysis
on this question has not been matched by its
theoretical sophistication, and bedeviled by
what is sometimes known as the benchmarking
problem, the studies can seldom provide persuasive evidence about whether the disparities are or
are not attributable to legitimate practices, and
moreover, they are not oriented toward explanation and offer little insight into the role of race in
police decision-making.
The mixed findings on the role of race may
reflect the complex, contingent nature of the
effects of race, but they might also be symptomatic of drawbacks to the conventional approach to
3605
P
analyzing
police
decision-making.
This
approach, described above, rests on an assumption that police decisions, like structural regression models, are a weighted sum of the postulated
decision cues: legal seriousness, strength of
evidence, complainant preference, suspect
demeanor, and the like. It is unlikely that the
process of police decision-making resembles
this computational model, however. And the
regression models seldom explain more than
a small fraction of the variation in behavior.
Different kinds of data and a different kind of
model might yield deeper insights into the forces
that shape police decisions.
A different approach, that of “protocol analysis” or “process tracing,” promises to shed further
light onto decision-making by opening the black
box of police officers’ cognitive processes (see
Worden and Brandl 1990, and the sources cited
therein). For research based on protocol analysis,
decision-makers are asked to think aloud as they
perform decision tasks, or they might be asked to
recount their thinking as soon as possible after
performing a decision task. Research subjects’
verbal reports of their thinking are data on their
decision processes. Protocol analysis of police
has not been conducted often, but several studies
demonstrate that it is both feasible and illuminating. For example, Stalans and Finn (1995) discovered that the perceptual and cognitive
processes that officers apply in interpreting and
evaluating domestic violence incidents vary by
experience: less experienced (“novice”) officers
tend to evaluate domestic violence situations in
terms of blame, while their more experienced
colleagues are more pragmatic, putting more
emphasis on the sparing use of arrest (and their
time) and on an assessment of the immediate
potential for further violence and the prospects
for prosecution. Both novice and experienced
officers, however, tended to assess situations as
typical or atypical. Stroshine et al. (2008) enumerate informal “working rules” that serve to
guide officers in performing their jobs, and they
observe that “different officers look for different
things and respond to suspects and situations
differently” (2008: 335). They describe the cues
P
P
3606
to which officers attend in forming suspicion
(e.g., time and place, and citizens’ appearance
and behaviors), the rules of thumb that they
apply (e.g., with respect to the seriousness of
violations, or the citizen’s demeanor), and more
general approaches that they take to their work
(seeking out or avoiding opportunities for traffic
enforcement). The potential of protocol analysis
might be best realized through the construction of
process models, which resemble flow charts.
Compared to the structural models that are normally estimated, process models are probably
more accurate representations of how people
make complex decisions, especially under conditions of ambiguity and uncertainty.
Finally, an overriding question revolves
around the normative dimensions of police decision-making and the use of police authority.
Research has little to say about the extent to
which, and the circumstances under which, police
use their authority well. Even research on the use
of physical force has, for the most part, lost
a focus on the improper use of force. Research
will be far more useful if and when it can draw
empirically grounded conclusions about when
police should use what authority and why.
Research will also be more useful if and when it
can establish the connections between what the
police do – what authority they exercise and how
they exercise it – and how it is subjectively experienced by citizens; survey research has dwelt on
citizens’ perceptions, but it has not estimated the
relationships between citizens’ perceptions and
overt police acts.
Recommended Reading and References
Bittner E (1970) The functions of the police in modern
society. National Institute of Mental Health, Bethesda,
MD
Bittner E (1974) Florence nightingale in pursuit of Willie
Sutton: a theory of the police. In: Jacob H (ed) The
potential for reform of criminal justice. Sage, Beverly
Hills
Bittner E (1983) Legality and workmanship: introduction
to control in the police organization. In: Punch M (ed)
Control in the police organization. MIT Press, Cambridge, MA
Police Discretion in Law Enforcement
Black D (1971) The social organization of arrest. Stanford
Law Rev 23:1087–1111
Black D, Reiss AJ Jr (1967) Patterns of behavior in police
and citizen transactions. In: Studies of crime and law
enforcement in major metropolitan areas, vol II, Sec I.
GPO, Washington, DC
Brandl SG, Stroshine MS, Frank J (2001) Who are the
complaint-prone officers? an examination of the relationship between police officers’ attributes, arrest
activity, assignment, and citizens’ complaints about
excessive force. J Criminal Just 29:521–529
Brown MK (1981) Working the street: police discretion
and the dilemmas of reform. Russell Sage, New York
Brown MK (1988) Working the street: police discretion
and the dilemmas of reform. Russell Sage, New York
Cuttler MJ, Muchinsky PM (2006) Prediction of Law
enforcement training performance and dysfunctional
Job performance with general mental ability, personality, and life history variables. Criminal Just Behav
33:3–25
Fyfe JJ (1979) Administrative interventions on police
shooting discretion. J Criminal Just 7:309–323
Fyfe JJ (1988) Police use of deadly force: research and
reform. Justice Quart 5:165–205
Harris CJ (2010) Pathways of police misconduct: problem
behavior patterns and trajectories from two cohorts.
Carolina Academic Press, Durham
Klinger DA (1994) Demeanor or crime? Why ‘Hostile’
citizens are more likely to be arrested. Criminology
32:475–493
Klockars CB (1996) A theory of excessive force and its
control. In: Geller WA, Toch H (eds) Police violence:
understanding and controlling police abuse of force.
Yale University Press, New Haven
Mastrofski SD, Worden RE, Snipes JB (1995) Law
enforcement in a time of community policing.
Criminology 33:539–563
Moore MH (2002) Recognizing value in policing: the
challenge of measuring police performance. Police
Executive Research Forum, Washington, DC
Muir WK Jr (1977) Police: streetcorner politicians. University of Chicago Press, Chicago
National Research Council (2004) In: Committee to
Review Research on Police Policy and Practices,
Skogan WG, Frydl K (eds) Fairness and effectiveness
in policing: the evidence. Committee on Law and
Justice, Division of Behavioral and Social and Behavioral Sciences and Education. Washington, DC:
National Academies Press
Paoline, Eugene A. III, William Terrill (2005) The Impact
of Police Culture on Traffic Stop Searches: An Analysis of Attitudes and Behavior. Policing: An International Journal of Police Strategies and Management
28(3):455–472
Reiss AJ Jr (1971) The police and the public. Yale
University Press, New Haven
Scrivner EM (1994) Controlling police Use of excessive
force: the role of the police psychologist. National
Institute of Justice, Research in Brief, Washington
Police Discretion in Providing Services and Assistance
Stalans LJ, Finn MA (1995) How novice and experienced
officers interpret wife assaults: normative and
efficiency frames. Law Soc Rev 29:287–321
Stroshine M, Alpert G, Dunham R (2008) The influence of
‘working Rules’ on police suspicion and discretionary
decision making. Police Quart 11:315–337
Terrill W, Paoline EA III, Ingram J (2011) Assessing
police use of force policy and outcomes. Michigan
State University, East Lansing
Terrill W, Paoline EA III, Manning PK (2003) Police
culture and coercion. Criminology 41:1003–1034
Toch H (1980) Violent men, Revth edn. Schenkmen,
Cambridge, MA
Van Maanen J (1974) Working the street: a developmental
view of police behavior. In: Jacob H (ed) The potential
for reform of criminal justice. Sage, Beverly Hills
Van Maanen J (1978) The asshole. In: Manning PK, Van
Maanen J (eds) Policing: a view from the street. Goodyear, Santa Monica, CA
Wilson JQ (1968) Varieties of police behavior: the management of Law and order in eight communities.
Harvard University Press, Cambridge, MA
Worden RE (1989) Situational and attitudinal explanations of police behavior: a theoretical reappraisal and
empirical assessment. Law Soc Rev 23:667–711
Worden RE, Brandl SG (1990) Protocol analysis of police
decision making: toward a theory of police behavior.
Am J Criminal Just 14:297–318
Worden RE, Shepard RL, Mastrofski SD (1996) On the
meaning and measurement of Suspects’ demeanor
toward the police. J Res Crime Delinq 33:324–332
Police Discretion in Providing
Services and Assistance
John D. McCluskey
Department of Criminal Justice, Rochester
Institute of Technology, Rochester, NY, USA
Overview
Consideration of police discretion as it relates
to the provision of services to the public is an
area of much speculation but little hard data. It is
widely recognized that police are called upon to
do much more than fight crime and enforce the
law (e.g., Cumming et al. 1965; Guyot 1991).
Patrol officers, for example, spend little time
arresting individuals for crimes in any given
work shift and are often observed not arresting
under circumstances when authorized to do so
3607
P
by the law. Nevertheless, the academic literature
focuses heavily on the decision to arrest and
the decision to use force, because they address
issues that are at the core of democratic governance: How does the state use its coercive power?
This leaves a substantial gap in understanding
what police choose to with regard to offering
help and assistance as necessary and as directed
by the public. The gap is especially apparent
with regard to enumerating and explaining the
actions police undertake outside the scope of
their legally mandated and more clearly
authorized law enforcement duties (Skogan and
Frydl 2004).
The issues surrounding police discretion in
helping and serving the public will be discussed
along several dimensions. First, the history of
police and their role, particularly in the United
States, will be considered. The mandate of police,
as a social control mechanism, offers an opportunity to understand broader patterns of police
behavior over the long term and how police relate
to and define their work. Second, service and helping behaviors will be defined. Since service and
helping constitutes, in the most rigorous analysis,
a variable, it is necessary to locate reasonable
boundaries and dimensions upon which service
and helping can be arrayed. No precise definition
will be offered, as technological changes and task
requirements continuously shape police work, but
a set of considerations will be explored. Third,
measurement of service including attempts and
future directions will be considered. Fourth, a set
of preliminary propositions regarding police exercise of discretion in allocating or engaging in
helping behaviors will be outlined. Fifth, the transformation of the police mandate toward recognition, if tacit, of the importance of service activity is
examined in light of problem-oriented policing
and the community policing movements and the
influence on discretion regarding helping behaviors. Sixth, the service and help behaviors of police
will be related to the specific client population of
mentally ill citizens whom which they very frequently encounter. Finally, the entry is summarized and future directions for consideration in
the areas of service and helping behavior and
police discretion are offered.
P
P
3608
History
US municipal police forces, according to some
historians, were established as general service
provision agencies and had wide-ranging
responsibilities including dealing with lost children, inspecting boilers, and feeding and housing
the homeless. Lane (1967) documents the primary
role that the Boston police, for example, played in
running soup kitchens and providing temporary
lodging for the homeless. This service was based
on the observation that “misery bred crime and
soup helped to prevent it” (Lane 1967: 193). Data
gathered by historians also indicate that activities
involving lost children, for example, were
recorded with some care, thus illustrating the
important nature of that task. Monkonnen (1981:
119) argued that the centralized communication
capacity of the police made them the primary
agency in assisting citizens with their problems.
Others have argued, more cynically, that police
were more easily understood as paid agents of
the dominant order than a public agency aimed at
help and service (Harring 1983).
Historical analysis derived from a Boston, MA
police officer’s diary entries in the year 1895 indicates the panoply of service provided to
constituents and the rarity of activity that explicitly
involved legal channels (Von Hoffman 1992).
Rather daily work would more closely resemble
regulatory attention to maintaining social order
and largely via service. Even crime-related entries
represented mobilizations by citizens who desired
the service of the officer in maintaining security or
apprehending wrongdoers. The focus on service is
likely due in part to some problems not being
easily defined as legal or criminal problems and
the cost that is borne, in terms of time lost, when an
officer has to take the time to invoke arrest, especially on foot. Nevertheless it is imperative to
recognize that police have traditionally been
asked to address a wide variety of problems as
a general service agency, which was probably
cast best by Egon Bittner’s (1990) hyphenated
summary of what becomes police business:
“something-that-ought-not-to-be-happening-andabout-which-someone-had-better-do-somethingnow.” Thus, there is an expansive definition,
Police Discretion in Providing Services and Assistance
broadly shared among the public, of what constitutes a situation that police ought to handle, even if
among the police some of those tasks are seen as
not police work or treated with disdain.
Kelling and Moore (1988) make a cogent
argument that three eras of US policing, the political, the reform, and the community problemsolving eras, can be mapped to changes in the
authorization and the function of the police. In the
political era, in which municipal police agencies
were formed, the police derived their legitimacy
from the community and their function was to
serve that community’s needs. In the reform era,
the authorization of the police was narrowed to
matters of law, and the police function was dominated by formal law enforcement. In the community and problem-solving era, which Kelling
and Moore were forecasting, it was expected that
there would be a return to service in the community and recognition of community needs as
authorizing police actions. Clearly, service and
helping is at the root of much police work, but its
emphasis has undergone dynamic change as
agencies have grappled with the role of police in
a changing society. In some ways, these differing
emphases of police agencies represent discretion
writ large, as managers have defined, refined, and
redefined “what police do” over time. Such institutional variation over time foreshadows contemporary interorganizational differences in coming
to grips with what services should be provided to
the public.
Fundamentals
Definition: What Is Service, Help, and
Assistance?
How to define police service and helping behaviors is a fundamental question that confronts
police researchers and police managers alike.
Perhaps it is useful to consider what is not service: Arresting a wrongdoer or using force
against a suspect, for example, do not appear to
be service. However, under careful scrutiny, this
quickly turns on an issue of perspective. If police
are acting at the behest of another when applying
an arrest or forcefully evicting an agitated spouse
Police Discretion in Providing Services and Assistance
from a household, then clearly such actions
constitute service, at least from the perspective
of the requesting citizen. Mastrofski and colleagues (2000) did much to outline police helping
behaviors, especially those in which police were
mobilized to use their legal powers against
another citizen. Furthermore, arrests of individuals who may injure themselves (e.g., the intoxicated or mentally ill) might, broadly defined,
constitute service and help.
The definition of service and help would seem
to hinge on using police power or expertise to
advance the well-being of a citizen or
a community through formal and informal
mechanisms. Service and help should not be
understood as a strict and formal application of
the law, law enforcement, or deterrence powers
police possess. Rather service should be considered to be dominated by, but not exclusively, the
informal application of police power. As noted
above, however, helping and service is dependent
upon perspective and can occasionally emanate
from arrests, citations, and other enforcement
actions.
A second approach to defining police service
and helping behaviors is to generate illustrations
by way of examples. Police researchers, for example, have engaged in studies of how police spend
their time and document what police do for citizens. These actions include taking reports; providing assistance to motorists; making referrals to
agencies (such as a juvenile in need of supervision,
mentally ill person in need of psychiatric care,
victim in need of counseling, or other person in
need of assistance outside of the police expertise),
transportation (e.g., for juveniles out after curfew),
and directed actions (banish, counsel, arrest) on
behalf of a requesting citizen; attending to traffic
accidents and directing traffic; counseling citizens
about self-protection or self-help actions for crime
prevention; and providing comfort and first aid to
victims. The list is extensive, illustrates great
variety, but it is certainly not exhaustive. Police
agencies, however, spend little time measuring or
systematically rewarding the extent to which
police engage in such activities. Thus, little is
known about the extent and variety of ways that
the public engages the police in service.
3609
P
A national snapshot, derived from the
National Crime Victimization Survey’s sampling
frame, has been collected periodically in the form
of the Police Public Contact Survey (PPCS). The
PPCS contacts approximately 60,000 citizens
over age 18 and asks them about contacts with
the police during the previous year. The survey
has been administered most recently in 2008,
where it was determined that police had approximately 40 million contacts with citizens aged 16
and older. Three categories of contact have broad
overlap with police service activities: police
response to a traffic accident (12 % of total),
police contact involving the reporting of a crime
or problem (21 % of total), and police contact
emanating from a request for assistance (6 % of
total). In all these, three contact types represent
almost 16 million or 40 % of the police-citizen
contacts estimated from the survey results (Eith
and Durose 2011). Though ill-defined, these
contacts represent a substantial proportion of the
public investment in public safety, yet little is
known about the content of these contacts.
Measuring Service and Help
What can be measured regarding police service
and helping behaviors? This question requires
a commitment to a level of analysis, and for
current purposes, it would be prudent to consider
face-to-face contacts between citizens and the
police as the unit for consideration. This allows
for aggregation to higher levels, such as neighborhoods and organizations for comparisons,
and allows for a more thorough consideration of
the encounter between police and the client as the
unit of work or service delivery.
This level of analysis leads to consideration
of two aspects of helping behavior, one quantitative, the other qualitative. With respect to quantitative help, clearly one is required to count the
presence or absence of a report taken, a referral
made, a citizen given first aid, or a child given
transportation to his family’s home. The quality
of help (service with a smile?) is less amenable to
counting but perhaps more influential in how
citizens experience police service. That one
P
P
3610
aspect of service is amenable to counting makes
the complex appear deceptively simple. What
should be counted? Mastrofski and colleagues
(2000) examined how police respond to requests
for help in terms of not fulfilling, partially fulfilling, and completely fulfilling requests by citizens
asking police to control another person. Thus an
enumeration of requests and whether they are
fulfilled could be a measure of police reactive
service to the public. However, much of what
police do may in fact be proactive or at least
self-directed service to, or in consultation with,
the public. For example, police may follow up
a case and refer a domestic violence victim to
a legal service, a shelter, or some other services as
necessitated by situational dictates, victim preferences, and available resources. This proactive
approach to service speaks to both counting as
something done in a particular case and also the
variety of actions that an officer may take.
Capturing variety requires deeper consideration than can be accommodated here, but knowing the possible referrals, resources available to
be accessed, and the needs of the individual
would be a useful starting point. Thus, accounting
for individual instances of services rendered
would require attention to both possibilities
(referrals to agencies is a good example, since
the proliferation and quality varies across cities
and even neighborhoods) and appropriateness.
Such an approach to measurement would depend
on a needs assessment among the service population under study and would be onerous except
in the most narrowly conceived execution, such
as the services provided by a crisis intervention
team to mentally ill individuals encountered. As
a general approach to measuring service, it would
appear to be too cumbersome to match what
police did with what they should have or possibly
could have done. Thus, only in specific and welldefined incident or client population types (e.g.,
the aforementioned provision of service to
a mentally ill citizen) would this be a feasible
approach. In most efforts to measure the quantity
of help, it would appear that counting actions by
the police (referrals, assistance offered, aid
given) would be the extent to which help could
be quantified. Questions of appropriateness and
Police Discretion in Providing Services and Assistance
whether needs were addressed become value
questions, which would require researchers to
know more about what a person needs than may
be possible except in all but a few research
settings.
As Mastrofski (1999) has noted, the quality of
service delivery is often as important as the outcome itself. A larger body of procedural justice
literature points to quality of treatment and
quality of decision-making as being vital elements for understanding police-citizen contacts
as antecedents of legitimacy and cooperation.
Thus quality of service delivery is separate from
whether requests are fulfilled or appropriate
referrals are made. Comfort of victims (Foley
and Terrill 2008), for example, could be considered a qualitative aspect of the service that police
offer to victims of crime. Similarly, respectful
treatment of citizens (Mastrofski et al. 1996) is
an important part of the quality of service delivery. Measurement of the quality of police-citizen
interactions (whether police listen to citizens,
whether they inform them of their decisionmaking thought process, whether citizens are
allowed to participate in police decision-making,
whether police are fair in their decision-making,
and so on) has been conducted both via in-person
observations and through post contact surveys.
The burgeoning literature on the quality of treatment during police-citizen contacts, even when
restricted to service type activities, indicates that
more just treatment yields citizens who cooperate
and comply with the police. Thus, although service contacts are little studied, they represent an
important interface where quality policing can
have substantial consequences for future support
and cooperation with the police.
Clearly there is variation in the nature, extent,
quality, and variety of services that police can
offer to citizens in their everyday encounters.
The simplest way to conceive of the encounter
is that a person was in need of something and
police either did nothing or they did something to
ameliorate the situation. As the preceding considerations illustrated, measuring help or service
is more complex and opens questions of appropriateness, alternative courses of action, extent of
help offered, and the manner in which services
Police Discretion in Providing Services and Assistance
are delivered. These dimensions should be
considered when the measurement of service is
undertaken. One may surmise that measuring
service, at the encounter level, is difficult to do
well and impossible to do perfectly. However, the
challenge should not inhibit researchers or police
managers, since this aspect of police work is
extensive and a deeper understanding would aid
in recognizing it, targeting it, explaining it, and
making it a more productive aspect of policing.
Explaining Service and Helping
If one had a measure of service and help directed
toward citizens, what might explain the choices
police make in the delivery of services and help?
Since the allocation of services by police, in some
sense, represents a legal (if informal) intervention, the literature on how police allocate legal
outcomes such as arrests or tickets will be
considered as a background for understanding
discretionary decision-making in the allocation
of service. Service could then be understood as
an extension of this behavior, and hypotheses
developed to explain arrest, use of force, or police
issuance of citations would provide a useful
framework when extended to service. Thus, the
discussion below adopts the following distinctions drawn from this more general literature:
need, legal factors and situational factors, officer
characteristics and attitudes, neighborhoods, and
organizations. These are arrayed from
proximal (at the encounter level) to the distal
influences (neighborhood routines, organizational resources) and discussed below.
Need
Mastrofski and colleagues (2000) conducted an
analysis involving 396 police-citizen encounters
in which citizens requested police help dealing
with another person at the scene. An important
area of consideration, regarding whether police
took actions, was the perceived need of the citizen. This research focused on youthful or elderly
status, gender of the requestor, gender of the
target, whether the target was intoxicated, and
seriousness of the situation as indicators of
3611
P
need. Interestingly, in the analysis of citizendirected police assistance, only whether the target
was intoxicated had an impact on whether police
fulfilled a citizen’s request.
Regardless of these initial empirical results,
several elements of need are considered here as
imperative for discussing, generally, what may
influence an officer’s decision to help a citizen.
The seriousness of the situation has dual meanings as a motive for police helping behaviors.
One is the seriousness related to legal standards,
treated below. The second meaning of seriousness would entail the social or physical consequences or harms facing a citizen, which may
have no relation to legal seriousness. Not precisely a “legal” variable, the police would likely
tend to ameliorate serious problems via helping.
Intoxication of citizens and apparent mental illness are thus likely indicia of need that would
motivate help or service from the police.
The mobilization of the police may be aligned
with need as well and may be important for
predicting the level of service received in an
encounter. Predictions of greater service in proactive encounters or reactive encounters, holding
other elements constant, are not straightforward.
Where police are mobilized by citizens (i.e., reactive encounters), their presence has greater legitimacy for intervention, especially legal
intervention (e.g., Reiss 1971). The reactive
encounter may also signal a greater need for
resolution in the form of help or service to the
citizens. There does not appear to be a clear foundation for whether this legitimacy translates into
more formal solutions or more service solutions.
Proactive encounters, in contrast, may be more
likely to yield services as police initiate such
encounters to help, assist, or otherwise provide
service to citizens in need. More simply, mobilization of police is likely linked with need, but it is
unclear if mobilization has a clear relationship to
whether police provide service.
Legal and Situational Factors
Several legal and situational factors likely shape
whether police engage in helping or service
behaviors in a particular instance. First, to the
extent there is more evidence of wrongdoing,
P
P
3612
police likely feel activated and obligated to
demonstrate a response. In work by Mastrofski
and colleagues (2000), when there was greater
evidence of wrongdoing, police tended to
respond to requests made by citizens desiring
a consequence for the wrongdoer. Second, with
regard to more serious legal implication, police
are likely to help, to a point, and at the highest
level of seriousness, they are likely to have less
discretion to bring informal sanctions or solutions
to bear on the problems they confront. Thus one
may speculate an inverted “u” response shape,
whereby at low and very high seriousness police
have little discretion. In the middle ranges where
discretion is high, they likely would engage in
more helping and assisting behaviors. Similarly,
the prior history of an individual, in terms of
misbehavior, would likely influence whether
police engage in helping behaviors or revert to
more formal handling of situations. In the modern
age of technology, the requestor is likely to be
“run” against computerized files and evaluated in
terms of history as well, both as a complainant
and as a potential target against whom which help
is mobilized.
Situational factors cover a wide variety of
variables which may influence discretion. For
purposes of this discussion, three are considered:
race, social class, and demeanor of the target.
Race and ethnicity may be linked with
whether police offer more services or help to
citizens. Research on arrests, for example, indicates that minorities are more likely to be
arrested. In light of those findings, one might
expect that minority citizens will receive less
informal help and service from the police.
Social class is also linked to criminal justice
outcomes, and as such, we might expect that
lower social class individuals are less likely to
receive services relative to their more well-off
counterparts. Clearly neither race nor class
should matter, but research on other facets of
discretionary police decisions lead to hypotheses
regarding race and class influencing decisions to
give services to individuals.
Demeanor of the citizen who is seeking or is in
need of service is likely more important than
either race or social class. Demeanor refers to
Police Discretion in Providing Services and Assistance
whether a citizen is respectful or disrespectful
toward the police. In many settings, across eras,
demeanor has shown to be a relatively consistent
and moderately strong predictor of what legal
agents do, in the sense that those citizens with
disrespectful demeanors tend to get more negative outcomes (more disrespect from the police,
more likely to be arrested, and so on). Thus it is
expected that citizens with disrespectful
demeanor will receive less service from the
police, especially if that service is conceived of
as a reward such as provision of service or help.
Individual Officer Characteristics
Evidence regarding the effect of officer characteristics on discretionary decisions is mixed but
certainly it points to no more than a modest to
weak effect on legal outcomes. With regard to
how officer discretion might be influenced by
personal characteristics, we focus on four: education, experience, gender, and officer’s attitudes. Officer education has been shown to have
limited effect on performance. With regard to
service, it might be expected that officers with
greater educational attainment might be more
inventive and expansive in their approaches to
problems requiring help or service as a possible
solution. Research examining whether officers
comfort citizens indicates, however, that officers
with college degrees provided significantly less
comfort to victims when compared to those without college degrees (Foley and Terrill 2008).
Similar to the prediction regarding college
education, experience, which is arguably the
best teacher, should allow for more inventive
and helpful behaviors to be routinely developed.
These would manifest as possibilities for solving
presenting situations and become routine for
officers with longer times on the job. Again,
however, the research of Foley and Terrill
(2008) found officers with more experience less
likely to comfort victims. Gender is controversial, in the sense that associating female officers
with a greater capacity for helping appears, on its
face, sexist. However, Foley and Terrill
(2008) found that female officers were no more
likely to comfort victims than their male counterparts. Nevertheless, gender may be a proxy for
Police Discretion in Providing Services and Assistance
attitudes regarding the expansiveness of the
police mandate and reflect both socialization
and attitudinal patterns.
Officer attitudes toward work vary and would
appear to be important for determining how individuals choose to execute facets of their work.
For example, Michael Brown (1988) argued that
selectivity and aggressiveness were two key
dimensions for typifying officer styles. Among
these styles some officers would be much more
likely to narrowly define the appropriate tasks for
“police work.” In fact, Brown argued that there
was an identifiable “service style” in which some
officers adopted a more helping role rather than
strictly law enforcement roles. Similarly,
William K. Muir’s observations of officers’
orientations toward service and helping, in the
1970s, found officers who defined some work as
civil issues (e.g., landlord-tenant disputes) and
explicitly sought to avoid such situations. The
emergence of the community policing model in
the 1980s and 1990s more broadly defined police
mandates and thus provides cover for officers to
adopt more helping orientations toward their
work.
Neighborhood Context
The neighborhoods in which police encounter
citizens provide the backdrop or setting for
face-to-face interactions. Aspects of neighborhoods have been shown to influence what police
do during encounters. Those neighborhoods that
are characterized by higher poverty, often measured by researchers as concentrated disadvantage which combines indicators of race, poverty,
and employment, have been linked with greater
likelihood of police arrests and use of force
against citizens in prior research. These same
neighborhoods are likely to demand greater service from the police. In the aggregate they may
receive greater numbers of police helping services. However, at the level of individual interactions between police and citizens, it is likely
that those in disadvantaged neighborhoods
receive less service relative to those in wealthier
or more advantaged neighborhoods. In part this
could be a consequence of police formal authority undermining many of the sources of local
3613
P
social regulation. Or it could be that there is an
absence of collective efficacy, upon which police
can draw informal solutions that can be enforced
by informal social control networks in these
neighborhoods. In either scenario the literature
leads to the prediction that police will, when
confronted with similar situations, exercise less
helping behavior in the more disadvantaged
neighborhoods.
Organizational Context
In a classic study of police at the organizational
level, James Q. Wilson’s Varieties of Police
Behaviors examined how departments molded
their responses to the political environment. It is
telling that in low-crime, more wealthy locales,
the police were considered primarily a service
function. Though the research is more than
40 years old, the contrast between Oakland, CA,
and Nassau County in Long Island and the legalistic style (defining work in terms of application
of the law) and the service style (defining work in
terms of what the police can do for the public and
what the public wants) anticipated the movement
toward community policing and the explicit recognition of “what the public wants.” This
research also indicates that organizational commitment to service likely influences officers’ discretion in their day-to-day work.
Contemporary organizational adaptation to
servicing public needs may be most comprehensively studied in Chicago, IL, (Skogan 2006)
where neighborhood policing was specifically
aimed at understanding community needs and
trying to accommodate them. The Chicago Alternative Policing Strategy (CAPS) explicitly
tapped the community to assess needs and
targeted police activity toward problems such as
graffiti abatement projects and efforts at removing abandoned cars from neighborhoods. This
menu of responses is a far departure from the
strictly conceived law enforcement mandate and
illustrates how variation at the organizational
level is likely to influence police discretion. The
mechanism by which this operates is the
expanded mandate of what police should do
coupled with an effort to understand community
needs. Officers working under such regimes are
P
P
3614
likely to develop a larger palette of helping
responses when compared to those working in
more legalistic environments.
At the organizational level, one could thus
measure variations in how organizations array
themselves across a legal/service continuum.
Organizational commitment to service could be
measured in a variety of ways including information sources collected in assessing needs, varieties of services offered by police, interfaces by
police agencies with other service agencies, and
the extent to which performance measures are
guided by service roles of the agency (i.e., does
the agency count “problems solved”).
It should be clear that police managers and
researchers alike have little direct experience in
measuring service and predicting “who gets
what.” For example, among those variables considered to be important characteristics of officers,
experience and college education are both hypothesized to relate to greater levels of helping. However, at least with respect to comfort to victims,
Foley and Terrill (2008) drew the opposite conclusion. In more recent research, Rossler and
Terrill (2012) examined police responses to citizen
requests to file reports, act on their behalf with an
agency, provide information, or provide physical
help. In these analyses, officers’ college education
had no impact on whether or the extent to which
officers fulfilled the request. But college-educated
officers were less inclined to provide an explanation when denying a request. This is a sign that
caution should be exercised in forecasting how
police may exercise their ability to help in faceto-face encounters and a clear indicator for more
data collection, measurement, and theory development. Although some propositions are introduced
here regarding need, situational and legal factors,
officer characteristics, and neighborhood and
organizational context, none of them have been
sufficiently tested against observations to a point
where one could make statements of any empirical
regularity with confidence. The absence of this
knowledge is welcome territory for future scholarship but currently a barren landscape from which
to make suggestions for police managers, policy
for organizations, or to yield strong theory from
which to begin analyses.
Police Discretion in Providing Services and Assistance
Current Issues
Considering Problem-Oriented Policing,
Community Policing, and Service
Responsibilities
Herman Goldstein (1990) observed that when the
unit of work patrol for patrol was defined as
a “case,” then police often missed larger problems
that are frequently the underlying generator of
calls for service. Such an underlying problem can
generate many cases which are resolved by
a response, a report, and perhaps an arrest.
A drug market, for example, might generate calls
of nuisance behaviors and police may respond and
disperse suspected dealers, take a report from
a complainant, or perhaps make an arrest of an
individual for possession, then return to duty. If the
unit of work is the call, then this call is satisfied.
If the unit of work is defined as a problem, then the
problem has not been addressed by the actions
described. The problem-oriented policing
approach outlines a variety of sources of problems
such as homelessness, bar hours, public health
hazards such as litter and trash, and a broader
palette of incivilities both physical and social that
may be considered within the sphere of problems.
These larger categories may have an ambiguous
relationship to the legal power of police and
therefore are easy for organizations and officers
to avoid, if so inclined. However, once engaged in
problem-oriented policing that is consistent with
the scanning, analysis, response, and assessment
approach outlined by Eck and Spelman (1987), the
likelihood that services and helping behaviors and
referrals will be accessed increases. For example,
interagency partnerships can form the backbone of
these approaches as illustrated by the Beat Health
program in Oakland, CA. Here the Oakland police
department addressed decay and disorder via partnerships with other public agencies as well as
private place managers such as landlords. In this
case the service provided is creating a more
ordered and civil public space in which conventional citizens may engage in legitimate daily
pursuits. Conversely, confronting the problems in
the local spaces discouraged the illegitimate use of
spaces by, for example, street-level drug dealers.
Thus a legal problem was solved by a service
Police Discretion in Providing Services and Assistance
approach. So the gray area between civil and criminal was straddled in the Oakland experiment as
documented by Mazerolle and Roehl (1999),
whereas, two decades earlier, Muir (1977) working with the Laconia (a pseudonym) department
found that officers tended to deny responsibility
for cases that were civil in nature.
The blending of the civil and criminal authorities within local law enforcement, from a legal
perspective, may be another area for future consideration. The nature of public regulation, compliance, and obedience is dramatically different
in the 2010s than in the 1970s when the theoretical foundations for much criminal justice
research on police and their behavior was being
established (Reiss 1971; Brown 1988; Muir
1977). This transformation is undoubtedly related
to the community policing movement and its
explicit desire to deactivate the legalistic mindset
that defined police work in terms of criminal law
(e.g., Kelling and Moore 1988). More expansive
definitions of police work will broaden what
police can do and will also provide opportunities
to study and understand why police make choices
to help or provide service in some instances, but
not others.
Populations in Need of Services:
The Mentally Ill
Lurigio and colleagues (2008) have outlined the
contemporary scene with regard to the extent to
which police process the mentally ill and point
out that the jail systems in Chicago, New York
(Rikers Island), and Los Angeles represent the
largest mental health facilities in the United
States. Consideration of the mentally ill clients
as persons in need of help mirrors some of the
civil/criminal tension that has characterized the
consideration of police service and helping
behaviors. The police contact with the mentally
ill is not a new subject for academic study as these
contacts have long been considered an important
part of police repertoire and a very difficult problem to handle with only legal tools such as arrest
and coercion (Muir 1977). In many ways, bridging service and legal processing for this population has been a continuous challenge for streetlevel policing. Adoption of crisis intervention
3615
P
teams, collaboration built around psychiatric services and police, and emphasis on training officers for the difficulties inherent in encounters of
the mentally ill point to efforts to shape discretion
with respect to this particular population. Interestingly the movement has first recognized that
police are often the first line of contact for mentally ill individuals, and to handle this exigency,
agencies have trained and developed resources
targeted specifically toward this population’s
needs. Equipping police with greater structure
and resources shapes discretion but broadens possibilities of what police can do.
Summary and Future Directions
Police have always engaged in helping and service behaviors. Despite this, modern scholars and
managers alike are hard pressed to measure, theorize, and evaluate the effectiveness of these
actions. This stems from recognition that the
mandate upon which helping behaviors rests is
clearly not as strong as the law enforcement mandate, which is rooted in law and linked to fundamental issues involving the exercise and control
of coercive force. Thus systematic study of
helping has been largely overlooked,
unmeasured, and, unfortunately, unrewarded. As
the public comes to expect more service from
police agencies, especially in times of economic
hardship and social consequences that accompany it, there is likely to be greater pressure to
systematically measure and evaluate this aspect
of police work. Among police researchers, similarly, there is likely to be a movement to test
propositions about “who gets what” from police
in terms of services rendered. As noted here it is
possible to speculate about how neighborhood
characteristics, citizen race and ethnicity, and
citizen social class relate to service delivery
patterns; in truth, these are open research
questions yet to be thoroughly addressed. Nevertheless, efforts to adopt the community and
problem-solving approaches have driven police
organizations and the officers within them to
embrace a greater service role, absent information on how such service is delivered. Given this
P
P
3616
situation it is likely that much research will be
targeted at best practices of service delivery,
especially with regard to services targeted at special populations, such as the mentally ill.
Related Entries
▶ Community Policing
▶ Measuring Police Performance
▶ Police Discretion and Its Control
▶ Police Discretion in Law Enforcement
▶ Police Family Violence Services
▶ Police Legitimacy and Police Encounters
▶ Policing Special Populations
Recommended Reading and References
Bittner E (1990) Florence Nightingale in pursuit of Willie
Sutton: a theory of the police. In: Bittner E (ed)
Aspects of police work. Northeastern University,
Boston
Brown MK (1988) Working the street: police discretion
and the dilemmas of reform. Russell Sage, New York
Cumming E, Cumming I, Edell L (1965) Policeman as
philosopher, guide and friend. Soc Probl
12(1):276–286
Eck JE, Spelman W (1987) Problem-solving: problemoriented policing in Newport news. Police Executive
Research Forum, Washington, DC
Eith C, Durose MR (2011) Contacts between police and
the public, 2008. Special report. Bureau of Justice
Statistics, Washington, DC
Foley T, Terrill W (2008) Police comfort and victims. Vict
Offender 3(2):192–216
Goldstein H (1990) Problem-oriented policing. McGrawHill, New York
Guyot D (1991) Policing as though people matter. Temple
University Press, Philadelphia
Harring S (1983) Policing a class society. Rutgers University Press, New Brunswick
Kelling GL, Moore MH (1988) The evolving strategy of
policing. Perspectives on policing, vol 4. National
Institute of Justice, Washington, DC
Lane R (1967) Policing the city: Boston 1822–1885.
Harvard University Press, Cambridge, MA
Lurigio AJ, Smith A, Harris A (2008) The challenge of
responding to people with mental illness: police officer
training and special programmes. Police J 81:295–322
Mastrofski SD (1999) Policing for people. Police Foundation, Washington, DC
Mastrofski S, Snipes J, Supina A (1996) Compliance on
demand: the public’s response to specific police
requests. J Res Crime Delinq 33:269–305
Police Family Violence Services
Mastrofski S, Snipes J, Parks R, Maxwell C (2000) The
helping hand of the law: police control of citizens on
request. Criminology 38:307–342
Mazerolle L, Roehl J (1999) Controlling drug and disorder
problems: a focus on Oakland’s beat health program.
National Institute of Justice, Washington, DC
Monkkonen E (1981) Police in urban America,
1860–1920. Cambridge University Press, New York
Muir WK (1977) Police: streetcorner politicians. University of Chicago Press, Chicago
Reiss AJ Jr (1971) The police and the public. Yale
University Press, Connecticut
Rossler MT, Terrill W (2012) Police responsiveness to
service-related requests. Police Quart 15(1):3–24
Skogan W (2006) Police and community in Chicago: a tale
of three cities. Oxford University Press, New York
Skogan W, Frydl K (eds) (2004) Fairness and effectiveness in policing: the evidence. The National Academies Press, Washington, DC
Von Hoffman A (1992) An officer of the neighborhood:
a Boston patrolman on the beat in 1895. JSoc Hist
2(Winter):309–330
Wilson JQ (1968) Varieties of police behavior: the
management of law and order in eight communities.
Harvard University Press, Cambridge, MA
Police Family Violence Services
Robert C. Davis and Sarah Greathouse
RAND Corporation, Arlington, VA, USA
Overview
In recent years a number of programs have been
developed in which the initial police response to a
family violence incident is followed by a visit
from a trained team of officers or by an officer
and a social worker. These “second responders”
attempt to help victims find long-term
solutions to recurring abuse (e.g. see Dean et al.
2000; Mickish 2002).
Second responder programs are based on the
premises that family violence often recurs and
that victims are likely to be especially receptive
to crime prevention opportunities immediately
following victimization. That is, there is
a “window of opportunity” during the first hours
or days after a crime during which victims feel
vulnerable and are willing to seriously
consider behavioral and lifestyle changes
Police Family Violence Services
(Davis and Smith 1994; Anderson et al. 1995).
The response often takes the form of a home visit
but other times consist of a phone call to the
victim. The team provides the victim with
information on services and legal options and
(in some models) may warn those perpetrators
present at the follow-up of the legal consequences
of continued abuse. The purpose of working
directly with the victims is to reduce the
likelihood of a new offense by helping them to
understand the cyclical nature of family violence,
develop a safety plan, obtain a restraining order,
increase their knowledge about legal rights and
options, and provide shelter placement or other
relocation assistance. A secondary aim of the
intervention with victims may be to establish
greater independence for victims through
counseling, job training, public assistance, or
other social service referrals. The purpose of
conversations with abusers is to ensure that they
understand that assaulting an intimate is
criminal and that further abuse will result in
(additional) sanctions.
Key Issues
History of Family Violence Officers
The role of police family violence services has
evolved over time. Prior to the 1970s, law
enforcement’s response to family violence was
significantly limited. Departments rarely
contained specialized teams to respond to
incidents of family violence. Responses by
general patrol to family violence incidents were
often slow or nonexistent (Ford 1983; Manning
1988). When police did respond to service calls,
departmental policies often dictated that officers
should only attempt to diffuse the immediate
situation and subsequently refer the individuals
to social service agencies; arrests of the
perpetrator were rarely made and follow-ups to
the initial call for service were rarely conducted
(Hutchison et al. 1996; Parnas 1967). This lack of
attention to family violence issues by law
enforcement reflected both societal beliefs about
family violence and traditional law enforcement
culture. Family disputes were commonly viewed
3617
P
as private affairs, and it was believed that police’s
role in handling family disputes should be one of
mediators and references to social services
(Breslin 1978). As such, police officers often
viewed responses to domestic disputes as not
“real” police work (Buzawa and Buzawa 2003).
Surveys of police officers’ attitudes toward
domestic violence, specifically, showed that
police believed arrests to be futile because
victims would rarely follow through to press
charges and that arresting the male head of
household would negatively impact the family
as a whole. Police also endorsed the commonly
held but incorrect belief that responses to
domestic disputes placed the officer in
greater danger compared to other calls
(Buzawa and Buzawa 2003). As a result, law
enforcement devoted little attention to family
violence issues; even less attention and resources
was devoted to taking actions that reduced further
incidents of violence.
The law enforcement response to family
violence began to become significantly more
proactive in the 1970s and 1980s. Several events
were catalyst for the change. First, victim’s rights
advocacy groups and feminist groups began to
challenge the police’s lack of forceful responses
to domestic disputes and their failure to protect
female victims of domestic violence (Buzawa
and Buzawa 2003). Second, a series of legal
cases in the 1970s and 1980s ruled that police
departments could be held liable and face
financial penalties if they failed to protect victims
of family violence (Bruno v Codd 1977; Tracey
Thurman et al. 1984). Finally, empirical research
around this same time began to demonstrate that
the actions taken by police had the potential to
reduce incidents of family violence.
One of the most influential empirical studies
of police response to incidents of family violence
was an evaluation of police departments
that employed violence family crisis intervention
units (FCIU’s) (Bard 1975). Originally
developed by psychologists, officers in this
program received training in mediating and
diffusing domestic disputes. While the officers
still responded to all calls for service, they were
specifically called in to handle family violence
P
P
3618
calls. The National Institute of Justice sponsored
a training and demonstration program that
employed 10 family crisis intervention units in
police departments across the country. The
evaluation of these programs reported some
positive effects of FCIU’s, including decreases
in arrests and officer injuries (Bard 1975).
Although this evaluation also found that calls
for service to family violence incidents increased,
this finding was explained as evidence of the
FCIU’s increasing victims’ faith in the ability of
the police to effectively handle incidents of
family violence. As a result of the reported
success of these programs, many major US police
departments began adopting either FCIU’s or
specialized units comprised of law enforcement
and social service workers (Liebman and
Schwartz 1973). Often funded with federal
funds, these officers were specially trained in
handling victim needs and problems and in
keeping themselves safe during home visits.
However, responses to FCIU’s were not all positive. Liebman and Schwartz (1973) reported that
New York City’s highly praised FCIU’s program
had resulted in increased homicides and domestic
disputes. Critical research findings, coupled with
strategic difficulties in implementing specialized
units, led many departments to completely cut or
severely limit the role of family violence
response units.
In the 1980s, police departments shifted focus
from specialized family violence policing
services to arrest and punishment of those
individuals that perpetrate family violence.
Several research studies were major forces in
changing the manner in which police responded
to family violence. The most prominent of these
studies was an experimental evaluation of police
responses to domestic violence conducted in
Minneapolis, Minnesota (Sherman and Berk
1984). In this study, police who responded to
domestic violence calls were randomly assigned
to handle the dispute in one of three ways:
(1) separate the parties (2) mediate the dispute
(3) make an arrest. The researchers reported that
making an arrest resulted in significantly fewer
further domestic disputes between the couple.
Although
the
study
received
some
Police Family Violence Services
criticism from the scientific community
(Binder and Meeker 1988; Lempert 1989), the
results from this experiment were highly
publicized and spurred mandatory arrest laws in
many jurisdictions and arrest oriented policies in
police departments across the country. This focus
on deterring future violent behavior through
arrest took precedent over employing
specialized family violence units or secondary
responder programs.
Although the criminal justice system has been
largely focused on reducing repeated incidents of
family violence through arrest and prosecution,
in recent years there has been recognition that
these policies may not be enough to effectively
address and reduce incidents of family
violence. Research studies have not always
reported positive effects for pro-arrest policies
(Dunford et al. 1990; Hirschel et al. 1992). In
addition, despite a strong emphasis on pro-arrest
policies, prior research reports a high chance of
revictimization for victims of domestic violence,
particularly immediately after the original
incident of abuse (Lloyd et al. 1994). This
research raised recognition that more needed to
be done to assist and empower victims of family
violence. As a result, a growing number of
police departments began to adopt or re-adopt
specialized family violence service units that
provide secondary responses to victims following
the initial call for service. Although the total
number of such programs is not known, we do
know a number of specialized programs across
the United States and Great Britain have been
implemented and evaluated. The research
examining the effects of these specific programs
will be discussed later in this article.
In recent years, law enforcement has also
become increasingly aware of the need for
specific units to address a special type of family
violence, elder abuse. Traditionally, elder abuse
was not considered a serious criminal problem.
Law enforcement officers did not often receive
specialized training to handle reports of elder
abuse, and departments rarely seriously pursued
elder abuse cases (Plotkin 1988). Recent research
has demonstrated that incidents of elder abuse are
much more prevalent than previously believed
Police Family Violence Services
(Tatara et al. 1997). As a result, a growing
number of departments have established
specialized units to handle cases of elder abuse
(Heisler 2000). Because cases of elder abuse are
often complex, specialized law enforcement
often coordinates with local prosecutors and
social service agencies to investigate reports of
elder abuse and work with the victims.
Training of Family Violence Officers
Historically, police officers have received very
minimal training about how to handle calls to
incidents of family violence. A survey of law
enforcement’s training practices during the
1970s revealed that most often police officers
received a limited amount of training on handling
family violence during a half-day to 1 day
training on responding to calls with disturbed
individuals (Buzawa and Buzawa 2003). This
module of training addressed a variety of
situations and didn’t solely focus on handling
family disputes. Many of the training programs
also simply reinforced the idea that police
officers should only play a minimal role in
handling incidents of family violence. Once on
the job, officers rarely received continued
training on handling family violence. The most
common method of educating young officers
came through observations of more experienced
officers’ responses to incidents of family
violence.
Today, more structured training programs
focusing on intervention strategies, arrest
policies, and attitudes toward domestic/family
violence are in existence. The International
Chiefs of Police and other national organizations
have produced recommended requirements for
domestic violence training, and an increasing
number of states are requiring their law
enforcement officers to receive family violence
training (Miller 1997). Although officer training
on domestic violence may be more detailed, the
amount of training officers typically received has
not significantly increased. One survey found that
on average officers receive 10 h of initial
domestic violence training. Once in the field,
most officers are not required to shadow officers
who are experienced in handling family violence,
3619
P
and the quality of on-going training programs
can vary significantly, from short videos to
in-class trainings with role playing activities
(Buzawa and Buzawa 2003). A growing number
of states are also requiring officers to receiving
training on elder abuse (Heisler 2000). Although
national organizations such as the Police
Executive Forum have distributed recommended
practices for training officer on elder abuse, little
information exists on the training practices of
specific departments. Little information also
exists on the typical methods used to train
individuals within any specialized family
violence or secondary responder units. Because
there are not standardized methods for operating
specialized family violence service units,
methods of training these team members vary
by department.
Functions of Domestic Violence Officers
Redlands, CA Police Department had a robust
domestic violence officer program that was the
subject
of
an NIJ-funded
evaluation.
Consequently, detailed information on what
these officers did are contained in the evaluation
report. The following description of the functions
of family violence officers draws heavily on the
report by Davis et al. (2010).
In Redlands, a team of officers, including
a trained female domestic violence detective,
attempted to visit households within either 24
h or 7 days of a domestic complaint depending
on the severity of the incident. According to the
officers conducting the second responses, contact
was made with the victims at their homes in 84 %
of the cases in which it was attempted. In cases
where the home visit attempt was not successful,
literature was left with information about
community services.
The visits typically lasted 30–45 min,
depending on the victim’s receptiveness to
assistance. The goals of home visits were to
ensure that the victim had information about
and access to resources and services, to answer
any questions they had about the complaint or the
justice process, and to encourage a sense of trust
in the police and the criminal justice system as
a whole.
P
P
3620
A written protocol guided the officer or
officers making home visits. The visits began by
the officer talking to the victim about the recent
incident and any immediate safety concerns that
she had. The officer discussed with the victim the
nature of domestic violence and the very real
possibility that the incident she experienced
would recur if no action was taken. The officer
tried to make the victim understand that the
police department took the matter seriously and
was there to assist her. She also asked the victim
a series of questions about her relationship with
the abuser, history of abuse, and the presence of
children and weapons in the home.
Once preliminaries were taken care of, the
second response officer tried to ensure that the
victim had information about resources and
services; offered practical assistance; worked
with the victim to develop a safety plan; and
instructed the victim in how to document future
abusive or stalking behaviors. Before leaving,
the officer provided the victim with a written
description of local resources to assist domestic
violence victims, including housing relocation,
counseling, domestic violence shelters, medical
help, civil legal assistance, information about
the criminal justice process, aid in applying for
an order of relief, and emergency financial
assistance. Referrals were most often made to
counseling programs or parenting classes;
smaller numbers of victims were referred to
shelters, civil legal assistance, assistance in
obtaining a restraining order, and district attorney victim advocates.
One concern about conducting second
response visits was that the visit might trigger
anger in the perpetrator. To avoid this, there was
an intentional practice to call ahead to make sure
perpetrators were not present when the officers
came to the home. However, since officers were
not always able to reach victims by phone ahead
of time, there were some instances in which the
perpetrator was there when the officers arrived.
While partners were home during just a handful
of visits, nearly half (46 %) of victims said in
subsequent interviews that their partners
were aware that the visit occurred. Of these,
Police Family Violence Services
approximately one in four (28 %) reported that
their partner had a negative reaction to the visit.
The Chicago Police Department has a similar
second responder program for victims of elder
abuse. Each Chicago police district has
a designated elder abuse officer whose job it is
to provide services to those victims. During the
course of their home visits, the elder abuse officers ensure that victims’ medical and physical
needs (food and shelter) are being met. When
they are not, officers arrange for temporary
shelter. They make referrals to social service
programs, including counseling, independent
living services, Meals on Wheels, and programs
for Alzheimer’s patients. Officers also enroll
victims in emergency identification bracelet
program so police responding to future incidents
can quickly grasp history and information on
victim, assist victims in completing forms to
receive state compensation, and provide
assistance getting to court. Finally, a report is
made to adult protective services.
Effectiveness
There have been a number of high quality evaluations of second responder programs yielding
conflicting results. Several randomized field
trials of second responder programs were
conducted in New York City public housing projects. Each tested the same intervention model:
persons who reported family violence to the
police were randomly assigned to receive or not
to receive a follow-up visit from a domestic
violence police officer and a social worker. This
follow-up visit was not immediate, as is the case
with most second responder programs, but
occurred an average of 2 weeks later.
The sample for one of the studies
(Davis and Taylor 1997) included instances
where someone had called the police in response
to a family violence incident (this could be
violence between romantic intimates, sibling
violence, elder abuse, or other forms of violence
between persons related or living under the
same roof). The incidents were minor in nature
(only 7 % of the incidents resulted in arrests and
just 14 % of victims reported any form of injury).
Police Family Violence Services
Four hundred and thirty-five victims were
randomly assigned to receive a home visit as
a follow-up to the patrol response. The control
group received only the initial police patrol
response. Additional calls for police services
were tracked for both groups over the next 6
months. At the end of the tracking period,
researchers interviewed victims to ask about
new abuse, about satisfaction with the police
response, and about victims’ knowledge and use
of social services.
According to law enforcement records,
households that received the home visit
intervention were more likely to call the police
during the subsequent 6 months than households
that did not receive the interventions. Yet,
according to victim survey data, there were no
differences between the two groups in abuse
during the 6 months following the trigger
incident. In the literature on the effectiveness of
arrest on curbing violence, victim reports and
calls to the police usually are both treated as
imperfect indicators measuring an underlying
construct of actual violence. However, the two
measures clearly are not synonymous. Many
family violence victimizations are not reported
to the police. Davis & Taylor interpreted this
pattern of results to mean that the experimental
interventions did not affect actual violence levels
but did increase victims’ confidence in the police
and made victims more willing to report violence
when it occurred.
A second experimental investigation (Davis
and Medina 2001) of the same intervention was
conducted several years later, this time using
a sample of 402 public housing residents who
had reported elder abuse incidents to the police.
Like the cases in the first field test, incidents in this
study were also relatively minor (5 % of the
abusers were arrested, just 4 % of victims reported
any injuries, and in only 22 % of the cases was
a crime alleged to have occurred). Once again, law
enforcement records for these households were
tracked for the next 6 months. As in the first
experiment, it was found that victims who
received the home visit intervention called the
police sooner and more often than controls.
3621
P
Pooled analyses of these and a third
unpublished experiment indicated that the
interventions were associated with an increase
in reporting of new abusive incidents not only to
authorities (which could indicate simply greater
confidence in the police), but also to research
interviewers (Davis et al. 2006). The New York
field tests suggested that second response
programs might actually increase the likelihood
of new abuse.
Going into these studies, it had been assumed
that the effects of the interventions would be to
empower victims through information about their
situation, available services, and legal options.
The program logic model posited that new
abuse would decline as victims extracted
themselves from self-defeating relationships or
worked with social services and criminal justice
staff to develop strategies to end the abuse while
staying in the relationship. However, researchers
in the New York studies found no evidence that
those who received the interventions were more
likely to avail themselves of social or legal
services, so the intervention could not have
worked – at least not in the way intended.
Some evaluations of other second responder
programs found results similar to the New York
field tests. A subsequent study by one of the
authors of the New York evaluations, this one
conducted in Redlands, CA, found that households that received a second response had worse
outcomes on seven measures of new abuse,
although the results were not statistically
significant (Davis et al. 2010). Two studies
conducted in New Haven, CT by Stover and her
associates (Stover et al. 2009, 2010) found
between them that persons who received
a second response were more likely to call the
police again, but less likely to report new abuse
on victim surveys. A study by Hovell et al.
(2006) found more reports of abuse to the police
among victims who received a second response.
Other studies however, reported a positive
effect of second responder programs. Greenspan
et al. (2003) found that victims who received
a second response in Richmond, VA were less
likely to report victimization on a subsequent
P
P
Model
3622
Study name
Davis and Taylor, 1997
Taylor, n.d.
Davis and Medina, 2001
Davis et al., 2007
Pate et al., 1992
Casey et al., 2007
Hovell et al., 2006
Stover et al., unpublished (1)
Fixed
Random
Police Family Violence Services
Statistics for each study
Std diff
in means
Standard
error
p-Value
0.10982
−0.07428
0.28159
0.02293
0.10586
−0.60351
0.29255
0.35616
0.11605
0.08162
0.10987
0.20002
0.12347
0.15211
0.09207
0.17643
0.10849
0.17587
0.04545
0.08813
0.31753
0.71036
0.02257
0.88015
0.25027
0.00062
0.00700
0.04286
0.01067
0.35438
Std diff in means and 95% CI
-1.00
-0.50
0.00
Favours Control
0.50
1.00
Favours Treatment
Police Family Violence Services, Fig. 1 Fixed and random effects models for reports of abuse to the police:
experimental and quasi-experimental designs
survey. Pate et al (1992) also found a decrease in
subsequent violence reported on a survey
following a second response in Dade County,
FL. Casey et al. (2007) reported fewer calls
made to the police among victims who received
a second response in New Haven, CT.
Davis et al. (2008) conducted a meta-analysis
of second responder programs. The analysis
concluded that the odds of reporting new abuse
to the police were about 1-1/4 times higher for
households assigned to a home visit treatment.
The meta-analysis found no difference between
treatments in reports of new abuse on research
surveys (see Fig. 1 below).
In sum, then, the weight of the evidence does
not indicate that second responder programs
reduce new instances of abuse and may, in fact,
increase subsequent calls to the police – possibly
because the intervention generates more actual
abuse or possibly because people who
receive a second response have more confidence
in the police.
Related Entries
▶ Domestic Violence
▶ Measuring Police Unit Performance
▶ Randomized Experiments in Criminology and
Criminal Justice
Recommended Reading and References
Anderson D, Chenery S, Pease K (1995) Biting back,
tackling repeat burglary and car crime. Crime
Detection & Prevention Series Paper 58. Home Office,
London
Bard M (1975) Role of law enforcement in the helping
system. In: Coffey AR, Renner JE (eds) Criminal justice as a system: readings. Prentice-Hall, Englewood
Cliffs, pp 56–66
Binder A, Meeker JW (1988) Experiments as reforms.
J Crim Justice 16:347–358
Breslin WJ (1978) Police intervention in domestic
confrontation. J Police Sci Adm 6:293–302
Bruno v. Codd, 396 N.Y.S. 2nd 974, NY Sup Ct (1977)
Buzawa ES, Buzawa CG (2003) Domestic violence: the
criminal justice response. Sage, Newbury Park
Casey RL, Berkman M, Stover CS, Gill K, Durso S,
Marans S (2007) Preliminary results of a police-advocate home-visit intervention project for Victims of
domestic violence. J Psychol Trauma 61:39–49
Davis RC, Medina J (2001) Results from an elder abuse
prevention experiment in New York City. National
Institute of Justice Research in Brief. U.S. Department
of Justice, Washington, DC
Davis RC, Smith B (1994) Teaching victims crime prevention skills: can individuals lower their risk of
crime? Crim Justice Rev 19:56–68
Davis RC, Taylor B (1997) Evaluating a proactive police
response to domestic violence: the results of a randomized experiment. Criminology 35:307–333
Davis RC, Maxwell C, Taylor B (2006) Preventing repeat
incidents of family violence: analysis of data from
three field experiments. J Exp Criminol 2:183–210
Davis RC, Weisburd DW, Taylor B (2008) Effects of
second responder programs on repeat incidents of
family abuse. Campbell Collaboration Systematic
Police in Nazi Germany
Review. http://db.c2admin.org/doc-pdf/Davis_2ndresp_review.pdf
Davis RC, Weisburd D, Hamilton EH (2010) Preventing
repeat incidents of family violence: a randomized field
test of a second responder program. J Exp Criminol
6:397–418
Dean CW, Lumb R, Proctor K, Klopovic J, Hyatt A,
Hamby R (2000) Social work and police partnership:
a summons to the village strategies and effective practices. A report of the Charlotte-Mecklenburg Police
Department and the University of North Carolina at
Charlotte. North Carolina Governor’s Crime Commission, Raleigh
Dunford FW, Huizinga D, Elliot DS (1990) The role of
arrest in domestic assault: the Omaha police
experiment. Criminology 28:183–206
Ford DA (1983) Wife battery and criminal justice: a study
of victim decision-making. Fam Relat 32:463–475
Greenspan R, Weisburd D, Lane E, Ready J, CrossenPowell S, Booth WC (2003) The Richmond/Police
Foundation domestic violence partnership. Police
Foundation, Washington, DC
Heisler CJ (2000) Elder abuse and the criminal justice
system: new awareness, new responses. Generations
24:52–58
Hirschel JD, Hutchison IW, Dean CW (1992) The failure
of arrest to deter spouse abuse. J Res Crime
Delinquency 29:7–33
Hovell MF, Seid AG, Liles S (2006) Evaluation of a police
and social services domestic violence program: empirical evidence needed to inform public health policies.
Violence Against Women 12:137–159
Hutchison IW, Hirschel JD, Pesackis CE (1996) Family
violence and police utilization. Violence Vict
9:299–313
Lempert R (1989) Humility is a virtue: on the
publicization of policy relevant research. Law Soc
Rev 23:145–161
Liebman DA, Schwartz JA (1973) Police programs in
domestic crisis intervention: a review. In: Snibbe JR,
Snibbe JM (eds) The urban policeman in transition.
Charles C. Thomas, Springfield, pp 421–472
Lloyd S, Farrell G, Pease K (1994) Preventing repeated
domestic violence: a demonstration project on
Merseyside, vol 49, Crime prevention unit paper.
Home Office, London
Manning P (1988) Symbolic interaction: signifying calls
and police response. MIT Press, Cambridge, MA
Mickish JE (2002) Domestic violence. In: Byers BD,
Hendricks JE (eds) Crisis intervention in criminal justice/social services, 3rd edn. Charles C. Thomas,
Springfield, pp 77–118
Miller N (1997) Domestic violence legislation affecting
police and prosecutor responsibilities in the United
States: inferences from a 50-state review of state
statutory codes. Institute for Law and Justice,
Alexandria
Parnas RI (1967) The police response to the domestic
disturbance. Wis Law Rev 2:914–930
3623
P
Pate A, Hamilton EE, Anan S (1992) Metro-Dade spouse
abuse replication project technical report. Police Foundation, Washington, DC
Plotkin MR (1988) A time for dignity: police and domestic
abuse of the elderly. American Association of Retired
Persons (AARP), Washington, DC
Sherman LA, Berk RA (1984) The specific deterrent
effects of arrest for domestic assault. Am Sociol
Rev 49:261–272
Stover CS, Poole G, Marans S (2009) The domestic violence home visit intervention: impact on police
reported incidents of repeat violence over twelve
months. Violence Victims 24(5):591–606
Stover CS, Berkman M, Desai R, Marans S (2010) The
efficacy of a police-advocacy intervention for victims
of domestic violence: 12-month follow-up data. Violence Against Women 16(4):410–425
Tatara R, Kuzmeskus MA, Duckhorn E (1997) Domestic
elder abuse information series. National Center on
Elder Abuse, Washington, DC
Taylor B (n.d.) Do home visits reduce repeat domestic
abuse calls to the police? Unpublished paper
Tracey Thurman et al. v. City of Torrington, Connecticut,
595 F. Supp. 1521 (D. Connecticut 1984)
Police Federations
▶ History of Police Unions
Police in Nazi Germany
1
Herbert Reinke and Thomas Roth
1
Bergische Universit€at Wuppertal,
Wuppertal, Germany
2
NS-Dokumentationszentrum der
Stadt Köln, Köln, Germany
P
2
Overview
Dictatorial regimes, not only in Europe but in
many other countries outside Europe as well,
have used and still use today the police for
terrorizing and for controlling political opponents
and for maintaining the power of the regime.
The Nazi movement went far beyond this when
getting hold of the police after their seizure of
power.
P
3624
While the Nazi regime particularly at its
beginning drew on traditional police forces and
started from authoritarian goals and strategies of
police work, it rapidly generated a radical new
kind of police organization, with a specific corps
of officers, committed to Nazi ideology, and
a radical concept of policing, focused on racial
categories and the goal of systematically cleansing society of all groups labeled as deviant.
During the Nazi years, the maintenance of
“law and order” and crime control and the persecution of political opponents became increasingly intertwined with policies of racial, often
total, exclusion. A police force evolved not only
supporting but promoting politics of genocide.
Police in Nazi Germany: Institutional
Contexts
Since the beginnings of modern policing in
Germany during the nineteenth century, and due
to federalism as a core characteristic of the
German political system, the police had always
been under the command of the federal states.
There has never been, except for the Nazi period,
a large-scale and centralized national police.
“Police” meant, and still means primarily,
the polices of the L€ander, federal states (i.e.,
before 1933 Prussia, Bavaria, Saxonia, etc.).
This institutional context remained unchanged
until the end of the Weimar Republic. Although
demands were put forward during the 1920s
for more centralization, in particular for fighting
crime more effectively, a centralized police
apparatus did not emerge before 1933.
In July 1932, the conservative German
Reichsregierung, the national government, had
the middle-left Prussian government on the
basis of an emergency decree dismissed, constituting a kind of prelude to the institutional
changes, which took place after the Nazi seizure
of power. In 1932, a presumed incapacity
for maintaining law and order was given as
a reason for the replacement of the Prussian
government (senior police officers included) by
commissioners, installed by the national
government.
Police in Nazi Germany
After the Nazi seizure of power at the end
of January 1933, a leading and prominent figure
of the Nazi party, Hermann Göring, was commissioned to direct the Prussian Ministry of the Interior. This position served as a basis for getting hold
of the Prussian uniformed police and the
Kriminalpolizei (Kripo, Criminal Police) as
power instruments for consolidating the Nazi
regime. Two emergency decrees, issued in
the weeks after the Nazi seizure of power,
suspended the basic constitutional rights of the
citizens, thus providing the Nazi regime with
the capacities for prosecuting ruthlessly its political opponents. These two emergency decrees
issued in February 1933 served until the end of
the Nazi regime as a pseudo-legal basis for the
unlimited police power of the regime.
Another leading and prominent figure of
the Nazi party, Heinrich Himmler, became head
of the Bavarian Political Police soon after the
Nazi seizure of power. Thereafter, until the first
months of 1934, Heinrich Himmler gained the
command of most of the political police departments of the L€ander and finally the command of
the Prussian Political Police as well, which had
been labeled since the end of November 1933 as
“Geheime Staatspolizei.” In November 1934, all
political polices, existing by that time in
Germany, were collected under the directorship
of Himmler. In a parallel process, the command
over the polices was transferred from the L€ander
to the national government. This development
toward a centralized national police came to
a central point, when Himmler was given in
June 1936 the title of “Reichsf€uhrer SS and
Chef der Deutschen Polizei,” indicating his dual
directorship as head of the SS and the German
police. In the aftermath of this entitlement, the
centralization and the “Vereichlichung” (nationalization) of the different police branches were
further enhanced; under Himmler, a national
police apparatus had been established, which
had not existed before.
But the centralization of the police was only
one aspect on the police agenda of the Nazi
regime. A further, very important development
was the merging of the police apparatus with the
SS, in order to create a novel kind of security
Police in Nazi Germany
force effectively and vigorously serving the
ideological ends of the Nazi regime. This meant
a growing detachment of the police from its traditional institutional context, while it was more
and more interleaved with the organizations of
the Nazi movement. Under the organizational
roof of the “Reichssicherheitshauptamt,” a sort
of holding for the different security forces of the
Nazi regime, which was established in 1939 after
the beginning of the war, this merging of the
police with the SS apparatus did find its
institutional imprint (Topographie des Terrors
2010; Wildt 2009).
Police in Nazi Germany II: From
Dictatorial State Police to Ideological
Security Force
After the Nazi movement got hold of the police in
1933, it took them only a few years to establish
a new approach of policing, clearly exceeding the
politics of persecution and suppression known
from other dictatorships and authoritarian
regimes being in power in Europe at that time.
Under the Nazis, the police’s functions were
no longer focused on “traditional” target groups
only, such as criminal perpetrators and political
opponents. The police became instead increasingly, and in a growing proactive manner,
focused on specific categories of people, who,
based on racial (Jews, Roma, and Sinti) and to
a significant extent on criminal-biological
criteria (social outsiders, habitual criminals),
were labeled as enemies of the racially defined
ethnic
community
of
Germans,
the
“Volksgemeinschaft” (Herbert 2011).
Policing was conceptualized as a policy of
social
cleansing,
meant
to
rid
the
“Volksgemeinschaft” of political and criminal
“enemies of the people” (“Volksfeinde”) and of
all “community aliens” (“Gemeinschaftsfremde”),
not fitting into or not willing or capable to adapt to
the racial and biological scheme of the Nazi
regime. This conceptualization was expressed in
the image of the police as a “doctor” cutting out all
symptoms of sickness out of the “social body.”
Racism, concepts of “racial hygiene,” and criminal
3625
P
biology, increasingly popular since the 1920s, now
penetrated nearly all layers of police work.
The identification and persecution of members of
“foreign races” and the stigmatization and exclusion of “hereditary” or habitual criminals and of
other categories of people labeled as deviant
became a core component of a police-driven, for
those concerned increasingly deadly social
engineering. The new police concept constituted
the basis for an active participation of the police
forces in the murderous and genocidal politics of
the Nazi regime.
Policing during the Nazi regime rested
upon a familiar yet specifically radicalized
division of labor: While the State Police
(Geheime Staatspolizei, Gestapo) organized the
repression of political opponents and took part
in the elimination of “racial aliens,” such as
Jews, Poles, or Russians, the Criminal Police
(Kriminalpolizei, Kripo) became responsible for
the prosecution and the elimination of criminals,
social outcasts, and the group of Roma and Sinti.
The uniformed police (since 1936 under the label
Ordnungspolizei, Orpo), responsible for the
overall maintenance of law and order in Nazi
society, was involved in the surveillance of
everyday life and played an important role in
ensuring the populace’s conformity and identifying deviants. It thus provided essential assistance
to the Security Police making a significant
contribution to the Nazi regime’s political persecution and social exclusion.
Phases of Development
The historical research on the police during
the Nazi regime has identified five phases for
summarizing the growing involvement of the
police in the Nazi regime. During the first
phase, primarily during the months after the
Nazi seizure of power, the efforts of the Nazi
regime were much focused on getting hold of
the police as a power instrument, indispensible
for suppressing political opponents and for
maintaining the Nazi regime during its first
months of being in power. The second phase,
consisting more or less of the years 1934 to
P
P
3626
1936, could be described as the phase when
foundations for a merging of the police and the
SS apparatus of the Nazi movement were laid.
This merging was started by putting Heinrich
Himmler, the head of the SS apparatus, and
other SS leaders at the head of the German police;
by linking the Kriminalpolizei and the Gestapo
under the label of a national socialist Security
Police (Sicherheitspolizei); and by placing
the Security Police and the Security Service of
the SS, the Sicherheitsdienst (SD), under a joint
institutional roof. The third phase, that is, the
years from 1936 to 1939, was very much marked
by specific forms of radicalizations of police
practices. During these years, the Kripo and
the Gestapo began to implement the racialbiological ideologies of the Nazi regime into
police work and started to select systematically
those groups labeled as enemies of the “community” for special police measures, for new forms
of confinement, and eventually for deportation
into the expanding concentration camp system.
With the beginning of the war in September 1939
and the establishment of the “Reichssicherheitshauptamt” in that year until late 1944, the police
in its fourth phase of its involvement in the Nazi
regime turned into the most important instrument
for the completion of the Nazi racial ideologies:
By its participation in the killing units of the
“Einsatzgruppen,” the police engaged as main
actor in the mass murder of the European Jews
in Europe. So did the units of the uniformed
police, especially after the war against the Soviet
Union had begun (1941). And back home within
Germany, it was the police which summoned
the German-Jewish population of the cities and
the countryside for the deportation into the
concentration camps and which accompanied as
guards the trains and transported the victims to
the death camps. By the same time, the Kripo
again radicalized its practices by sending in
more and more people into concentration
camps, thus occupying next to the Gestapo
a central position in the overall strategy for the
cleansing of the “community” from its enemies.
When carrying out this, the Gestapo and the
Kripo were not only instruments used by the
Nazi regime but engaged in it beyond obeying
Police in Nazi Germany
orders by putting forward own initiatives for
improving the efficiency of the cleansing of the
“Volksgemeinschaft,” that is, of mass murder.
During the fifth and final phase, which lasted
from 1944 to the end of the war and the end of
the Nazi regime in 1945, the radicalization of
police practices intensified once again. During
this final phase, the Security Police focused its
control efforts on the foreign workers, who lived
by that time in enormous number in Germany and
had mostly been forced to work for the “Third
Reich.” Moreover, indicators on increasing discontent and disloyalty among the Germans within
the “Reich” turned into a major concern for the
Security Police. Starting in the last months of the
war and the regime and continuing in its final
days, the Security Police tried to cope with
the decreasing power of the regime and the
increasing disloyalty of the population with
a dramatically growing brutalization, resulting
in the mass murders of foreign workers and
political opponents and selected killings of
ordinary Germans no longer loyal to the regime
during the final days of the war.
The Security Police and Its Personnel
In order to turn the Security Police into the police
of the “Volksgemeinschaft,” it became organizationally realigned, centralized, and provided with
new hierarchical structures, guidelines, and
standards. The transformation and renewal of
the staff played a key role as well. While the
personnel of the Kripo with about 13,000
employees was mainly transformed with the
normal turnover of staff, the Gestapo was
established fundamentally new. On the basis of
the political police of the Weimar Republic, with
its approximately 1,000 officers, the Gestapo
grew to an apparatus with over 30,000 employees
until the end of World War II.
Senior positions within the Gestapo and
the Criminal Police were held since 1933 by
men who openly showed their loyalty to the
Nazi regime or had even been active in the Nazi
movement for a long time already. The few police
officers, who had declared themselves against the
Police in Nazi Germany
Nazis and had represented a decidedly republican, democratic police during the Weimar years,
were dismissed or demoted. New police recruits
had to indicate, at least through membership in
the Nazi party, that they were supporting the
regime permanently. The police training combined political-ideological indoctrination and
professional police instruction. Finally, with the
inclusion of police officers in the SS, and with
joint spheres of action and social contacts
between the Security Police and the SS Security
Service (SD), close relations between police and
the Nazi apparatus emerged (Banach 1998;
Browder 1996).
In this way, the Sipo established the model of
a new type of police officer, combining expert
knowledge and professional ambition with ideological radicalism and the commitment of
a “political soldier.” This new type of police
officer became a dominating pattern inside the
head office of the Gestapo and the Criminal
Police in Berlin, as well as among the senior
officers of the regional headquarters of the Sipo
branches. This leadership group consisted mainly
of younger cadres with middle-class background,
well-educated, academically trained, often active
in extreme right-wing organizations before 1933
and striving for implementing the societal models
of the Nazi regime (Wildt 2009). Another form of
“Nazification” took place among the lower ranks
of the Security Police, particularly within
the Gestapo, which, due to its growing staff
requirements, recruited members of Nazi organizations with low-level formal education only and
without specific police skills. These activists
played an important part in the implementation
of the Nazi movement’s political and racial stereotypes and its violence into police practice.
Despite these fundamental changes, patterns of
regular, traditional police work remained present
in the Sipo (Dams and Stolle 2012; Roth 2010).
Established procedures and routines of criminal
investigation and administration were not
abolished but pursued for the purposes of Nazi
regime. In addition, older senior detectives, who
had entered the police service during the German
Empire or after World War I, played a significant
role even after 1933. Although the Nazi leadership
3627
P
had promised to purge the police of politically
“unreliable” officers after the seizure of power,
the rate of dismissals in 1933/1934 was relatively
low (<5 %), partially due to the lack of candidates
for replacing the officers to be dismissed. Thus,
detectives, who were professionally trained and
socialized in the Weimar Republic, had
a significant impact on the apparatus of the Sipo,
in particular during the early years of the Nazi
regime. In the middle and lower ranks of the
Sipo, the local posts, and particularly inside the
Kripo, they remained an influential and supporting
element until the end of the regime. With their
experience and technical knowledge, they contributed significantly to the functioning and effectiveness of the Security Police.
Reasons for Support and Initiative
The support that the Nazi regime gained
from various groups of police officers was due to
political indoctrination, opportunism, and careerism; instead, shared values and professional
expectations were key factors in that respect as
well. Nationalism, antiliberalism and anticommunism, radically stereotyped views on political outsiders or marginalized groups,
authoritarian conceptions of state intervention,
and social order (“law and order”) had been part
of the mainstream police philosophy during the
1920s. The rejection of parliamentary democracy
and of the rule of law, demands for extensive and
harsh control of political opponents, social
outcasts and criminals, and the expansion, unification, and centralization of the police forces came
upon a broad consensus within the police apparatus. The Nazis reacted positively on the demands
and expectations circulating inside the police by
suspending or abolishing existing legal restrictions
and by giving the police the means for realizing
their ideas of radical social control and social
cleansing. In the “fight against criminality,”
against “enemies of the Volksgemeinschaft” or
against “antisocial elements,” the police and Nazi
movement, senior detectives, and the new leadership of the Sipo shared the goals to pursue (Paul
and Mallmann 1996; Wagner 1996).
P
P
3628
As far as the cultural integration of the police
in the Nazi state was concerned, celebrations or
public events played a role, too. Of particular
importance was that the Sipo was given a new
image, which was perceived positively by most
police officers. While public criticism of the
police nearly disappeared after the suppression
of the free press, the Nazi regime and senior Sipo
officers developed an image of the police, which
depicted the detectives as modern, scientifically
trained, effective investigators and provided the
Gestapo with the myth of an omnipresent and
omnipotent power apparatus.
For the participation of Kripo and Gestapo
officers in the terror of the Nazi regime, other
factors and processes were important as well:
increasing routine and familiarity with practices
of exclusion and violence, the dehumanization of
the victims through propaganda and bureaucratic
procedures, and division of labor and responsibilities and social distance to those persecuted,
but also the emergence of a specific Nazi ‘cop
culture’ that was based on harshness and soldierly
masculinity, camaraderie, and group pressure.
The Security Police: Methods and
Instruments of Power
The alignment of the Sipo with the overall policies of the Nazi regime and the emergence of
a concept of “racial policing” was accompanied
by an expansion of instruments of power and
a radicalization of police methods. Limitations
of police work by legal bonds and external controls were abolished, and fundamental civil and
human rights were suspended. Those affected
were exposed to nearly unrestricted police
arbitrariness.
While the Gestapo and Kripo detectives continued to work as a part of the law enforcement
process and as an investigative administration for
the public prosecutor, an independent police
penal justice steadily expanded since 1933.
Its base was the introduction of an independent police custody, the so-called protective
custody (Schutzhaft) used by the Gestapo and
the “preventive custody” in the case of the
Police in Nazi Germany
Kripo. Both instruments allowed for detention
without formal criminal justice proceedings and
judicial review and resulted in permanent confinements of regime opponents and offenders in
prisons or concentration camps. Those affected
by these forms of confinement could be exposed
to permanent drill, exhausting labor, hunger and
disease, corporal punishments, and mortal
danger. “Protective” and “preventive custody”
enabled the police to punish, even on the ground
of mere suspicion and poor evidence. Both forms
of custody were used to “correct” court decisions
and became instruments used massively for the
intimidation and elimination of political opponents and people labeled as deviant. By applying
this police custody, Gestapo and Kripo developed
a policy of social exclusion that, until 1945,
took several hundred thousand people into the
concentration camps.The radicalization of police
methods went even further within the Gestapo.
While the use of physical violence against
suspects by the Criminal Police was rather
the exception, the Gestapo had the right to use
violence for extorting statements. Against
the backdrop of World War II, the Gestapo
implemented its own prison camps, the so-called
corrective labor camps (Arbeitserziehungslager),
where prisoners were usually held for several
weeks. In the 1940s, groups such as Jews, Poles,
or Russians were completely removed from the
realm of the judiciary and brought completely
under the control of the Gestapo. Toward
the end of the war, local and regional Gestapo
offices were finally empowered to execute
prisoners.
Arbitrary arrests and deportations, penetrations
of the private sphere, extensive surveillance measures, and the use of informers (V-M€anner) were
the main features of the work of the Security
Police, but routine police methods of investigation,
detection, and forensic collection were continued
as well, being the basis for the deployment of
police terror. Despite the image of being omnipresent and omnipotent, the police force was dependent
on the support of other institutions of the Nazi state.
When heading for a comprehensive control of the
“Volksgemeinschaft,” the Sipo’s own resources
and personnel were insufficient (Gellately 1990;
Police in Nazi Germany
Paul and Mallmann 1996). But the Security Police
could count on the cooperation of other instances
of social control: Together with subunits of the
Nazi party, the uniformed police, government and
municipal administrations, health and employment
services, the tax administration, postal and railroad
services, as well as the customs authorities and
large parts of the economic sector, a cooperative
network for enhancing and enforcing the
policies and the terror of the Nazi regime was
created. Significant support came also from the
public. For controlling and suppressing core political opposition groups such as communists, or
for controlling so-called criminals, the Gestapo
and the Kripo could fall back on their own investigative work, but for the discovery of small
individual deviances and minor violations, for the
observation of everyday life, for the penetration
into the privacy of suspected people, and for the
pervasion of so-called closed milieus, the Sipo had
to rely on denouncers and informers. For the
discovery of a spontaneous protest, for the exclusion of the Jews from the “German” society, and
for the close observation of minorities, input from
the population and party organizations was essential. The power of Gestapo and Kripo derived not
only from the results of its own investigations but
from a network of individual and institutional
supporters.
Key Aspects of Gestapo Work
Central to the work of the Sipo were permanent
dynamics of radicalization (Dams and
Stolle 2012; Mallmann and Paul 2000). When
heading for an overall cleansing of the
“Volksgemeinschaft” by eliminating political
opponents’ “racial enemies” and all sorts of deviance within the population, new enemy groups
and accordingly new security requirements were
continuously constructed. Initially the most
important activity of the Secret State Police was
the suppression of the left-wing workers’ movement, in which the regime saw the main potential
for opposition and unrest. The Gestapo’s
attack was directed primarily against communists, moreover against socialists, social
3629
P
democrats, and representatives of the Free Trade
Unions. By 1933, tens of thousands of opponents
have been arrested, mistreated and tortured,
handled by criminal courts, or transferred to –
by that time – early forms of concentration
camps. Other measures were repeated raids,
searches, and the confiscation of literature, printing machines, and illegal publications; the ban of
organizations and newspapers; and the closure of
the workers’ movement’s associations. Although
the left opposition was constantly trying to
reorganize illegal groups, the Gestapo managed
to dismantle their resistance until 1936/1937.
Additionally, the Gestapo also developed
a supporting role in the control of the Christian
churches and religious groups. The aim was to
stem criticism from the churches, to promote the
adaptation of the believers to the Nazi ideology,
to force the churches out of the public, and to
reduce their impact on the population. Here, the
repressive approach of the Gestapo was clearly
more differentiated than toward communist
resistance and was aimed mainly at integration
into the “Volksgemeinschaft.” While in Protestant-dominated areas, the Gestapo only monitored a critical minority within the church; in
Catholic areas, where the Gestapo suspected
a widespread reluctance to Nazism, the State
Police developed a wide range of measures: It
reached from the observation of meetings and
church services and larger propaganda
campaigns against the church to the ban on Christian associations and the punishment of individual representatives of political Catholicism and
dissident clerics.
Another important field of Gestapo work was
the control of public opinion. In addition to the
suppression of critical statements from the labor
movement and the churches, the communication
of “ordinary” citizens should be regulated. Based
on new decrees and laws putting the criticism
of the representatives and actions of the government and the Nazi party under punishment as
“treachery” (Heimt€ucke), the Gestapo initiated
tens of thousands of investigations. Since the
beginning of the war, expressing doubts about
the Nazi propaganda or pessimistic statements
about the war were persecuted more intensely
P
P
3630
and partly criminalized as “undermining of
the military strength” (Wehrkraftzersetzung).
The same was true of the media consumption of
the population, as listening to foreign radio
stations and distributing demoralizing information could be persecuted since 1939 under the
label “radio crime” (Rundfunkverbrechen).
Since the mid-1930s, the State Police exerted
increased pressure on nonconformist juveniles
who rejected national socialist cultural hegemony
by insisting on their particular subcultures and by
turning against the Nazi youth organization. In
addition, the Gestapo initiated wide-ranging
prosecution actions against homosexuals, whose
behavior was considered a threat to Nazi population policy. It punished violations of economic
policy and labor discipline as well and was
responsible for criminal proceedings against
deviant members of the Nazi movement. On all
these fields, the Gestapo sometimes acted in
competition but usually in close consultation
with the Kripo.
In several of these fields of action and against
occasional violations of the norms of the community the Security Police, especially during
the prewar years, acted selectively (Gellately
2001; Roth 2010). So it not only applied hard,
negative sanctions but used warnings and
lighter penalties to intimidate and discipline
violators, in order to make them get in line with
the “Volksgemeinschaft.”
As far as the German-Jewish population was
concerned, the police aimed at a radical marginalization and exclusion. While the political,
social, and economic exclusion of Jews had
been enhanced during the first years after the
Nazi seizure of power very much by Nazi party
subunits, the Gestapo thereafter took more and
more the initiative. Since legislations in 1935,
which turned German Jews into second-class
citizens, the Gestapo went into organizing the
discrimination, isolation, and exclusion of this
population group. The police conducted the
registration of German Jews, restricted their freedom of movement, were continuously involved
in the criminalization of relations between Jews
and non-Jews, and also contributed to the
economic expropriation of Jewish citizens.
Police in Nazi Germany
As of November 1938, the Gestapo began with
the systematic exclusion from German society.
During and after the November 1938 pogroms,
the Gestapo throughout Germany arrested more
than 30,000 German-Jewish men and deported
them into concentration camps, where the victims
were released only when they had agreed upon
abandoning their properties and upon leaving
Germany. Just a few weeks earlier, in
a nationwide campaign, the Gestapo had arrested
about 17,000 Polish Jews and deported them over
the German-Polish border – a test run for the
mass deportations of the 1940s.
From the spring of 1941, the Gestapo organized the final spatial segregation of the
remaining Jews within Germany. In the cities
Jews were evicted from “Aryan” houses and
sent to special “ghetto houses” or local camps.
People of Jewish origin, who were kept in
asylums, prisons, and other institutions, were
separated from other inmates and handed over
to the Gestapo. From autumn 1941, the State
Police organized the deportations from the
“Reich” to the ghettos and death camps in occupied Eastern Europe. By 1943 some 150,000
people were deported from Germany, while
only a few were able to survive as “Mischlinge”
(crossbreeds), partners in so-called mixed
marriages or in hiding.
The Work of the Kripo
Unlike the Gestapo, the Criminal Police was
initially focused more on traditional police activities such as criminal investigation and crime
detection. During the 1930s, however, its practice
was more and more determined by the methods
and goals of a police of the “Volksgemeinschaft”
(Browder 1996; Wagner 1996). In late 1933,
the senior ranks of the Kripo opened a new field
of activity under the label of “preventive fighting
against crime” (vorbeugende Verbrechensbek€ampfung). Under the pretext of reducing
crime, the Kripo rendered recidivists into
“preventive detention” by deporting them for an
indefinite time into concentration camps, even in
the absence of a specific offense or criminal court
Police in Nazi Germany
proceedings. Since 1934, the policemen were
also enabled to subject known criminals to
a “systematic observation” (“planm€aßige
€
Uberwachung”)
and rigid rules of conduct.
At the beginning, the policy of preventive
fighting against crime targeted at a relative
small group of a few thousand criminals, which
should be kept from further criminal activity by
deterrence and confinement. With the vision of
a “Volksgemeinschaft” without crime in mind,
the Kripo constantly enlarged its net. According
to a nationwide campaign of arrests against 2,000
so-called professional and habitual criminals
(“Berufs-” and “Gewohnheitsverbrecher”) in
March 1937, the local Kripo offices started systematically detecting recidivists and sending
them in very large numbers to concentration
camps. The criminal deportations soon were
directed not only against professional burglars
and violent or sexual offenders but met more
often petty criminals. At the same time, the persecution of homosexuals now shifted from the
Gestapo to the Kripo. Even when acting against
other social outcasts, the Kripo officers acted
with increasing harshness. By doing so, the Criminal Police took up and radicalized the strategies
and practices, pursued already by labor, social,
and health administrations. As early as 1933, the
Kripo had initiated numerous raids against
prostitutes, beggars, and vagrants, demonstrating
that social “disorder” and public “indecency”
should no longer be openly tolerated. The next
step came when the practice of “crime prevention” in December 1937 was extended to marginalized groups, and the Kripo was empowered to
apply indefinite confinement on all individuals
“endangering the general public by asocial
behavior.” Thereupon the concentration camp
became a common measure against outsiders,
while the Kripo turned into the dominant actor
of social exclusion (Ayaß 1995; Gellately and
Stoltzfus 2001).
In 1938, the police leadership prompted two
nationwide sweeps against so-called work shy, in
the course of which the Criminal Police arrested
more than 9,000 people (the Gestapo about
1,500). Since then the local Kripo posts consistently put more and more “asocials” into
3631
P
“preventive detention” – next to prostitutes,
homeless and vagrants welfare recipients, alcoholics, men neglecting their family duties,
women labeled as promiscuous and women with
venereal diseases, dealers, peddlers, etc. Even
biographically, “crime prevention” radicalized:
Since the early 1940s, 1,000 “criminal and antisocial youths” were interned, terrorized, and biologically examined in special camps. Especially
involved was the female unit of the Kripo
(Weibliche Kriminalpolizei), which not only
served as kind of welfare police for (female)
adolescents but was responsible for the elimination of “criminal offspring.” Until the end of the
Nazi era, the Kripo under the label of “crime
prevention” delivered about 80,000 offenders to
the terror of the concentration camps (Wagner
1996).
In addition, there was the fight against “alien
races.” The Criminal Police assisted the Gestapo
in its anti-Jewish policy, for example, by sending
Jews with criminal records to camps or persecuting relationships between “Aryans” and “nonAryans” as “race defilement” (“Rassenschande”).
To core projects of the Kripo belonged the
exclusion of the Roma and Sinti, who, as
“gypsies,” had been particularly discriminated
against and placed under special surveillance
long before 1933 (Zimmermann 1996). Since
the mid-1930s, they were deprived of civil rights,
prohibited
further
mobility,
separately
registrated, examined by racial experts, and
marked as “racial aliens.” After May 1940 the
Kripo started several deportations, taking thousands of German and Austrian Roma and Sinti to
the annexed and occupied Poland, where most of
them lost their lives. From March 1943, about
23,000 “gypsies” from the “Reich” and occupied
Europe were transported to the extermination
camp at Auschwitz. As the practice of Gestapo,
the work of the Kripo since the 1940s was closely
associated with national socialist mass murder.
Gestapo and Kripo During World War II
The beginning of World War II provided
additional rationalizations for the radicalization
P
P
3632
of policing and increased the willingness of
policemen to act harshly and murderously against
political, social, and racial “enemies.” With the
expansion of the Nazi regime into large parts of
Europe, new spaces of social control and
eliminatory rule were created, while millions of
people who were defined as dangerous by the
Nazi regime fell in the hands of the Gestapo and
the Criminal Police (Mallmann and Paul 2000;
Topographie des Terrors 2010).
The further radicalization of the Sipo occurred
in the occupied territories, where the Sipo
exercised unprecedented and unbound physical
violence. Since September 1939, more and more
Gestapo and Kripo detectives had to leave their
offices within Germany and put themselves at the
service of Nazi occupation and warfare. Some of
them participated in the “Einsatzgruppen,” who
entered the occupied territories behind the
Wehrmacht to “cleanse” the conquered areas of
indigenous elites, oppositional groups, Jews, and
minorities like gypsies and the disabled. The
“Einsatzgruppen,” already established as mobile
killing units in the Polish campaign 1939 and
considerably extended since the invasion of the
Soviet Union in 1941, have been responsible for
the direct killing of hundreds of thousands of
people especially in Eastern Europe (Klein
1997). In these units, and the offices established
in the occupied territory, the institutional
merging of police and SS was further promoted.
Members of the Security Police were involved
at all levels in Nazi crimes: in the invention of
killing techniques for the gas vans and death
camps and the “administrative execution” of
genocide; in the surveillance of ghettos and economic plundering of the occupied countries; in
reprisals against political opponents, mass executions of Jews, political officials, prisoners of war,
or “partisans”; and in the killing of “criminals,”
“asocials,” or “gypsies.” In the violent environment of the East, it could be practiced without
restraint and directly, what the police in the
“Reich” before the war could achieve only
indirectly through deportation and imprisonment
into concentration camps: the annihilation of
those categories of people defined as enemies of
the “Volksgemeinschaft.” When doing this, the
Police in Nazi Germany
Security Police was not only supported by the
Germany army or the civil administrations of
the occupied countries. An important element of
German police power in Europe was the collaboration of local police forces, radical nationalist
organizations, and auxiliaries, which contributed
significantly to the execution of mass murder.
The Security Police on the “Home Front”
In the course of World War II also inside
Germany, the Sipo radicalized its scenarios of
crucial enemies and possible dangers (Mallmann
and Paul 2000). Political opposition and social
deviance were now seen as a threat to the Nazi
war effort and the stability of the “home front.”
But while this resulted in a further tightening of
sanctions, gaps in the system of police control
were widening, as many officers were withdrawn
from the “internal front,” the police functions were
continually extended, and the prolonged war and
ongoing air raids, especially in big cities, lead to
a decline of law-abidance and loyalty to the
regime. The impending loss of control in the war
did not lead the Security Police to a strategy of deescalation but was answered with a further intensification of terror. To this added, that police officers who had returned from the occupied
territories brought their experiences of violence
and concepts of “enemy combat” with them to
the “internal front.” Due to the expansion of control requirements and the social distractions of war
society, the Kripo focused more and more on the
disciplining of those refusing to work for the war
community, on surveilling the conduct of young
people and single women, on halting the spread of
sexually transmitted diseases among prostitutes
and soldiers, or on the persecution of the black
market, “looting” and other war-related crimes
(Roth 2010; Wagner 1996). These control efforts
also met conservative and bourgeois opponents of
the regime – as the mass arrests after the attempt to
assassinate Hitler in 1944 illustrate.
However, on the “home front” within
Germany, racial politics stood at the center of
the State Police’s activity. In addition to the
exclusion of the Jewish population, the control
Police in Nazi Germany
of “foreign workers” became a main task of the
Gestapo – especially since the intensification of
the use of foreign workers in the German
Economy (Ausl€andereinsatz) and the massive
recruitment of forced laborers from the occupied
Soviet Union since 1942. The Gestapo not only
had to monitor the work discipline of foreign
workers and to ensure the functioning of the war
economy, it also was to suppress resistance and
sabotage and punish forbidden relations between
Germans and foreigners. That was mainly
applied to workers from Poland and Soviet
Russia: They were classified as “racial aliens”
and placed at the bottom of the racial hierarchy
of the Nazi system, were regarded as a special
threat to the order of the “internal front,” and
were almost entirely subject to the power and
violence of the State Police. Since the middle of
the war, the measures of the Gestapo even
included the killing of East European laborers.
Numerous Polish, Ukrainian, or Russian civilian
workers, who have had intimate relations with
locals, were publicly executed due to transgression of “racial barriers” (Gellately 2001). This
indicates what State Police practice characterized
in the last months of the war: A policy of open
terror, which brought mass murder to the “home
front.” In the final phase of the war, the Gestapo
inside the “Reich” increasingly followed the
models and methods of occupation. Hierarchies,
bureaucratic procedures, and the formal allocation of responsibilities became less important, as
much of the work of the State Police was taken by
Sonderkommandos (special task forces) operating with high mobility and flexibility.
They worked increasingly independent to the
central offices in Berlin, with almost unlimited
instruments of terror. In the last months of the
war, executions, initially dependent upon authorization from Berlin, could be imposed by
regional commanders of the Security Police and
eventually the local offices themselves. Before
the collapse of the Nazi system, the work of the
Gestapo in the “Reich” was marked by mass
arrests, targeted killings, and “combat missions.”
They were directed primarily against “foreign
workers,” young people, deserters, and criminals
who were in hiding in the ruins of the big cities.
3633
P
Since the Gestapo with its flexible organizational
structures and unlimited powers assumed nearly
all of the Security Police’s competencies, the
officers of the Criminal Police lost importance
and influence in the last months of the Nazi
regime. However, at least some of the Kripo
officers participated in arbitrary killings and
supported the “final battle” of the Gestapo.
The Uniformed Police: Everyday
Practices and Participation in the
Holocaust
During the Weimar Republic Period, the
uniformed police tried to change the image it
had in the public. While an image, deriving
from the police before World War I, lingered
on, which portrayed the police very much as
authoritarian, militaristic, and emphasizing the
social distance between the uniformed police
and the general public, senior police officers
tried to change this image for a portrait, which
displayed a police much closer to the public,
a police, which was no longer primarily
a protector of the state but a protector of the
“people” as well. After the Nazi seizure of
power, the leadership of the police attempted to
give this portrait even more contours by popularizing a label, which showed “The police, your
friend and helper.” The uniformed police
organized during the 1930s and even during the
first years of the war a number of “Tage der
Polizei” (days of the police), which, as a public
relations action, were to propagandize the label.
Other activities, inaugurated by the police during
the 1930s, focused on improving the traffic
circulation in the cities and on preventing the
growing risk of traffic accidents, by training the
general public, in particular children, how to
behave adequately on the roads of the city.
All this was intended to make the slogan “The
police, your friend and helper” popular, but this
label covered only partially the reality of
the uniformed police during the Nazi regime.
The uniformed police was “friend and helper”
only to those, who were members of the
“Volksgemeinschaft.” Jews, for example, were
P
P
3634
not members of this community, which comprised only Germans along very narrowly defined
social, racial, and biological criteria. Some of
those, who were opponents of the new regime, or
were not members of the “Volksgemeinschaft,”
experienced immediately, others very soon a new
repressive side of the uniformed police and a new
specific, selective approach to law and order,
influenced by the political strategies of the Nazi
regime. The uniformed police did not intervene,
when political opponents of the Nazi regime were
arrested, maltreated, or even killed but conceded
instead to a partition of labor with the police doing
the work, which looked professional and members
of the paramilitary groups of the Nazi movement,
often in the position of auxiliary police personnel,
doing the “dirty” work, such as torture and killings
of political opponents. And in 1933, soon after the
Nazi movement had come into power, the police
did not intervene, when the Nazi movement staged
a boycott against German-Jewish shopowners and
merchants, thus indicating law and order might not
pertain to German Jews any longer. These attitudes became increasingly enhanced, when
a growing loading of police training with patterns
of Nazi ideology and anti-Semitic contents
occurred (Matth€aus 2003; Westermann 2005).
Institutionally and organizationally, the
uniformed police experienced a drawback, when
in 1936, after a organizational reconstruction,
which had started very early after the Nazi
seizure of power already, the units of the
uniformed police were transferred to the German
army to increase the army’s numerical strength.
This weakened the organizational resources of the
uniformed police and reduced the number of its
personnel to about fifty percent of its original
strength. Until the end of the 1930s, the uniformed
police gradually managed to increase again the
number of its officers and to reestablish police
units. The personnel of these newly established
units, later named as police battalions (Polizeibataillone), was comprised to large extents of police
reserve recruits without any professional police
background instead of professional police officers.
These police units were used alongside the military, when the Nazi regime annexed Austria
(1938) and Czechoslovakia (1938/1939). These
Police in Nazi Germany
participations of units of the uniformed police
still ran under the heading of securing the occupied
territories, but with the start of the war against
Poland in 1939, the uniformed police turned into
an instrument executing the racial and genocidal
ideologies of the Nazi regime. The uniformed
police finally became a murderous key institution
in the Holocaust. Already during the starting phase
of the war against Poland, the units of the
uniformed police, following the advancing German army, committed atrocities among the Polish
population, Jewish and non-Jewish, thus becoming part of the racial strategies the Nazi regime
started pursuing already during the first phase of
the war. Especially the violent transfer of
populations, the settlement of “ethnic Germans”
in the occupied parts of Western Poland and the
brutal eviction of Polish and Polish-Jewish
inhabitants of these parts of the country into the
occupied central Poland (the “Generalgouvernement”) became one of the first elements of
the participation of the uniformed police in the
geostrategical and racial policies of the Nazi
regime.
With the beginning of the war against
the Soviet Union in June 1941, the involvement
of the units of the uniformed police gained
unprecedented genocidal dimensions. The socalled Einsatzgruppen, operating in the hinterland of the advancing German army within the
Soviet Union, killed more than 500,000 victims
until the end of 1941, most of them Jews. These
mobile killings units not only were composed of
members of the Security Police and the Security
Service of the SS (SD) but could rely to large
parts on military SS-formations (Waffen-SS) and
members of the uniformed police. These mobile
killing units advanced within the Soviet Union as
far as the German army, that is, into the territory
of the former Baltic states, and from there close to
Leningrad, they covered Belorussia and moved
forward into the vicinity of Moscow, covered
the whole of Ukraine and proceeded as far as
the northern fringe of the Caucasus. Most of the
direct killing actions and massacres on the
Eastern Front, as the killing of about 33,000
Jews from Kiew in Babi Jar, were carried out by
the “Einsatzgruppen” and supporting units.
Police in Nazi Germany
While the overall number of members of
the “Einsatzgruppen” remained relatively
small (there were up to 3,000 men in the
“Einsatzgruppen” following the advancing
German army in the war against the Soviet
Union), the overall figures for the members of
the units of the uniformed police participating in
war, occupation and genocide were much
higher: In occupied Poland and in the occupied
parts of the Soviet Union, approximately 50,000
men served in different units of the uniformed
police, the number of these units amounting to
90. This high figure of men serving in mobile
units of the uniformed police was due to the fact
that the Order Police was not limited to its participation in the murderous actions of the
“Einsatzgruppen.” The units of the uniformed
police carried out a great number of executions
of their own, quite often when the Ghettos in the
cities of Eastern Europe, where the Jewish population was forced to live in, were dissolved and
its inhabitants being sent into the death camps.
But the radius of action of the uniformed police
was not limited to Eastern Europe. In most of the
countries occupied by Nazi Germany, units of
the uniformed police held a significant position
in the occupation apparatus, often charged with
a more general maintenance of the security in
the occupied countries. In the later phases of the
war, units of the uniformed police became
heavily involved in fighting guerrilla activities
against the German occupation (“Partisanen-” or
“Bandenbek€ampfung,” terms often used for
masking executions of Jewish men, women,
and children; the elimination of resistance
groups; and the brutal suppression of the local
population). During the final phases of the war,
units of the uniformed police took part in the
indiscriminate mass killings of hostages all over
Europe, leading to the extinction of whole village populations (Oradour-sur-Glane).
The overall death toll, for which the
uniformed police of the Nazi regime could be
held responsible, is difficult to calculate
precisely. But calculations in different publications, older ones (Gutman 1990) and more recent
ones (Curilla 2011), count approximately
1,200,000 victims, who were either murdered by
3635
P
the units of the uniformed police or by members
of this police as participating in mass executions
as members of other units (“Einsatzgruppen”).
After 1945: Research and Institutional
Reflections
Elaborated research about the role and the functions of the police as part of the Nazi regime did
set in late in Germany. The few studies available
until the 1980s either gave limited descriptions of
the police’s institutional settings and the development of this settings during the years from
the Nazi seizure of power until the end of the
war, or they focused on the Gestapo, emphasizing
a picture of the Gestapo as omnipresent and
omnipotent, terrorizing political opponents and
the so-called racial enemies of the German people and, since the beginning of the war, covering
the whole of Nazi-occupied Europe with its
murderous activities. This early research concentrated very much on a top-down perspective, by
looking primarily at the main goals, leadership,
and institutional context of the Gestapo’s activities without analyzing the whole personnel
involved and without detailing the Gestapo’s
impact and societal background. This changed
for a bottom-up perspective, when the social history turn within historiography touched upon the
writing about the Gestapo as well. This meant
primarily to find out in detail what the Gestapo
had meant for its victims. But next to the attempts
of giving answers to this question, another feature
turned up during the 1970s and 1980s: A new
social history of the Gestapo revealed how important input from other institutions and the population had been for the work of the police during the
Nazi regime, in particular input in the form of
denunciations and input deriving from all sorts of
informers. This led some historians to postulate,
that Germany under the Nazis had been a selfpolicing society (Gellately 1990; Paul and
Mallmann 1996). More recent research after the
year 2000 however has refocused the attention on
the terror exercised by the Gestapo and emphasized the initiative and specific politicalideological radicalization of the Gestapo corps.
P
P
3636
At the same time earlier, descriptions of the
Gestapo focusing mainly on Himmler and
Heydrich or paradigmatic figures like Eichmann
were replaced by more complex and differentiated analyses of the personnel, emphasizing factors like generational background, collective
values, organizational culture, and institutional
dynamics (Mallmann and Paul 2004; Paul 2002;
Wildt 2009).
If the research about the Gestapo did set in
relatively late already, critical studies of the other
polices, that is, about the criminal and the
uniformed police, did set in even later. The longlasting silence surrounding the Kripo’s crimes
during the Nazi period had much to do with the
fact that those victims affected by national socialist “crime fighting” were continuously marginalized after 1945, mostly remained objects of
repressive policing and were not seen as “legitimate” victims of the “Third Reich.” This changed
during the 1970s as critical social scientists and
“grassroots historians” developed a more comprehensive approach toward the exclusionary politics
of the Nazi regime and claimed recognition for the
formerly “forgotten victims” (Ayaß 1995; Wagner
2002). The research on the Kripo, which was
clearly expanding since the 1990s, has given its
attention on the politics of the “fight against criminality” and the Kripo’s growing inclusion into the
biological and racial strategies of the Nazi regime.
While it elaborated the characteristics of crime
fighting in the “racial state” of the “Third Reich,”
this research yet also made clear in what extent
police work during the Nazi regime adapted
values, images, policies, and techniques from
pre-Nazi times – and what continuities of policing
can be identified between the German Empire and
the Weimar Republic – the Nazi regime, and the
postwar years.
The interest, research has taken into the history of the uniformed police, has grown significantly in recent years. In particular its
involvement in the mass executions of the Jewish population in Nazi-occupied Eastern Europe
has been analyzed in great detail. Due to this
focus, research on ordinary policing of the
uniformed police in German cities is only gradually emerging. A core issue of the research
Police in Nazi Germany
about the uniformed police is asking for the
factors, which turned policemen from members
of the police force, loyal to the Nazi regime, into
murderous perpetrators. There is currently
a consensus among researchers that a whole
spectrum of factors has to be taken into account
for explaining this shift, ranging from the effects
of political indoctrination to the habituation to
murderous violence, from psychological factors
such as group pressure to situational time- and
location-based factors (Browning 1992; Welzer
2005).
In 1945, when Nazi Germany was defeated,
the victorious allied forces declared the Gestapo,
together with a number of Nazi organizations,
a criminal organization, while the Kripo and the
uniformed police were left out from being
accused of a specific affinity to the Nazi regime.
In those parts of Germany, occupied by the
Americans, the British, and the French, purges
of the police took place, but the overall continuity
remained relatively high. Very many of the detectives and the officers of the uniformed police
continued to work in the police service (F€urmetz
et al. 2001). As the public – at least during the first
decades of (Western) postwar Germany – was
reluctant to a detailed and critical examination
of the Nazi regime, its leading actors, and supporters and as strong political, institutional, and
legal barriers were impeding systematic investigations against Nazi perpetrators, only few of
the former members of the Gestapo, Kripo, and
Orpo were accused and sentenced for their
participation in racial politics and genocide
(Mallmann and Andrej 2009; Ullrich 2011).
It has taken the German police considerable
time to deal with the legacies of its Nazi past.
First projects did set in during the early 1980s
already. These projects, which quite often had the
form of joint teams of police officers and academic historians, were much driven by so-called
“critical” members of the police force, who saw
police history as being part of a police reform
strategy. This strategy aimed at a more civil,
service-orientated, and self-reflecting police.
But as this reform turned into a contested field,
which did not find undivided acceptance among
members of the police force, some of these early
Police in Nazi Germany
police history enterprises came under the
critic from the police, especially as not everyone
among senior ranks of the police was already
“open” for critical views on the history of the
police during the Nazi period. More or less
a decade later, during the 1990s, a second wave
of police history activities, involving big city
police forces, did set in. The – in many aspects –
most prominent among these big city police force
histories was carried out under the auspices of the
Cologne police (Buhlan and Jung 2000). A major
result of this project on the Cologne police had
been to show how deeply and intensively everyday policing had been involved in pursuing the
exclusionary, racial policies of the Nazi regime.
Further projects on the police of major German
cities during the Nazi period confirmed these
findings. In 2007, the Bundeskriminalamt
(BKA) initiated a history project, which was to
deal with the history of the BKA during the 1950s
and 1960s, thereby focussing on the question to
what extent continuities existed within the BKA:
“Continuities” either meant asking for the members of the Bundeskriminalamt, who had served
already in the security forces of the Nazi regime,
or it meant trying to find out what impact (if at all)
these characteristics of the BKA-personnel
had on the practical work, the practices of the
Bundeskriminalamt (Baumann et al. 2011).
Since very recently (2011), projects dealing
with the Nazi legacies within the Bundesnachrichtendienst (Federal External Secret Service)
and the Bundesamt f€
ur Verfassungsschutz (Federal Internal Secret Service) are on the way as
well. The initiatives for these projects are very
much driven by the following logics: Firstly,
dealing with the Nazi legacies of an institution
such as the police, even if highly problematic
aspects might emerge, is increasingly seen as
important for a reflective police culture.
Secondly, the German public generally recognizes positively efforts to deal with the Nazi
past. And thirdly, and last but not least, within
many institutions in the Federal Republic,
a self-understanding has evolved, which sees the
development of the institutional infrastructure of
the Federal Republic as part of a democratic success story – which is the more successful, as this
3637
P
success had evolved on the background of
a terrible past. That is why many of the research
questions put forward in this context are not only
directed on scandalizing continuities (personnel,
practices) beyond the 1945 line within these institutions but to show how successful the respective
institutions had moved toward democracy and
“Rechtsstaat,” the rule of law, despite the burdens
of the Nazi past.
Related Entries
▶ Comparing Police Systems Across the World
▶ Conceptualizing of Police
▶ German Police Until 1918
▶ History of Criminal Investigation
▶ History of Criminological Theories: Causes of
Crime
▶ Police in the Police State
▶ Political Crime
Recommended Reading and References
Ayaß W (1995) ‘Asoziale’ im Nationalsozialismus. Cotta,
Stuttgart
Banach J (1998) Heydrichs Elite. Das F€
uhrerkorps der
Sicherheitspolizei und des SD 1936–1945. Schöningh,
Paderborn
Baumann I, Reinke H, Stephan A, Wagner P (2011)
Schatten der Vergangenheit. Das BKA und seine
Gr€
undungsgeneration in der fr€
uhen Bundesrepublik.
Luchterhand, Köln
Browder GC (1996) Hitler’s enforcers. The Gestapo and
the SS Security Service in the Nazi revolution. Oxford
University Press, New York/Oxford
Browning CR (1993) Ordinary men: reserve police battalion 101 and the final solution in Poland. Harper Collins, New York
Buhlan H, Jung W (eds) (2000) Wessen Freund und
wessen Helfer? Die Kölner Polizei im Nationalsozialismus. Emons, Köln
Curilla W (2011) Der Judenmord in Polen und die
Ordnungspolizei. Paderborn, Schöningh
Dams C, Stolle M (2012) Die Gestapo. Herrschaft und
Terror im Dritten Reich, 3rd edn. Beck, M€
unchen
F€
urmetz G, Reinke H, Weinhauer K (eds)
(2001) Nachkriegspolizei. Sicherheit und Ordnung in
Ost- und Westdeutschland. Ergebnisse, Hamburg
Gellately R (1990) The Gestapo and German society.
Enforcing racial policy 1933–1945. Oxford University
Press, New York, Oxford
P
P
3638
Gellately R (2001) Backing Hitler. Consent and coercion
in Nazi Germany. Oxford University Press, New York/
Oxford
Gellately R, Stoltzfus N (eds) (2001) Social outsiders in
Nazi Germany. Princeton University Press, Princeton/
Oxford
Gutman I (ed) (1990) Encyclopedia of the Holocaust.
Macmillan, New York
Herbert U (2011) Best. Biographische Studien €
uber
Radikalismus, Weltanschauung und Vernunft.
1903–1989, 5th edn. Dietz Nachfolger, Bonn
Klein P (ed) (1997) Die Einsatzgruppen in der besetzten
Sowjetunion, 1941/42. Die T€atigkeits- und
Lageberichte des Chefs der Sicherheitspolizei und
des SD. Edition Hentrich, Berlin
Mallmann K-M, Andrej A (eds) (2009) Die Gestapo nach
1945. Wissenschaftliche Buchgesellschaft, Darmstadt
Mallmann K-M, Paul G (eds) (2000) Die Gestapo im
Zweiten Weltkrieg. ‘Heimatfront’ und besetztes
Europa.
Wissenschaftliche
Buchgesellschaft,
Darmstadt
Mallmann K-M, Paul G (eds) (2004) Karrieren der
Gewalt. Nationalsozialistische T€aterbiographien.
Wissenschaftliche Buchgesellschaft, Darmstadt
Matth€aus J (2003) Ausbildungsziel Judenmord?
‘Weltanschauliche Erziehung’ von SS, Polizei und
Waffen-SS im Rahmen der ‘Endlösung’. Fischer,
Frankfurt/M
Paul G (ed) (2002) Die T€ater der Shoah. Fanatische
Nationalsozialisten oder ganz normale Deutsche?
Wallstein, Göttingen
Paul G, Mallmann KM (eds) (1996) Die Gestapo. Mythos
und Realit€at. Wissenschaftliche Buchgesellschaft,
Darmstadt
Roth T (2010) ‘Verbrechensbek€ampfung’ und soziale
Ausgrenzung im nationalsozialistischen Köln.
Kriminalpolizei, Strafjustiz und abweichendes
Verhalten
zwischen
Macht€
ubernahme
und
Kriegsende. Emons, Köln
Topographie des Terrors (2010) Gestapo, SS und
Reichssicherheitshauptamt in der Wilhelm- und
Prinz-Albrecht-Straße. Eine Dokumentation. Stiftung
Topographie des Terrors, Berlin
Ullrich C (2011) ‘Ich f€
uhl’ mich nicht als Mörder’. Die
Integration von NS-T€atern in die Nachkriegsgesellschaft.
Wissenschaftliche
Buchgesellschaft,
Darmstadt
Wagner P (1996) Volksgemeinschaft ohne Verbrecher.
Konzeptionen und Praxis der Kriminalpolizei in der
Zeit der Weimarer Republik und des Nationalsozialismus. Hamburg, Christians
Wagner P (2002) Hitlers Kriminalisten. Die deutsche
Kriminalpolizei und der Nationalsozialismus
zwischen 1920 und 1960. Beck, M€
unchen
Welzer H (2005) T€ater. Wie aus ganz normalen Menschen
Massenmörder werden. Fischer, Frankfurt/M
Westermann EB (2005) Hitler’s police battalions
enforcing. Racial war in the East. UP of Kansas,
Lawrence
Police in the Police State
Wildt M (2009) An uncompromising generation. The Nazi
leadership of the Reich Security main office. University of Wisconsin Press, Madison
Zimmermann M (1996) Rassenutopie und Genozid. Die
nationalsozialistische ‘Lösung der Zigeunerfrage’.
Christians, Hamburg
Police in the Police State
Justice Tankebe
Institute of Criminology, University
of Cambridge, Cambridge, UK
Overview
“Police state” is a translation of the German word
polizeistaat. The first citation of the term by the
Oxford English Dictionary comes from the Times
(London) of 1851: “Austria has become more of
a police state than before.” The War Illustrated
followed suit in 1939: “spies are everywhere;
indeed, Germany is the modern exemplification
of ‘the police state’ in action.” One way to understand the category of the police state is to frame it
in terms of the question of the relationship
between the “police” and the “state” (Neocleous
2000). Two, albeit, crude distinctions can be
drawn; on the one hand, there are regimes in
which the relationship is intimate to the point
where the police institution is merely
a handmaiden of the state and concerned almost
exclusively with the task of executing regime
orders; the aim is to protect the regime rather
than serve citizens. On the other hand are
regimes that claim to operate in such a way as
to maintain a distance between the state and the
police institution. These two types of regimes
correspond, respectively, with non-liberal and
liberal political cultures. In liberal political
cultures, the police institution is seen as identical
with civil society. For example, it has been
argued that the source of legitimacy for the
British police is a strong moral and emotional
identification between the police and the British
people: Thus, Reith (1956, p. 287) stressed
Police in the Police State
what he considered to be “the historic tradition
that the police are the public and that the public
are the police.” This alleged independence of the
police from the state is the converse of what
happens in police states or totalitarian states.
This entry summarizes three main issues:
First, it examines what is known about the nature
of police states. Second, how do police states
emerge? Third, it discusses what the police do
and how well they do it in the police state. For
example, what is the nature of crime in the police
state, and what are the technologies of crime
control? How reliable is it to assess the effectiveness of the technologies used? Fourth, how do
citizens respond to the police state? Were the
actions of the police considered “legitimate,” or
police states lie outside the consideration of legitimacy? Did citizens live in fear of the police or
they expressed popular support for the police?
Problems of Definition1
The emergence of the modern state in Europe was
associated with the birth of what has been termed
the “well-ordered police state” in the sixteenth
century. It is not that police states were unknown
prior to this period. Indeed, some scholars make
a distinction between the traditional police state
and the modern police state, the latter referring to
the well-ordered police state and the former
describing the state prior to the era of the modern
state.
Why are certain regimes classified as
“police states”? Stated differently, what is the
“police state”? Prima facie, the answer to both
questions is straightforward, and many will easily
cite examples of such regimes. However,
a careful consideration of the literature leads to
a more cautious appraisal of the concept, that one
cannot simply make a binary categorization of
regimes into police states and nonpolice states.
Rather it is more useful to speak of a continuum
of police statehood, with different regimes
1
The essay draws heavily on the work of Arendt (1968)
and Evans (2005).
3639
P
(both democratic and nondemocratic) positioned
on different points of the continuum: totalitarian
regimes at one extreme end of the continuum
and (liberal) democratic regimes at the other
end. One of the few attempts to engage conceptually with the idea of the police state is Brain
Chapman (1970). He set out a number of criteria
for measuring the applicability of the police state
model.
The first is politicization. This is not to be
equated with the idea that the police are political.
All police forces, whether in totalitarian or liberal
democracies, are political in the sense that they
are created by the state and used by the state
to maintain law and order. By politicization,
Chapman is referring to police involvement in
partisan politics and to a situation where police
actions are dictated by partisan political considerations rather than the rule of law. Under
such circumstances, the vicissitudes of police
legitimacy are closely tied up with politics.
Second is militarization of the police; rather
than relying on the army for armed response
where necessary, the police organization chooses
to augment the capacity of its riot unit by providing it with armored vehicles and other equipments to enable it to operate independently of
the army. This is a deeply problematic criterion,
especially when Chapman argues that a state is
on the road to becoming a police state if its
police force weans itself from the control of
the army and operates as an independent state
institution (p. 119). There is no liberal democracy
in which police operations are under the control
or supervision of the army.
Chapman’s third criterion is the Centralization of police services. It is hard to see how can
be seen as a distinctive feature of the police
state. For example, the structural configuration
of a police force does not necessarily correspond
to the character of the political culture in that
country. Countries such as Belgium and Sweden
have centralized police forces, yet they are
scarcely the kind of countries we might describe
as being totalitarian. A decentralized police
institution does not guarantee against the emergence of localized despots, whether such despots are police officers acting on their own or
P
P
3640
under the dictates of local politicians. The Royal
Commission on Police in Britain considered and
rejected this approach to defining the police
state. The Commission argued that the proper
criterion for the police state should be “whether
the police are answerable to the law,” and that in
the police state, “the government acknowledges
no accountability to a democratically elected
parliament, and the citizen cannot rely on the
courts to protect him” (1962, p. 45). But even
this test of police statehood is not without difficulty. For example, it fails to differentiate
between formal or procedural accountability
and substantive accountability. Police states
such as Nazi Germany had a parliament (the
Reichstag) which, it was claimed, represented
the will of the German public and passed laws to
legitimate Nazi actions often in retrospect (The
question of legality and its legitimacy will be
considered in final part of this entry).
Finally, Chapman discusses what he labels
penetration. This, he asserts, entails an encroachment into the judicial domain, with the police
obtaining powers of arrest, supervision, and
detention, and a right to inflict penal sanctions
outside the control of the normal judicial
machinery. It is very hard to claim that this characterization is typical of totalitarian regimes
and that such encroachment is alien in liberal
democracies, even in the twenty-first century;
police forces in all societies are vested with
powers of arrest and detention, although the
duration of detention before trial varies widely.
Nor is it unknown in democratic societies for
police to have powers to administer penal sanctions without recourse to the courts. Take the
example of Britain: Criminal Justice Act 2003
effectively granted police officers the power of
punishment. Specifically, the Act allows officers
to issue cautions to certain categories
of offenders, a legal provision that has been
criticized for its procedural and substantive
unfairness (Brownlee 2007).
While dictionary definitions should not drive
criminological analysis, a look at the dictionary
can often illuminate our grasp of certain
concepts. According to the Oxford English
Dictionary the meaning of a police state is
Police in the Police State
“a totalitarian state run by means of a national
police force, using repressive methods such as
covert surveillance and arbitrary arrest and
imprisonment to control the population.” Here
again, the idea of a national or centralized
police force is set out as a feature of totalitarianism, yet it is never entirely clear why centralization should be anathema to democratic
governance. Even the notion of “repressive
methods” is not without difficulties. Although
a liberal democracy, the United Kingdom is
considered to be one of the most closely surveilled countries in the world. For example, the
City of London has 69 closed circuit television
(CCTV) cameras per 1,000 population; Wandsworth, an area of 4.6 miles, has a total of 1,113
CCTV cameras which is far more than the
CCTV cameras of police departments of Boston
(USA), Johannesburg (South Africa), and
Dublin City (Ireland) combined (BBC 2009).
This level of surveillance was unparalleled by
the experience of Nazi Germany or any other
police state.
In the end, all regimes target certain sections
of their populations (for example, migrants and
suspected terrorists), and embark on elaborate
surveillance against them. It may be a bold
claim but it may not be delusional to suggest
that it is not the fact of particular actions or
modes of operation of particular regimes that
ipso facto make those regimes totalitarian or
police states; many liberal democracies share
with totalitarian regimes the use of intensive
covert surveillance, imprisonment as a means of
control of some sections of their population, and
encroach upon the powers of the judiciary. Far
more important might be how these technologies
are employed and what mechanisms exist for
substantive democratic accountability. In other
words, deploying the category of police state to
label non-liberal states may not help to illuminate
the difference between them and liberal states
(Neocleous 2000). It is far more useful to pay
attention to how the police in both liberal and
non-liberal societies treat citizens during everyday encounters. As a dictum, claims to democratic governance “would hardly be allowed to
go unchallenged if the police severely restricted
Police in the Police State
public meetings and political demonstrations or
resorted readily to physical force and intimation
in order to prevent crime” (Bayley 1969, p. 11).
Nor is it unreasonable to observe that even within
democratic societies, the experiences and perceptions that some sections of society have of
the police is akin to what citizens of police
states experience. Put colloquially, one person’s
democratic state might be another person’s
police state.
Emergence of Police States
The preceding comments about the conceptual
challenges in making sense of the category of
police state do not mean that such a category is
without merit. It has been argued that concepts
are the gateway to the empirical world of study
for empirical science, and therefore, the effective
functioning of concepts is a matter of decisive
importance (Blumer 1954). However, one cannot
hope to deploy concepts effectively withoutclarity in their meaning. That is not a task necessarily attempted here; the objective so far has
been more modest, and it is to draw attention to
the need for further work in order to understand
more fully its analytical utility. Nonetheless,
there are examples of states in recent history
widely described as police states: Nazi Germany
and Stalin’s Russia.
Under what conditions do police states
emerge? It is difficult to answer this question
without the risk of contextless generalizations.
The coming to power of Stalin and Hitler
occurred under completely different conditions.
Russian was a relatively backward country, with
a weak economy and history of despotic rule.
Germany, on the other hand, was a welldeveloped economy; it had a highly educated
population with reasonably strong institutions
and a burgeoning democracy. Indeed, as
discussed below, Hitler came to power through
elections. It is important to expatiate on the
Germany situation further, because it demonstrates that economic development and education
do not necessarily suppress the emergence of
totalitarian rule (Fukuyama 1992).
3641
P
Toward the end of the nineteenth century,
Germany had embarked upon its first experiment
with democratic governance, and the experiment
had gone spectacularly wrong. Parliament was
toothless, unable to check executive power. The
rights of workers were not fully recognized and
industrialists had mounted strong opposition
against moves toward unionized labor, arguing
that unionized labor was an impediment to
economic growth. Nor could it restrain encroachment upon the civil liberties of some sections of
the population, in particular Catholics and Jews,
both of which were considered “enemies of the
Reich” (Evans 2005). At the same time, extremist
political parties were increasingly gaining prominence. These parties tapped into the anxieties
and resentment of ethnic Germans who felt
increasingly disillusioned by the major economic
and social transformation in German society at
the time. In many Western societies today,
migrants and asylum seekers are easy targets for
such anxieties (Bauman 2004). In Germany, Jews
were the target; they were seen as the source of all
societal problems. Extremist politicians argued
for their civil liberties and economic activities
to be restricted. Mainstream political parties
capitulated. Finally, defeat in the First World
War brought economic burdens, and there
was also increased violence. The judiciary was
a tainted institution, widely perceived as partisan
in favor of reactionary elements.
It was within this context that Nazis came
to power under the leadership of Adolf
Hitler “who possessed one great gift: the
ability to move crowds with his rhetoric”
(Evans 2005, p. 7). The first electoral success
for the Nazis was in September 1930 and
July 1932, largely with support from the middle
classes which had felt threatened by the
election promises of the Communists to restrain
capitalism. After various political maneuvers in
parliament, Hitler was appointed the head of
a new government in 1933. In the initial stages
of the development of the police state, there is
often a struggle for power and the focus of the
regime is to liquidate all forms of organized
resistance, both open and secret. Nazi Germany
was not an exception. It established an extensive
P
P
3642
surveillance system that tracked down, arrested,
and punished people opposed to it. So effective
was this pursuit that the regime succeeded in
crushing completely all forms of organized
opposition. The Nazis banned all political parties,
transforming Germany from a fledging
multi-democratic state into a one-party state
under the leadership of Hitler. In 1934, a
nationwide plebiscite approved a law that made
Hitler the Leader and Reich Chancellor; his
authority was “total and all-embracing. . .and subject to no checks or controls” (Evans 2005, p. 44).
One of the features of the police state is the
dominance of the police in domestic politics. The
military tends to occupy peripheral positions.
There are at least two interrelated reasons for
this state of affairs. First, regimes in police states
tend not to have confidence in the military to
make the necessary cognitive shift from its raison
d’être of dealing with external threats to that of
considering its own people as if they were foreign
combatants. Second, police states aspire eventually to establish a world government (Arendt
1968). Consequently, they tend to approach the
victims of their foreign aggressive adventures as
though they were rebels, and therefore prefer to
govern them with the police rather than the
military.
The Police State and Crime Control
To understand the effectiveness of the mechanisms of crime control, it is prudent to first
consider what law and crime mean under the
police state. A key feature of law in democratic
societies is that it brings a certain degree of certainty and predictability into aspects of relationships among citizens. The law is also applied
prospectively, thereby allowing those subject to
power to know in advance when they will be
subject to coercion and thus avoiding needless
interference in their lives. The opposite is true
of police states; unpredictability and uncertainty
were the main features of the law both in
Germany under Hitler and Russia under Stalin.
The law and procedures for its application are
always in “continuous flux.”
Police in the Police State
In Nazi Germany, two states existed contemporaneously: the normative state and the prerogative state (Fraenkel 1941). The former was
the formal institutions that existed prior to
Hitler coming to power, and were bounded by
long-standing rules, laws, and procedures. The
Prerogative state was an extralegal institutional
arrangement that derived its legitimacy from
Hitler. Although initially characterized by conflict and sometimes accommodation, the relationship between the two eventually became one of
subjugation; the spirit of the prerogative state
permeated the normative state as it abandoned
its legal procedures and gave approval to hitherto
illegal state actions. Hitler’s word was law, and
the legally correct procedure was as he deemed it.
At any rate, the courts of the normative state were
dominated by Nazi sympathizers; the Nazis
disregarded the law, including laws enacted by
themselves if this suited their convenience.
Within this context is to be expected shifting
definitions of crime and their punishments.
Criminologists have long recognized that
crimes are social constructions and that there is
a great deal of variation between societies in
perceptions of crime seriousness. Thus, the kind
of behavior, utterance, or dissent that would be
tolerated in liberal democratic societies is viewed
as criminal in police states. For example, under
the Malicious Gossip Law of 1934 in Nazi
Germany, it was an offense to make “spiteful or
provocative statements” in public against the
governing party, to criticize its policies, or complain about suppression of freedoms and civil
liberties. Whether the offense was for malicious
gossip or not, the social identity of the offender
was always an important consideration in determining the swiftness of police response and the
severity of sentences by the courts. Thus, Jews,
for instance, were punished more severely for
criminal offenses that were otherwise overlooked
or attracted very lenient sentences if committed
by non-Jews.
Once the police state has been established,
there is a shift from dealing with the “suspect”
to dealing with the “objective enemy” (Arendt
1968). The objective enemy is unlike the suspect
in many ways. Suspects are people whose
Police in the Police State
previous and present deeds and “dangerous
thoughts” give reason for the state to be suspicious about what they might do. They are people
who oppose and desire to overthrow the government, or are believed to have committed crimes.
The objective enemy, on the other hand, is the
product of government policy irrespective of the
orientations of those so defined. The objective
enemy is seen as a “carrier of tendencies” considered a threat to the state. The government
never runs out of objective enemies because
new ones are easily found depending on the circumstances (Arendt 1968). For example, the Nazi
regime had well-developed plans for the extermination of the Polish people when the extermination of Jews was near completion (Arendt 1968).
This was to involve, among other things, attempts
to impose regulations strikingly similar to what
had been done before implementing “the final
solution”: for example, change of names and the
death penalty for racially undesirable marriages,
such as those between Germans and Poles
(Evans 2005).
The secret police constitute the main channel
through which orders of the government
are transmitted. In addition to the change from
dealing with suspects to attending to objective
enemies is a corresponding change from
suspected offense to possible crime:
. . .every thought that deviates from the officially
prescribed and permanently changing line is
already suspect, no matter in which field of
human activity it occurs. Simply because of their
capacity to think, human beings are suspects by
definition, and this suspicion cannot be diverted
by exemplary behavior, for the human capacity to
think is also a capacity to change one’s mind.
Since, moreover, it is impossible ever to know
beyond doubt another man’s heart . . .suspicion
can no longer be allayed if neither a community
of values nor the predictabilities of self-interest
exist as social realities. (Arendt 1968, p. 430)
The result of this is to create a system of
ubiquitous spying, infusing hitherto healthy
social relationships with mutual suspicion. Each
person feels under constant surveillance; what
was once methods employed exclusively for
dealing with the population is now the methods
that ordinary citizens employ to deal with their
3643
P
neighbors in everyday encounters. As discussed
later, however, scholars differ on whether fear,
terror, and intimidation were as pervasive as it
is often portrayed in popular accounts, or it was
the case that citizens freely consented to the
decisions and directives of the police.
In the initial years of the establishment of the
police state, the focus is primarily upon eliminating all sources of organized opposition. Concern
with suspects of possible crimes assumes prominence when the regime is fully established. As the
police state enters its last and fully totalitarian
stage, it abandons the category of objective
enemy and possible crime. The new category of
“undesirables” replaces them. Here victims are
chosen at random and officially declared unfit to
live. In the well-documented case of the Nazis,
undesirables included the mentally ill and persons with certain kinds of diseases. This introduction of arbitrariness means that the police
state becomes far more efficient in suppressing
freedoms; both the innocent and the guilty
share the fate of being undesirable. But more
importantly, this change in the concept of crime
and criminals comes with new methods for
law enforcement. Consistent with their task of
executing executive orders, it is the responsibility
of the police to ensure criminals are punished and
undesirables disappear, often without a trace.
The only trace victims leave behind are the
memories of those who knew them; a challenging
task for the secret police is to ensure that even this
trace will disappear with the victims. The
Russian secret police, for instance, employed
elaborate network analysis of families, friends,
and acquaintances of suspects with the aim
to eliminating those believed to possess
“dangerous memories” of the disappeared. Hannah
Arendt has captured this well:
In totalitarian countries all places of detention
ruled by the police are made to be veritable holes
of oblivion into which people stumble by accident
and without leaving behind them such ordinary
traces of former existence as a body and a grave.
Compared with this newest invention for doing
away with people, the old-fashioned method of
murder, political or criminal, is inefficient indeed.
The murderer leaves behind him a corpse, and
although he tries to efface the traces of his own
P
P
3644
identity, he has no power to erase the identity of his
victim from the memory of the surviving world.
The operation of the secret police, on the contrary,
miraculously sees to it that the victim never existed
at all. (Arendt 1968, pp. 434–435)
In common with their counterparts in democratic societies, the lifeblood of police work in
the police state is information from the
public. Two decades of empirical analyses of
police legitimacy have shown that cooperation
with police forces rests mainly on the perceived
legitimacy of the police (Tyler 1990). Police
states and their secret police forces were fully
aware of this role of legitimacy in securing the
flow of intelligence, and they devoted much
energy and resources cultivating legitimacy
among their publics. In Nazi Germany, the
Gestapo relied upon the assistance of the German public, and there is strong evidence to show
that many Germans responded positively,
denouncing their neighbors, colleagues, and relatives to the secret police. The police in the
police state also rely heavily on extensive networks of informers, both impressed and voluntary informers (Chapman 1970). The use of
informers is not exclusive to police states; it is
a technique that is also widely used by law
enforcement agencies in democratic societies
(Natapoff 2009). But informers are not always
reliable sources of intelligence. The police therefore seek to operate their own sources of intelligence gathering by infiltrating areas of social and
political life. Again, the strategy of infiltration is as
much a practice of police states as it is of democratic states.
The police in Nazi Germany were effective
in destroying initial opposition to the regime
and subsequently perceived enemies and undesirables such as Jews and social outsiders. Levels of
petty criminality were however comparable to
those that pertain in democratic societies.
Johnson (2011) has shown that a large proportion
of Germans committed crimes on a frequent
basis, without much fear of being detected and
punished. The reason for that situation was not
ineffectiveness by the police; on the contrary, it
was because the police prioritize other crimes
over petty street-level offenses.
Police in the Police State
Citizens and the Police State
How do citizens react to police states? This
question divides historians of Germany under
Hitler and Russia under Stalin, the two quintessential cases of police state in the twentieth
century. There are two broad schools of thought.
One school of thought holds that citizens
consented voluntarily to the police state. Some
studies have produced evidence to show that the
Gestapo, for example, was a much smaller
organization than it was often thought. Hitler
was “so immensely popular among most
Germans that intimidation and terror were rarely
needed to enforce loyalty” (Johnson and Reuband
2005, p. 329). Gellately (1990) cites election
results in support of a view that the Nazi regime
was popular among the German people. The
coercion the regime applied was aimed at only
a small proportion of minorities but that was even
done with the approval of the majority of the
population. Apart from election results, the
other reason for the claim that citizens did not
experience the police state as an “all knowing, all
powerful, and omniscient” is the size of the secret
police. The ratio of secret police to citizens is
estimated at 1 per 10,000 residents in the cities
of the Third Reich, and usually with extremely
limited presence in rural areas (Johnson 2011).
The secret police, it is said, were after all not
always well resourced, and therefore could not
be said to have relied systematically on terror and
coercion to induce citizen compliance. As further
evidence, scholars argue that the reaction of the
ordinary population was that of voluntary supply
of information through denunciation of neighbors
(Johnson 2011).
A second school of thought contends that
police organizations of police states were omnipresent and omnipotent, and that they employed
arbitrary powers, repressive tactics, and unimaginable brutality against citizens. Consequently,
citizens lived in fear and terror, mindful of the
potentially ruthless repercussions from the
police. Elections, it is argued, lacked integrity
and therefore a poor measure of popular attitudes
toward the police state. Far from being a society
of “self-policing,” terror was widespread.
Police in the Police State
The Nazis regularly publicized executions, court
proceedings, and sentences. The aim of the
publicity was to deter potential offenders from
underestimating the risk of being caught and
punished (Evans 2005).
It is fair to say that police states thrived upon
a mixture of coercion and terror, and popular
support both from within the security forces
and the general population. There is no evidence
throughout history to show that any regime
relied exclusively on its ability to physically
intimidate its subjects or citizens into obedience,
not even the most “unjust and blood-minded
dictatorship” (Fukuyama 1992, p. 16). The
Nazis, for instance, “did not just seek to batter
the population into passive, sullen acquiescence.
They also wanted to rouse it into positive, enthusiastic endorsement of their ideals and their policies, to change people’s minds and spirits and to
create a new German culture that would reflect
their values alone” (Evans 2005, p. 118). The
question that rises is whether one can speak of
legitimacy under totalitarian regimes. It is
a question that recalls a long-standing debate
among scholars. Social scientists approach legitimacy (that is, recognition of the moral rightness
of power) within the particular historical societies rather than universally. They are fully aware
that what makes power legitimate in one society
may differ from others, and that the conditions
for legitimacy in one may be repudiated by
another. This approach does not require the
social scientist to make any judgment about the
appropriateness or otherwise of the social order
she investigates. A clear implication of this line
of thought is that it is possible for a researcher to
conclude that a police organization is legitimate
in the empirical sense (that is, it finds wide moral
acceptance among citizens) and yet for that
researcher to believe that that organization is
deeply unjust or even “evil.”
It is on this conclusion that moral philosophers
and some political scientists depart from
Max Weber. For them, legitimacy relates to
whether by some objective standards of ethical
evaluation, a claim to legitimacy can be recognized as valid. The contention is that one cannot
simply reduce legitimacy to a matter of fact, the
3645
P
fact that citizens hold a certain belief about
a regime. On the contrary, the concept should
signify a normative evaluation of the correctness
of the procedures, the justification for decisions,
and the fairness with which regimes treat their
subjects (Grafstein 1981). The full implication of
this approach is that there is a need to connect
analysis of legitimacy to theories about justice
(Bottoms and Tankebe 2012). That is
a discussion beyond the scope this entry. The
point to note here, however, is that a focus on
the public actions or behavior of citizens as the
measure of their reactions to a regime, especially
one that exercises unrestrained power, is unlikely
to tell the whole story about power relations;
indeed, it is likely to lead to the erroneous conclusion that “subordinate groups endorse the
terms of their subordination and are willing,
even enthusiastic, partners in that subordination”
(Scott 1990, p. 4). What is known so far about the
reactions of citizens to police states offers no
assurance against this error.
Conclusion
It is commonplace to read from newspapers, and
even from scholarly work, that some democratic
societies are on a pathway to becoming police
states. Such claims often arise from evidence
that those states have undertaken certain measures to facilitate a greater surveillance of the
population or a section of that population
(e.g., the US Patriot Act and the UK CONTEST
strategy for counterterrorism) and other forms of
intrusion into private liberties of citizens. By any
account, these are poor indicators of the police
state; no state, democratic and totalitarian, can
hope to respond effectively to threats in the modern world without some level of surveillance of
(some) its population.
The absence of conceptual clarity regarding
the category of police state suggests that the
greatest immediate need in assessing the utility
of that category has to start with attempts to
clarify its meaning. It appears that criminology
can contribute toward that endeavor from two
interrelated standpoints. First is to return to
P
P
3646
Edwin Sutherland’s well-known definition of
criminology as the study of “the processes of
law-making, of law-breaking, and of reacting
to the breaking of laws” (Sutherland 1939,
p. 1). At the heart of this triumvirate is the
question of power, and how it is exercised.
A second and related issue is for criminology
to engage with the political science literature
with the aim of procuring a proper understanding of democracy. That literature suggests that
democratic societies are characterized by
a universalistic ideological claim about human
equality, that all human beings are of equal
worth and therefore entitled to equal respect
and treatment. Consequently, any exercise of
power must do so with respect for the principles
of individual liberty and equality. Democratic
societies attempt to achieve this through mechanisms of democratic accountability of the
police, including external oversight of police
institutions that is independent of government.
How effectively such oversight works is open to
debate. What is true is that democracies differ
from police states on this check on police power,
and it is here that the search for conceptual
clarity should begin.
Related Entries
▶ Conceptualizing of Police
▶ Counterterrorism
▶ Democratic Policing
▶ Electronic Monitoring
▶ False Confessions and Police Interrogation
▶ Policing of Peacekeeping
▶ Surveillance Technology and Policing
Recommended Reading and References
Arendt H (1968) The origins of totalitarianism. Harcourt,
London
Bauman Z (2004) Wasted lives: modernity and its
outcasts. Polity Press, Cambridge
Bayley DH (1969) The police and political development
in India. Princeton University Press, Princeton
Blumer H (1954) What is wrong with social theory? Am
Sociol Rev 19:3–10
Police in the Police State
Bottoms AE, Tankebe J (2012) Beyond procedural justice:
a dialogic approach to legitimacy in criminal justice.
J Crim Law Criminol 102:119–170
Browder GC (1990) Foundations of the Nazi police state:
the formation of Sipo and SD. University Press of
Kentucky, Kentucky
Brownlee ID (2007) Conditional cautions and fair trial
rights in England and Wales: form versus substance
in the diversionary agenda? Crim Law Rev 129–140
Chapman B (1970) Police state. Pall Mall Press, London
Coicaud J-M (2002) Legitimacy and politics:
a contribution to the study of political right and political responsibility. Cambridge University Press,
Cambridge, UK
Evans RJ (2005) The third Reich in power: how the Nazis
won over the hearts and minds of a nation. Penguin,
London
Fraenkel E (1941) The dual state: a contribution to the
theory of dictatorship. Oxford University Press,
New York
Fukuyama F (1992) The end of history and the last man.
Penguin Books, London
Gellately R (1990) The Gestapo and German society:
enforcing racial policy, 1933–1945. Oxford University
Press, Oxford
Giddens A (1985) The nation-state and violence. Polity
Press, Cambridge
Grafstein R (1981) The failure of Weber’s conception of
legitimacy: its causes and implications. J Polit
43:456–472
Johnson EA (2011) Criminal justice, coercion and consent
in ‘totalitarian’ society: the case of national socialist
Germany. Brit J Criminol 51:599–614
Johnson EW, Reuband K-H (2005) What we knew: terror,
mass murder, and everyday life in Nazi Germany: an
oral history. Basic Books, Cambridge, MA
Lukes I (1996) The birth of a police state: the Czechoslovak ministry of the interior, 1945–48. Intell Natl Sec
11:78–88
MacCormick N (2007) Institutions of law: an essay in
legal theory. Oxford University Press, Oxford
Natapoff A (2009) Snitching: criminal informants
and the erosion of American justice. NYU Press,
New York
Neocleous M (2000) The fabrication of social order:
a critical theory of police power. Pluto Press, London
Raeff M (1983) The well-ordered police state: social and
institutional change through law in the Germanies and
Russia, 1600–1800. Yale University Press,
New Haven/London
Reith C (1956) A new study of police history. Oliver &
Boyd, London
Scott JC (1990) Domination and the arts of resistance:
hidden
transcripts.
Yale
University
Press,
New Haven/London
Sutherland EH (1939) Principles of criminology.
Lippincott, Philadelphia
Tamanaha BZ (2004) On the rule of law. Cambridge
University Press, Cambridge
Police Integrity
3647
P
Tyler T (1990) Why people obey the law. Yale University
Press, Yale
Weber M (1978) Economy and society: an outline of
interpretive sociology (Roth G, Wittich C, eds),
2 vols. University of California Press, Berkeley
Zureik E, Salter MB (2007) Global surveillance and policing: borders, security, and identity. Willan Publishing,
Cullompton
history on police integrity. The accompanying
methodological approach has been utilized to
survey police officers and to detect the contours
of police integrity in nearly 20 countries across
the world. Other research on police integrity
explores the mechanisms used to enhance integrity and accountability, such as the early warning
systems and citizen reviews.
Police Integrity
Police Integrity
Sanja Kutnjak Ivković
School of Criminal Justice, Michigan State
University, East Lansing, MI, USA
Introduction
In the 1990s, a wave of police misconduct
scandals shook the country, from the police
corruption scandal in New York, which prompted
the establishment of the Mollen Commission, the
Rodney King beating, which resulted in the
establishment of the Christopher Commission,
the Abner Luima sexual violence case in New
York, which resulted in $8.75 million dollar
settlement, and the Rampart Division scandal in
Los Angeles. In 1997, the Office of Community
Oriented Policing Services (COPS) and the
National Institute of Justice (NIJ) held
a national symposium bringing together policy
makers and social scientists to address this
emerging problem. A natural tendency would be
to anchor the conference on the events that signal
the lack of integrity. That is, the discussion would
center on various forms of misconduct and the
ways of controlling them. As Hickman and
colleagues (2004, p. 1.1) put it, “[a]ll too often
we infer integrity from its demonstrated absence.
That is to say, research that focuses on corruption
infers ideas about integrity absent its measurement.” Instead, the COPS/NIJ team decided to
use the term police integrity as the guiding theme
for the conference (Greenberg 1997, p. v):
Overview
Police integrity and police misconduct are
closely related; traditionally, the discussion of
police integrity typically would focus on the
events that signal the lack of integrity (e.g.,
forms of police misconduct and the ways of controlling them). Since the mid-1990s, the focus has
shifted on police integrity, which, in turn, allows
for the exploration of a broader concept – the
development of a high-integrity police agency.
Police integrity is defined as “the normative inclination among police to resist temptations to
abuse the rights and privileges of their occupation” (Klockars et al. 2006).
Although police integrity could be associated
with the moral virtues of individual police
officers, starting with the writings by Herman
Goldstein in the 1970s, the idea has started to
develop that this “bad apple” theory of police
corruption is inadequate and that a more
organizational approach should be used instead.
Klockars and colleagues (Klockars and Kutnjak
Ivković 2004; Klockars et al. 1997) built upon
this view and proposed their theory of
police integrity. The theory features four
dimensions:
organizational
rulemaking,
detection, investigation, and discipline of rule
violations, curtailing the code of silence, and the
influence of public expectations and agency
The focus on “police integrity” opened a whole
new domain. Although previous research, study,
and experimentation had focused on critical issues
such as corruption and excessive use of force, these
approaches, had, in fact, revolved around single
dimension. As a result, the solutions were
constricted in that they were derived out of a need
to control unwanted behaviors of individuals. In
comparison, police integrity guided the focus on
P
P
3648
the broader domain of developing a healthy organization that would serve to reinforce and maintain
the good character and constructive motivations of
many of the individuals joining the ranks of law
enforcement.
In July of 1996, about 200 police leaders,
politicians,
lawyers,
and
researchers
attended a three-day National Symposium
on Police Integrity in Washington, D.C.
(Gaffigan and McDonald 1997). Among the
participants who tried to define police integrity,
the preferred approach was to envision a police
officer with high integrity and list
the characteristics (e.g., virtues, values, and
character traits) that separate this police officer
apart from other police officers of lower integrity.
The characteristics explicitly listed at the
Symposium
were
allegiance,
courage,
honor, honesty, prudence, trust, effacement of
self-interest, intellectual honesty, justice,
morality, principled behavior, responsibility,
and dedication to mission (Vicchio 1997). Mark
Moore was in charge of summarizing the findings
of the symposium; he wrote that “[w]hat we mean
by integrity and professionalism is law-abiding
character, technical confidence, neutrality,
distance – in Steve Vicchio’s wonderful phrase,
‘the effacement of personal interest’ – and
probably some notion of courtesy and client
responsiveness” (Moore 1997, p. 63). Similarly,
Hickman and colleagues wrote in their
introductory remarks to a co-edited book on
police integrity that “police integrity refers to
the underlying values and ethical attachment of
the police and how those values and ethics affect
police behavior” (Hickman et al. 2004, p. 1.1).
These lists of characteristics contain expected
individual police officers’ character traits. As
Klockars and colleagues argue (2006, p. xiv),
“[t]he fact is that any virtue – charity,
compassion, decency, faith, loyalty, passion,
patience, and perspective, to add but a few to
the list mentioned at the Symposium – is at least
arguably as crucial to the integrity of
a professional police officer as any other and
this approach gives us no guidance whatsoever
in prioritizing this litany.” Yet, some of the
officers who exhibit these characteristics may
Police Integrity
decide not to act in accordance with high
integrity, just like the police officers who do not
possess many of these characteristics may decide
to act in accordance with high integrity. Finally,
these character traits are features of individuals
and they close the possibility that police integrity
could be an organizational feature as well.
Since 1996, the issue of police integrity has
been discussed at several conferences (“the 1999
NIJ/NYU Seminar on Police Integrity and
Democracies;
the
1999
Strengthening
Police-Community Relationships conference;
the 2000 biannual conference Policing in Central
and Eastern Europe: Ethics, Integrity, and
Human Right (Pagon 2000); the 1998 Sixth
International Conference on Ethics: Integrity at
the Public-Private Interface (Huberts and van
den Heuvel 1999”)), and continues to be explored
bi-annually at corruption conferences such as the
Global Forum against Corruption and
Safeguarding Integrity and the Transparency
International’s International Anti-Corruption
Conference. The US Department of Justice
provided a publication addressing the Principles
for Promoting Police Integrity (2001). The
co-edited book Police Integrity and Ethics
(Hickman et al. 2004) contains chapters
addressing various aspects of police integrity.
However, the understanding of what police
integrity is, and what should be covered under
this topic, varies substantially across conference
organizers, presenters, and authors. A substantial
portion of the literature seeks to develop integrity-enhancing mechanisms or mechanisms that
would control police misconduct (Pope 2000;
Transparency International 2001; U.S. Department of Justice 2001; See ▶ Control of Police
Misconduct).
Yet, it seems that even the basic debate regarding what constitutes police integrity has not been
resolved. Sam Walker, a leading expert on police
accountability, points out that “[t]he issue of
police integrity is extremely important and has
received an increasing amount of public attention
among policy makers and the general
public. Unfortunately, the academic literature
has not adequately addressed this very important
subject” (2004, p. vii).
Police Integrity
Definition
Klockars and colleagues (Klockars et al. 2006)
have provided a definition of police integrity that
goes beyond the traditional focus on individual
police officers’ character traits and allows for the
organizational approach toward it. They have
also accompanied the definition with a theory of
police integrity and the methodological approach
toward its measurement. Klockars and colleagues
(2006) define police integrity as “the normative
inclination among police to resist temptations to
abuse the rights and privileges of their
occupation.” According to the authors
(Klockars et al. 2006), this definition has six
critical components.
Normative – The normative aspect of the
definition emphasizes that integrity is based on
the moral norms of right and wrong
(Klockars et al. 2006, p. 1) that tell people what
they should or should not do. Thus, integrity
combines a belief in moral values with an
inclination to behave in accordance with that
belief; “[j]ust as a belief in honesty inclines one
to avoid lying and a belief in fidelity obliges one
to be faithful, integrity requires not only a belief
that certain behaviors are right or wrong, but also
actions that are in accord with those beliefs”
(Klockars et al. 2006, p. 2).
Inclination to Resist – This part of the
definition emphasizes the difference between
attitudes and behavior. The authors do not expect
that the person of high integrity will always
behave in accordance with these values. Rather,
they discuss a general problem that people who
believe in honesty sometimes lie; people
who believe in fidelity sometimes are unfaithful;
and people of integrity sometimes do things they
know are wrong. Three dimensions of the relation
between police attitudes of integrity and police
misconduct are addressed.
The first point the authors emphasize is
that, while integrity describes the normative
inclination to resist temptations, integrity is not
the only source (Klockars et al. 2006, p. 3).
Other possible candidates include the lack of
imagination, lack of opportunity, fear of
discovery and public humiliation, shame,
punishment, or a simple unfavorable risk/reward
3649
P
calculus as potentially critical in ensuring that
police officers behave honestly. The second
point the authors emphasize is that that attitudes
of integrity put at least some pressure on
police officers who share them to actually avoid
wrongful behavior. Because these attitudes of
integrity may lead police officers to comply
with the rules and, at the same time, may not
necessarily be the driving force behind compliant
behavior, the authors have decided to refer to
integrity as the inclination to resist rather than
the actual resistance to temptations. The third
point the authors emphasize is that the
direction of the causal relation between attitudes
of integrity and behavior of integrity is not
always absolutely clear. In particular, while
attitudes may cause behavior, behavior may also
cause attitudes. Requiring of police officers to
behave in accordance with integrity may lead
them to adopt the attitudes of integrity as well.
Police – Klockars et al. (2006) emphasize that
integrity is a characteristic of the “police,”
without specifying that this is a characteristic of
individual police officers or police agencies. This
reflects the view that integrity may be a feature of
an individual police officer, a group of police
officers, a subunit of a police agency, a whole
police agency, or police agencies. The authors
caution that the discussion about police integrity
easily lapses into the discussion of integrity at
the police officers’ individual level. Yet,
police integrity is also an organizational
feature – a feature of police agencies, organizations, and institutions – and we can discuss
organizational cultures of police integrity.
However, the discussion of the dynamics and
the relevant correlates of police integrity will
differ depending on the level of integrity under
discussion (e.g., individual, organizational);
“[h]ow one understands and explains the
psychology of integrity of an individual
police officer will most certainly differ from the
understanding and explanation of the evolution of
a culture of integrity in a police agency”
(Klockars et al. 2006, p. 3).
Temptations – The motivation or the reasons
why police officers engage in misconduct
could vary dramatically across forms of police
P
P
3650
misconduct, individual police officers, and
police agencies. The most obvious of these
temptations is gain, the monetary or
non-monetary reward for type of police
misbehavior, a typical feature of traditional
police corruption. However, other forms of police
misconduct are not necessarily motivated by
gain; different forms of police misconduct are
by no means the product of a similar, singular,
or even ignoble temptation. The example of use
of excessive force is particularly illustrative
(Klockars 1995, p. 17):
. . .[the excessive force] need not (and usually will
not) be the product of malicious or sadistic behavior. It can spring from good intentions as well as
bad, mistakes and misreading, lack of experience,
overconfidence, momentary inattention, physical
or mental fatigue, experimentation, inadequate or
improper training, prejudice, passion, an urge to do
justice or demonstrate bravery, misplaced trust,
boredom, illness, a specific incompetence, or
a hundred other factors that might influence an
officer to behave in a particular situation in a less
than expert way.
Klockars et al. (2006) conclude that the
methods used to control misconduct should differ
depending on the type of temptation and argue
that the methods used to control corruption could
be quite different for the methods used to control
the use of excessive force. They also infer not
only that the contours of integrity may be very
different in different police agencies, but also that
the contours of police integrity within the same
police agency could be different depending on
the form of police misconduct. They argue that
police integrity need not be a uniform phenomenon. Consequently, “it is not difficult to imagine
a police organization or subculture that was
highly intolerant of officer theft, soliciting bribes,
taking kickbacks and other acts of corruption and
at the same time was much more accepting of
discourtesy, excessive force, perjury, forging
records, fabricating evidence, or unwarranted or
illegal searches” (Klockars et al. 2006, p. 4).
The authors proceed to note that one could also
expect that police officers who succumbed to the
temptations of one variety (e.g., gain) may also
be more likely to succumb to the temptations of
another variety (e.g., prejudice).
Police Integrity
Abuse – A crucial element of the definition is
the concept of abuse. In many situations, the
abuse of office is obvious (e.g., a police officer
accepts a bribe in exchange for letting the citizen
caught violating the speeding limit). However,
the discussion of whether particular behavior is
abusive is confounded by two possible arguments. The first argument flatly denies that the
behavior in question is abusive. For example, the
police officers’ acceptance of discounts, holiday
gifts, free food, and liquor could be viewed as
gestures of goodwill, hospitality, and/or gratitude. The second argument recognizes that the
behavior could be abusive, but justifies and
excuses it as an expression of “street justice.”
As Klockars and colleagues (Klockars et al.
2006, p. 5) elaborate, “[a] common theme is that
police are ‘human’ and cannot be expected to
behave without normal human emotions in
situations in which they are insulted, defied,
assaulted,
deceived,
shocked,
repulsed,
disgusted, or horrified by the conduct of those
they police.”
The Rights and Privileges of Their Occupation – Policing is a highly discretionary, coercive
activity that routinely takes place in private
settings, out of the sight of supervisors, and
before witnesses who are often regarded as
unreliable. Policing as an occupation is rife with
opportunities for misconduct, as independent
commission reports and scholarly studies clearly
demonstrate (see, e.g., Christopher Commission
1991; Knapp Commission 1972; Manning and
Redlinger 1993; Mollen Commission 1994;
Pennsylvania Crime Commission 1974).
Causes of Police Integrity
Until relatively recently, the prevailing
administrative view of integrity (at least in the
United States) was to associate integrity with the
moral virtues of individual police officers;
consequently, the tendency was to fight
corruption, for example, by carefully screening
applicants for police positions, pursuing
defective officers aggressively, and removing
them from their police positions before their
behavior spreads throughout the agency (see,
e.g., Knapp Commission 1972). In the 1970s,
Police Integrity
Herman Goldstein (1975) embarked on the
pioneering work and argued that this “bad
apple” theory of police corruption is inadequate
and that a more organizational approach should
be used instead.
Klockars and Kutnjak Ivković and their colleagues (Klockars and Kutnjak Ivković 2004;
Klockars et al. 1997) built upon this view and
proposed their theory of police integrity.
This approach stresses the importance of four
distinct dimensions, each of which is profoundly
organizational in nature.
Organizational Rulemaking – The first
dimension of the organizational theory of police
integrity addresses organizational rule making
(see, e.g., Klockars and Kutnjak Ivković 2003;
Klockars et al. 1997) and relates police
misconduct to the ways police agencies create,
teach, and enforce rules explicitly prohibiting
misbehavior (Klockars and Kutnjak Ivković
2003, p. 1.4). In countries with highly
decentralized police organization, as is the case
with the United States, police agencies could
differ dramatically in the nature of the behavior
they prohibit. First, agencies could differ in terms
of whether they have official rules at all. Some
agencies may have no written rules, particularly
if they are very small (fewer than 10 police
officers), while others, particularly if they are
very large (with 500 sworn officers or more),
may have extremely lengthy official rules (e.g.,
standard operating procedures, official rules).
Second, when they do have the official rules
established, the behaviors those rules allow and
prohibit may differ substantially from agency to
agency. This is particularly the case for less
serious forms of misconduct such as mala
prohibita corrupt behavior (e.g., receipt of favors,
gratuities, small gifts, free meals, and discounts,
off-duty employment).
The problem of organizational rulemaking is
further complicated by the fact that in many
agencies, while an agency’s official policy
formally prohibits certain activities, the
agency’s unofficial policy tolerates them
(see, e.g., Knapp Commission 1972; Mollen
Commission 1994). The obligations of
rulemaking require police agencies not only to
3651
P
develop both formal and informal rules that
specify agency expectations of integrity, but
also to teach these rules and enforce them. As
Klockars and colleagues conclude (2006, p. 9),
“[i]n a police agency of integrity police officers
ought to know the agency’s integrity relevant
rules, understand the agency’s rationale for
them, and believe in the rightness of both.”
Detecting, Investigating, and Disciplining
Rule Violations – The second dimension of the
organizational theory of police integrity focuses
on both the creation and maintenance of activities
that permit detection, investigation, and discipline of misconduct (see, e.g., Klockars and
Kutnjak Ivković 2003; Klockars et al. 1997).
The activities are heterogeneous by nature.
They include not only reactive and proactive
investigations by police agencies, but also
inspections, audits, early-warning systems,
external reviews, reception of citizen complaints,
and integrity testing. The extent to which these
activities have been implemented and actually
used by a police agency varies substantially
across the agencies. According to Klockars and
colleagues (2006, p. 9), in a police agency of
integrity the occupational culture of the agency
will support the introduction and management of
the activities employed to detect, investigate, and
discipline rule-violating behavior.
Circumscribing “The Code” – The third
dimension of the organizational theory of police
integrity emphasizes the police agency’s obligation to circumscribe the code of silence (i.e., “the
blue curtain,” “the code”). The code of silence
refers to the norm of the police culture which
prohibits reporting of misconduct by fellow
police officers (see, e.g., Klockars and Kutnjak
Ivković 2003; Klockars et al. 1997). However,
although potentially viewed as a single norm that
could apply to all police officers, the reality is that
the code of silence could vary substantially. First,
exactly what behavior – in terms of its nature and
seriousness – is covered by the code varies
enormously across police agencies. The code of
silence in some agencies may cover only the least
serious forms of corruption (e.g., the acceptance
of gratuities and small gifts), while in others
it can cover even the most serious forms of
P
P
3652
corruption (e.g., thefts from crime scenes, the
acceptance of bribes from motorists caught
speeding). Second, to whom the benefit of the
code’s protection is extended could also vary
extensively. In some agencies, the code could
be limited to protect only misconduct by partners,
viewed a testimonial immunity that would be
a mirror image of traditionally privileged
relationships (husband and wife, physician and
patient, lawyer and client). In other agencies, the
code could protect misconduct not only by police
partners, but also by all police officers working
on the same shift, section, or unit. Klockars and
colleagues (2006, p. 9) write about the close
relation between the code of silence and integrity:
Many police administrators probably understand
that circumscribing both whom and what The
Code covers should be an administrative priority
. . . However virtually all police administrators
were line officers at some point in their careers,
and thus they have at least an appreciation, if not an
affection, for the bonds of collegial loyalty and
fraternal support that are part of the subculture of
policing. To the extent that circumscribing The
Code requires the weakening of those bonds of
loyalty and support, it is a task that not a few police
administrators approach with ambivalence.
A police agency of integrity is one in which the
occupational culture is intolerant of those who
abuse the rights and privileges of their office.
The Influence of Public Expectations and
Agency History on Police Integrity – The fourth
dimension of the organizational theory of police
integrity relates to the influence of the social
and political environment in which police
institutions, systems, and agencies operate
(see, e.g., Klockars and Kutnjak Ivković 2003;
Klockars et al. 1997). This theory argues that the
integrity of a police agency is affected by the
larger social and political environment in which
it operates (Klockars and Kutnjak Ivković 2003).
Although this understanding – the idea that the
police agency is affected by its environment – is
the underlying assumption of virtually all
historical studies of police, Reiss and
Bordua (1967; Reiss 1971) were the pioneers
who provided the first systematic exploration
of the topic.
When the society at large expects ethical
behavior of its officials, police agencies are also
Police Integrity
more likely to set high expectations and expect
ethical behavior from its employees. Not only do
the expectations of integrity differ enormously
across the world (see, e.g., Transparency
International Corruption Perceptions Index), but
they could also differ within the same country.
Using the United States as an example, Klockars
and colleagues (2006, p. 10) argue that some parts
of the country have long and virtually
uninterrupted traditions of persistent police
corruption (e.g., Chicago, New Orleans, Key
West), some areas have equally long traditions
of integrity (e.g., Milwaukee, Kansas City), and
other areas have undergone repeated
cycles of scandal and reform (e.g., New York,
Philadelphia, Oakland). In the end, Klockars and
colleagues conclude that “not only public
expectations about police integrity exert vastly
different pressures on police agencies in different
areas, but also police agencies of integrity may
effectively resist such pressures” (2006, p. 10).
Measuring the Extent of Police Integrity
Klockars and Kutnjak Ivković (2003) designed
a questionnaire that seeks to measure the extent
of police integrity in systematic, standardized,
and quantitative manner. The respondents were
provided with a letter asking them to assume that
the officer described in the scenarios had been
a police officer for 5 years, had a satisfactory
working record, and had not been disciplined in
the past. The first version of the questionnaire
contains 11 hypothetical scenarios focusing
mostly on corruption and ranging from those
merely giving an appearance of a conflict of
interest to those describing incidents of bribery
and theft (Klockars and Kutnjak Ivković 2004;
Klockars et al. 1997). The second version of
the questionnaire went beyond seeing police
integrity as the opposite of police corruption
and includes a range of police misconduct such
as the use of excessive force, failure to execute an
arrest warrant, and falsification of official record,
as well as police corruption. Each scenario is
followed by the same set of seven questions that
ask about police officers’ knowledge of official
rules, their opinion about the seriousness of
particular rule-violating behaviors, the discipline
Police Integrity
these behaviors should receive and would
actually receive, and their estimates of how
willing they would be to report such behavior.
The first version of the questionnaire has been
distributed across 30 police agencies in
the United States (Klockars et al. 1997, 2000),
characterized with a decentralized police. The
authors reported substantial differences across
the agencies. Although the sample of police agencies is not nationally representative, but
a convenience sample, it nevertheless illustrates
the point that the level of police integrity could
vary substantially across the police agencies.
Klockars and colleagues (2000) created
a ranking of police agencies based on their
levels of integrity. The authors (Klockars et al.
2006, p. 37) note:
In some police agencies in our sample police officers found nearly half of the behaviors described in
our sample to be of sufficient seriousness to merit
dismissal. In those agencies officers also claimed
that they and their police colleagues were highly
likely to report all but the least serious forms of
misconduct. In other agencies only theft from
a crime scene was, in the opinion of a plurality of
respondents, sufficient grounds to fire a police officer and not even that offence would motivate the
majority of officers in such agencies to break the
Code of Silence and report the misconduct of
a colleague.
A comparison of the results from the two large
municipal police agencies, one ranked at the top
and one at the bottom of the police integrity scale,
shows large differences in how serious police
officers perceive misconduct, what discipline
they expect and approve, and how wiling
they seem to be to report the fellow officers’
misconduct. With one exception, police officers
from the agency at the bottom of the police
integrity scale would not expect the police officer
to be dismissed for any of the behaviors described
in the questionnaire, while the police officers
from the agency at the top of the police integrity
scale expected dismissal in four most serious
cases. In addition, Klockars and colleagues
(Klockars et al. 2006) found stark contrasts in
the extent of the code of silence between the
two agencies: in the agency ranked at the top of
the integrity scale, the majority of police officers
3653
P
reported that their colleagues would report their
fellow officer who engaged in the behavior
described in all cases. On the other hand, the
majority of police officers from the agency
ranked at the bottom of police integrity scale
did not expect their fellow officers to report
misconduct in any of the described cases.
However, as the subsequent study of three
large municipal police agencies (identified to be
in the top third of the police integrity scale)
demonstrates, police agencies can be quite
diverse in the ways they create and maintain
their integrity (Klockars et al. 2006). The authors
studied each of these three agencies – Charleston,
South Carolina; Charlotte-Mecklenburg, North
Carolina; and St. Petersburg, Florida – for two
years and have developed their integrity profiles.
Since 1996, nearly two dozen studies have
used this first police integrity survey. Three
studies (Micucci and Gomme 2009; Burbach
Raines 2009; Marche 2009) analyzed the
30-agency data collected by Klockars and
colleagues,
and
another
four
studies
(Chappell and Piquero 2004; Hickman, et al.
2004; Schafer and Martinelli 2008, Gottschalk
2009) used the police integrity questionnaire as
the basis for their surveys of other US police
agencies. In addition to the US agencies, the
same questionnaire has been used in nearly 20
countries across the world, including Austria,
Bosnia and Herzegovina, Canada, Croatia, the
Czech Republic, Finland, Hungary, Japan,
the Netherlands, Pakistan, Poland, Slovenia,
South Africa, Sweden, and the UK
(see Klockars et al. 2004).
Although the list of the countries participating
in the survey is far from exhaustive, the findings
clearly show that the levels of police integrity
differ substantially and significantly, analogously
to the tremendous variation in the levels of police
integrity across the 30 US police agencies.
The nature and extent of police integrity
measured in the countries such as Finland and
Sweden is quite different from the nature
and extent of the police integrity measured in
countries such as South Africa and Pakistan
(see Klockars et al. 2004). The differences are
visible not only in the police officers’ perceptions
P
P
3654
of seriousness and willingness to report, but
also in their views of appropriate and expected
discipline (Klockars et al. 2004, p. 13):
It appears that in each country the seriousness of
officers’ misconduct is, in large part, determined by
the absolute level of discipline the organization is
expected to visit on the offending officer. In almost
every case, when the police organization is
expected to punish a offense very severely, officers
regard that offense as serious. Conversely, when
organizations do not punish misbehavior severely,
as is the case in Hungary, Pakistan, and South
Africa, officers seem to have little ability to distinguish among the levels of seriousness with regard
to misconduct. Perhaps the most dramatic finding
that emerges from examining the contours of integrity concerns the worldwide prevalence of the code
of silence.
The authors continue to note that, in five out of
14 countries included in the book “Contours of
Police Integrity,” the code of silence would cover
misconduct described in every scenario
(Klockars et al. 2004, p. 17). In addition, the
code of silence would protect behaviors such as
the acceptance of a bribe from a person caught
speeding – a violation of the penal code – in nine
out of 14 countries.
The heterogeneity of police integrity is
influenced by a host of reasons; a police agency’s
local social, political, economic, and legal
environments influence its level of police
integrity. Research indicates that countries that
belong to the same category along one dimension
(e.g., economic development, geographic
location) still have different levels of police
integrity. This is the case for both developed
democracies (e.g., the USA and the UK)
and countries in transition (e.g., Croatia and
Hungary), as is the case for countries from
Northern America (e.g., the USA and Canada)
and Europe (e.g., Croatia and the Netherlands).
The second version of the questionnaire,
exploring the resistance to various sources of
temptations (not only police corruption), has
been developed as well. The initial work by
Klockars and colleagues explored the extent of
police integrity in three US police agencies using
both versions of the questionnaire (see Klockars
et al. 2006). The comparative efforts are underway, with the first results coming from Croatia
Police Integrity
(Kutnjak Ivković 2009), South Korea (Kutnjak
Ivković and Kang 2011), and South Africa
(Kutnjak Ivković and Sauerman 2012).
Other Research on Police Integrity
The issues of police integrity and police
misconduct are closely interconnected. Please see
separate entries addressing different forms of
police misconduct (e.g., “police corruption,” “use
of excessive force,” “police lying”). A few writings
explicitly connect police misconduct (or any of its
specific forms) with police integrity. For example,
Garner and colleagues (2004) studied the patterns
in the police use of force as a way of measuring
police integrity. The authors (Garner et al. 2004,
p. 6.119) conclude that, although the average extent
of force used by the police during a typical arrest is
about equal across racial categories (and thus
should be taken as an indication of police integrity),
the extent of force used by the police during
a typical arrest of male suspects is larger than the
quantity of force used during a typical arrest of
female suspects (and thus should be taken as an
indication of the lack of police integrity).
In 2001, the US Department of Justice
published Principles of Promoting Police
Integrity. The publication, available in print and
on the Internet, lists the “best practices” for
promoting integrity. The first part of the
publication explores the use of force, complaint
and misconduct investigations, training, recruitment, hiring, and retention, as well as general
principles of promoting accountability and
effective police management. The second part of
the publication contains examples of promising
police practices and policies, as well as the
research projects on police integrity funded by
the US Department of Justice. Many of the practices that should be used to enhance integrity (e.g.,
complaint misconduct investigations, the early
warning systems, citizen reviews), at the same
time, are mechanisms of accountability. Publications like that, focusing on the specific mechanisms or practices that should be used to enhance
integrity, have significantly grown in numbers
since the late 1990s. A separate body of research
focuses on these mechanisms. Please see the entry
“Control of Police Misconduct.”
Police Integrity
Related Entries
▶ Control of Police Misconduct
▶ History of Police Unions
▶ Police and the Excessive Use of Force
▶ Police Corruption
▶ Police Culture
▶ Police Discretion and Its Control
▶ Police Lying and Deception
Recommended Reading and References
Burbach Raines J (2009) Ethics in policing: misconduct
and integrity. Jones and Barlett, Sudbury
Chappell AT, Piquero AR (2004) Applying social learning
theory to police misconduct. Deviant Behav
25:89–108
Gaffigan SJ, McDonald PP (eds) (1997) Police integrity:
public service with honor. U.S. Department of Justice,
Washington,
DC.
http://www.ncjrs.org/pdffiles/
163811.pdf. Accessed 8 Apr 2006
Garner JH, Maxwell CD, Heraux C (2004) Patterns of
police use of force as a measure of police integrity.
In: Hickman M, Piquero AR, Greene JR (eds) Police
integrity and ethics. Wadsworth/Thomson Learning,
Belmont, p vii
Goldstein H (1975) Police corruption: a perspective on
its nature and control. The Police Foundation,
Washington, DC
Gottschalk P (2009) Crime-based survey instrument for
police integrity measurement. Policing Int J Police
Strateg Manag 33(1):52–68
Greenberg SF (1997) Executive summary. In: Gaffigan SJ,
McDonald PP (eds) Police integrity: public service
with honor. U.S. Department of Justice, Washington,
DC. http://www.ncjrs.org/pdffiles/163811.pdf, p. v.
Accessed 8 Apr 2006
Hickman M, Piquero AR, Greene JR (eds) (2004) Police
integrity and ethics. Wadsworth/Thomson Learning,
Belmont
Huberts LWJC, van den Heuvel JHJ (eds) (1999) Integrity
at the public-private interface. Shaker Publishing B.V,
Maastricht
Klockars, C. (1995). A Theory of Excessive Force and Its
Control. In William A. Geller & Hans Toch (Eds.),
And Justice for All. Washington, D.C.: Police Executive Research Forum, pp. 11-29.
Klockars CB, Kutnjak Ivković S (2003) Measuring police
integrity. In: Hickman MJ, Piquero AR, Greene JR
(eds) Police integrity and ethics. Wadsworth, Belmont
Klockars CB, Kutnjak Ivković S, Harver WE,
Haberfeld MR (1997) The measurement of police
integrity. Final Report Submitted to the U.S. Department of Justice, Office of Justice Programs, National
Institute of Justice
3655
P
Klockars CB, Kutnjak Ivković S, Harver WE, Haberfeld
MR (2000) The Measurement of Police Integrity.
Research in Brief. U.S. Department of Justice, Office
of Justice Programs, National Institute of Justice.
Government Printing Office, Washington, DC
Klockars CB, Kutnjak Ivković S, Haberfeld MR (eds)
(2004) The contours of police integrity. Sage,
Newbury Park
Klockars CB, Kutnjak Ivković S, Haberfeld MR (2006)
Enhancing police integrity. Springer, New York
Kutnjak Ivković S (2009) The Croatian police, police
integrity, and transition toward Democratic policing.
Policing Int J Police Strateg Manag 32(3):459–488
Kutnjak Ivković S, Kang W (2011) Police integrity in
South Korea. Policing Int J Police Strateg Manag
35(1):76–103
Kutnjak Ivković S, Sauerman A (2012) Curtailing the
code of silence among the South African Police.
Policing: Int J Police Strateg Manag (forthcoming)
Manning PK, Redlinger L (1993) The invitational edges of
police construction. In: Klockars C, Mastrofski S (eds)
Thinking about police. McGraw Hill, New York,
pp 398–412
Marche GE (2009) Integrity, culture, and scale: an
empirical test of the big bad police agency. Crime
Law Soc Change 51:463–486
Micucci AJ, Gomme IM (2009) American police and
subcultural support for the use of excessive force.
J Crim Justice 33:487–500
Moore M (1997) Epilogue, police integrity. In: Public
service with honor. U.S. Department of Justice.
https://www.ncjrs.gov/pdffiles/163811.pdf. Retrieved
on Oct 10 2012
Pagon M (ed) (2000) Policing in central and Eastern
Europe: ethics, integrity, and human rights. College
of Police and Security Studies, Ljubljana
Pope J (2000) Confronting corruption. The elements
of a national integrity system (TI Source
book 2000). Transparency International, Berlin.
http://www.transparency.org. Accessed 8 Apr 2006
Reiss AJ Jr (1971) The police and the public. Yale
University Press, New Haven
Reiss AJ Jr, Bordua DJ (1967) Environment and
organization: a perspective on the police. In: Bordua
D (ed) The police: six sociological essays. Wiley,
New York
Schafer JA, Martinelli TJ (2008) First-line supervisor’s
perceptions of police integrity. Policing Int J Police
Strateg Manag 31(2):306–323
Transparency International (2001) The national integrity
system. Concept and practice. A report by
Transparency International (TI) for the global forum
II on fighting corruption and safeguarding integrity.
Transparency International, Berlin
U.S. Department of Justice (2001) Principles for promoting police integrity; examples of promising police
practices and policies. U.S. Department of Justice,
Washington, DC. http://www.ncjrs.gov/pdffiles1/ojp/
186189.pdf. Retrieved on 8 Apr 2006
P
P
3656
Vicchio SJ (1997) Ethics and police integrity: some
definitions and questions for study. In: Gaffigan SJ,
McDonald PP (eds) Police integrity: public
service with honor. U.S. Department of Justice,
Washington, DC
Walker S (2004) Forward. In: Hickman M, Piquero AR,
Greene JR (eds) Police integrity and ethics.
Wadsworth/Thomson Learning, Belmont, p vii
Police Interrogation
▶ Law of Police Interrogation
Police Interrogation
general theories of leadership and organizational
behavior might be used to explain aspects of
police organizations (Adlam and Villiers 2003).
This is often the approach taken in textbooks and
professional literature. The second approach
examines the styles and methods supervisors use
in seeking to influence subordinate personnel
(Engel 2001). The entry begins by briefly
reviewing the dominant implications arising
from these two traditions. It continues with an
examination of what research evidence suggests
about the ability of police leaders to influence
unit productivity and output. The entry concludes
by considering how different generations of
police personnel and leaders might respond to
different styles of leadership.
Police Leadership Styles
Joseph A. Schafer
Department of Criminology and Criminal
Justice, Southern Illinois University, Carbondale,
IL, USA
Overview
Leadership is of central importance to the operation of effective, efficient, and equitable police
organizations, yet it remains an elusive concept.
It is clear that good leadership is vital to ensure
that an agency operates in a manner that preserves public legitimacy and trust. Among the
compelling evidence in support of this assertion
is the fact that a crisis of leadership is often
associated with major scandals in police agencies. Despite recognizing that strong leadership
is of vital importance, scholars and police organizations have struggled to measure and understand a number of key questions, including what
leadership styles are commonly used by police
supervisors, when and how supervisors can influence subordinate performance and output, and
what distinguishes between effective and less
effective leaders.
This entry considers the topic of police leadership styles. Literature considering leadership in
police organizations can be separated into two
major approaches. First, some write about how
Introduction
At the onset it is important to distinguish between
police management and police leadership. Police
supervisors are often interchangeably referred to
as managers and leaders. Management is the control of routine processes and the maintenance of
the status quo. Leadership emphasizes changes
that will improve processes, personnel, and organizations. Effective supervisors tend to excel at
both management and leadership, because both
skill sets are important. However, they are not of
the same quality, and policing has tended to
emphasize management far more than leadership
to its own detriment. It should also be noted that
leadership is a behavior that can be independent
of formal rank or authority. Some very prominent
and influential leaders in police organizations are
front-line personnel. Much of what has been written about police “leadership” is actually describing aspects of managing, administering, and
supervising police organizations and personnel.
The emphasis of this entry is on true leadership, not what is often described in policing literature (including academic scholarship, which has
tended to confuse management and leadership).
Leaders innovate, take calculated (and often necessarily) risks, encourage and develop personnel,
and are willing to challenge tradition and status
quo. This does not imply that leaders blindly
Police Leadership Styles
challenge anything, everything, and everyone in
the organization. Rather, they have the courage to
confront the pointless, mindless, counterproductive aspects of their organizations and its practices. This tendency is sorely lacking in many
aspects of the modern public sector, particularly
policing.
Though much has been written about leadership
in broader settings, especially in corporate, military, political, and athletic contexts, policing
scholars have given limited empirical consideration to matters of leaders and leadership. The
traditional perspective found in police literature
tends to cast leaders as using a narrow range of
directive and controlling styles; as a consequence,
supervisor-subordinate relationships are characterized by tension and animosity (Rowe 2006; Van
Maanen 1984). More recently, research evidence
has suggested an alternative perspective. Leaders
have increasingly been framed as using a range of
styles and approaches when interacting with those
they seek to influence. These clusters of leadership
approaches go beyond the traditional, authoritarian, and autocratic supervisory approaches that
once dominated policing (Kuykendall and
Unsinger 1982). It is quite likely this change
reflects a very real transformation in leadership
styles observed in policing.
Major Theories Relevant to Police
Leadership
A wide array of theories has been offered to
describe and explain the behavior of individuals
and organizations. Some of the earliest studies of
leaders and leadership tended to focus on the
“great man/great woman” tradition. Wellregarded leaders were studied in a biographical
format in the hope of deriving an understanding
of the behaviors and actions that contributed to
their perceived success. Over time, studies
expanded this tradition by examining samples of
recognized leaders. Modern research continues to
seek an understanding of the casual links between
leader traits and subsequent efficacy; what characteristics, habits, and behaviors differentiate
leaders who are most effective from their peers?
3657
P
Because early studies were concerned with the
traits and habits of leaders, they tended to pay less
attention to followers. The latter were often seen
as liabilities and risks that organizations and
supervisors needed to manage through control,
policy, procedure, reporting requirements, and
similar of mechanisms. This was based on what
Douglas McGregor (1960) referred to as Theory
X assumptions, and it led to a very distinctive,
controlling, directive form of management. In
contrast,
McGregor
described
Theory
Y approaches that were predicated on the belief
that people can achieve good work outcomes
when they are given the proper motivation and
environment. As a result, Theory Y management
approaches emphasize creating the optimal environment in which employees can be engaged and
empowered, rather than seeking to regulate and
constrain their choices and behavior.
James Burns observed a similar duality in the
nature of leader-follower interactions, writing
about “transactional” and “transformational”
approaches to leadership (1978). Transactional
leaders were seen as emphasizing exchanges
and quid pro quo arrangements with the personnel they supervised. Followers were provided
with an understanding of what they were to
accomplish and how it was to be achieved; they
also understood what they would receive for
adhering to those expectations and the consequences of violating accepted practices. The
role of leaders in such arrangements was not to
provide leadership; in actuality, they were there
to monitor and manage followers, grant rewards
when earned, and issue discipline when needed.
Followers were given “carrots” (rewards) for
good behavior and faced the risks of “sticks”
(punishments) if they did not produce the
expected outcomes in the expected way.
In contrast, transformational leadership was
characterized as a process of seeking to improve
organizations, operations, and personnel. Leaders
truly lead rather than just managing and
maintaining the status quo. To accomplish this
outcome, transformational leaders sought to
inspire employees to embrace a shared vision
and empowered them to use discretion and creativity in the pursuit of that outcome. Followers
P
P
3658
were given the freedom to make choices, the
authority to develop new ideas, and the independence to identify and implement ways to pursue
the leader’s vision. In other words, followers
were viewed as being smart enough and creative
enough to be entrusted to solve problems, pursue
new ideas, and act in a way that was consistent
with the vision of the leader. This was done
without requiring followers to engage in burdensome reporting requirements, bureaucratic permission seeking, and mindless compliance
checks.
Though not universally embraced, there is also
a sense in more contemporary perspectives that
the models, styles, and traits that generate leadership efficacy are situational. This idea was first
expressed by Tannenbaum and Schmidt (1958) in
writing about management in the corporate
world. They arrayed leadership approaches on
a continuum from “boss centered” to “subordinate centered,” noting that a major challenge for
(then) modern managers was deciding where to
fall on the continuum in a given situation. In
policing contexts, this suggests that “what
works” for a leader is rarely universal, instead
varying based on context, culture, position,
objectives, the environment, the strengths and
predispositions of a leader, and the preferences
and characteristics of those the leader seeks to
influence. An officer or leader who is ineffective
in one agency or assignment might perform quite
well in a different set of circumstances or vice
versa. It has also been argued that leadership is
not the same at all levels of an organization. The
behaviors that make a leader an effective patrol
sergeant might not yield the same outcomes for
the leader when promoted to serve as a lieutenant
overseeing the records division. Leaders need to
be conscious of the fact that the approaches used
in a given situation and/or to achieve a given end
might not yield universal success.
Leadership is a challenging process, but it can
be the element that distinguishes organizations
that are “great” from those that are merely
“good.” In addition, leadership is a behavior,
not a position within an organization. Leaders
are people who demonstrate certain skills and
habits, regardless of their formal rank or office.
Police Leadership Styles
Too often it is presumed that all supervisors are
leaders and that only supervisors can lead.
Increasingly there is awareness and celebration
of the idea that every officer must be a leader, at
least in some contexts. Informal leadership is
very difficult to measure in any type of organization, but it can have a profound influence on
personnel and operations (Schafer 2001). Though
the influence of informal police leaders is generally evident in policing scholarship, research has
not been able to empirically define when and how
it matters.
Theoretical perspectives and explanations
regarding leadership extend far beyond the brief
ideas reviewed in this passage. What is of relevance for this discussion are the ideas that leadership is situation but that modern perspectives
tend to emphasize involving employees in the
decision-making process. It is sometimes
observed that leadership is something done with,
not to, others. Further, it should be recognized
that good leadership is actually quite difficult to
achieve. Effective leaders continually seek ways
to improve their performance as they confront
new and different challenges (Schafer 2012).
Leadership Styles Displayed in Policing
Despite the presumed importance and influence
of informal leaders in policing context, studies of
police leadership have focused on studying formal supervisors. As a consequence, what is
known about leadership styles and their influence
(or lack thereof) is based on studies of those who
have been awarded a rank. Such supervisors may
or may not demonstrate actual leadership behaviors as they carry out their duties. Far less is
known about leaders in police organizations,
who may or may not hold an official rank in
their agency (Van Maanen 1983). This section
reviews literature seeking to describe the management and leaders styles police supervisors
have displayed in prior studies. There is
a distinction between what scholars found when
they first began to study police supervisors and
what is more often seen in contemporary
agencies.
Police Leadership Styles
Early efforts to study the management styles
of police supervisors did not paint a favorable
picture of the level of actual leadership taking
place in organizations. Kuykendall and Unsinger
(1982) administered a leadership inventory to
police managers. Their results suggested that
most respondents used a variety of styles in the
course of performing their supervisory duties.
The most commonly reported management styles
were selling, telling, and participation. There was
little evidence that supervisors used actual leadership or other forms delegation as a component
of their managerial style, though their data was
collected in the late 1970s before these types of
approaches had entered the discourse of police
management. In other words, managers reported
a preference to use styles that were oriented more
toward “safe” methods emphasizing control and
direction of personnel, rather than responses that
might be considered more “risky” styles, such as
delegating important tasks to subordinate
personnel.
Mayo (1985) contended supervisors tended
not to trust the skills, loyalty, and judgment
of personnel in their agencies. This resulted in
the observed practice of police supervisors
engaging in micromanagement. Girodo (1998)
surveyed police executives from around the
world and found most reported leaning toward
what might be considered “Machiavellian”
approaches. End objectives were achieved
not through democratic and participative
approaches but through the manipulation of subordinate personnel. This is reflective of the tendency for police agencies to seek to operate as
paramilitary organizations that exert top-down
control over personnel, communication, and
decision making. These traditional styles of
police supervision have been subject to extensive criticism because they do not encourage
employees to be committed, productive, and
creative members of the workforce.
Allen (1982) found the type of supervision
(i.e., the use of incentives) was likely more influential than the volume of supervision. Quality
was more important than quantity (more contact),
though his measures were quite limited. It should
also be noted his research was conducted on
3659
P
a 1970s group of front-line police personnel.
The education, traits, experiences, and expectations of modern patrol officers are likely quite
different than what was seen in earlier eras.
Johnson (2006) studied supervisor influence
on traffic and drunk driving enforcement in
a collection of agencies in the Cincinnati metropolitan region. His results support that leadership
matters, both in terms of the priorities of agency
executives and also the priorities of patrol supervisors. Though leadership was deemed important, it was not the only factor influencing
enforcement productivity. Organizational factors
such as policies, supervision, and bureaucratic
requirements do influence when and how officers
enforce the law but so do situational factors,
officer preferences, and environmental considerations. In other words, even when leadership
matters, it is not the only force or factor influencing officers and the decisions they make.
Engel (2001) developed four broad categorizations of supervisory styles that help frame an
understanding of different leadership approaches
in policing. “Traditional” leaders were highly
task oriented and expected aggressive enforcement of the law from subordinates; they were
focused on traditional policing outcomes and presumably would have leaned toward transactional
relationships with followers (e.g., officers making a lot of arrests and issue a lot of citations
would presumably be ranked higher by traditional leaders). “Innovative” leaders were situated further onto the transactional end of the
leadership continuum; they shared power with
subordinates, sought to have friendly and productive relationships with subordinates, and emphasized community relations. “Supportive”
supervisors sought to be a buffer between frontline personnel and top agency leaders; they place
less emphasis on accomplishing traditional policing tasks. Finally, “active” supervisors were
themselves highly active in front-line policing
matters; they had positive views of their subordinates, tended to exemplify the idea of “leading by
example,” and did not see themselves as radically
different than those they supervised. Active
supervisors demonstrated the most influenced in
the two agencies included in Engel’s study.
P
P
3660
Beginning in the 1980s, there has been
a greater emphasis on democratic and participative approaches to police supervision and leadership (Wycoff and Skogan 1994), though limited
advancements have been realized. The promise
of such approaches is multifaceted. A leader who
engages subordinates in making decisions can tap
into the creativity, experience, and intelligence of
employees. It is expected that participative
approaches can make organizations more adaptable and responsive in the face of shifting and
dynamic social environments. Leaders who use
these types of approaches do not make all choices
unilaterally. They engage subordinates to identify problems, determine priorities, and derive
solutions. This does not suggest that all choices
and actions are subject to full participation by all
employees. Certainly there are aspects of police
operations that require command-and-control
authority and decision making, but such circumstances are the exception, not the rule. Though
there is limited evidence assessing these types of
practices, experiences to date suggest promising
implications (Steinheider and Wuestewald 2008).
Survey data from police officers finds personnel often express a preference for supportive and
participatory leadership styles (Witte et al. 1990).
Officers do not like directive and controlling
management styles. Instead, they prefer to be
led in a process that grants them a voice and
input into organizational decisions. This does
not suggest that officers do not understand and
support the needs of leaders to have ultimate
authority. Rather, it implies that, when possible,
officers prefer to be able to participate in organizational decision making in an open and democratic manner. Tentative evidence suggests that
many police executives and leaders are similarly
open-minded in employing nontraditional systems (Steinheider and Wuestewald 2008;
Tannenbaum and Schmidt 1958). It is less clear,
however, whether those expressions of support
translate into actual changes in supervisory style
as perceived by employees.
Consideration of supervisory styles in policing
should not overlook the role of poor or ineffective
leadership. Barbara Kellerman (2004) offered
one of the best considerations of this topic in
Police Leadership Styles
leadership literature. She argued that too often,
professional and scholarly discussions of leadership frame leader behavior as benevolent and
positive behavior. She and others have argued
consideration of leadership needs to acknowledge leaders sometimes use less-than-ideal
methods. Some supervisors achieve their objectives while using poor leadership techniques.
Some leaders are quite successful but pursue
ignoble or evil objectives (Adolf Hitler is the
common example of this outcome). People who
are generally very good and effective leaders
occasionally make mistakes. Leaders who have
been successful sometimes “derail” (McCall and
Lombardo 1983). Consideration of police leadership styles should not overlook how poor and
ineffective approaches can hamper the ability of
leaders to achieve their objectives.
The Influence of Leadership
It is generally accepted that police supervisors
(who are formally in positions of leadership)
can shape various organizational and individual
outcomes produced by employees (Engel 2001).
Supervisors are thought to have the ability to
influence the tone of police operations, helping
shape the style, tactics, and outcomes. In reality,
it is far easier to believe that supervisors have
such a strong and direct influence on officers and
the organizations than it is to prove that belief is
reality. This is partially a function of the limited
amount of quality research studying supervisory
and leadership influences in policing. It is also
a reflection of the nature of the police organization itself.
In many organizational contexts, supervisors
have routine and direct contact with their subordinates. The task environment is also such
that employees often have limited discretion,
established accountability mechanisms, and routine reporting demands that serve to structure,
regulate, and formalize the decisions they make.
Police departments are often characterized as
a quasi-military bureaucracies characterized by
command-and-control relationships between
supervisors and subordinates. This tends to
Police Leadership Styles
support the use of directive supervisory styles,
which are predicated on the ability of a manager
to closely monitor employee performance.
In reality, officers are frequently separated
from their supervisors by both time and space.
The nature of police work also means that officers
enjoy a wide degree of discretion in performing
their duties, particularly the low seriousness
events that typify much of police work (Brown
1988; Van Maanen 1983). Additionally, police
operations and personnel are influenced by cultural and external forces that serve to condition
where, when, how, and to what end police decisions are made. The result is that it has been
suggested police agencies are “loosely coupled
organizations” (Maguire and Katz 2002). Police
managers do not have the time or capacity to
effectively monitor the actions of officers on the
street. The unpredictable, dynamic, and complex
nature of police work also makes it quite difficult
for agencies to craft effective policies to govern
and guide how officers make decisions in the
field. As a result, police personnel are not nearly
as restricted by administrative impositions.
On paper and in textbooks, agencies are
orderly and control is clear; reality is quite different. This makes leadership (particularly transformational and democratic approaches) all the
more important. Directive, controlling, manipulative, and micromanaging approaches actually
function to push officers away from organizational objectives and discourage them from
supporting the mission and vision of executives.
Real leadership (as opposed to management) can
help engage, empower, and motivate personnel to
support and pursue organizational objectives.
Because officers are often performing their duties
with no direct supervision and little reliable evidence verifying when, where, and how they perform their duties, having engaged personnel is all
the more important. Creating work environments
that treat officers in accordance with McGregor’s
(1960) Theory Y assumptions would, in the
aggregate, be expected to produce better outcomes and results for the organization.
As a result of these factors, it is reasonable to
expect limitations on the ability of police leadership to influence police personnel. Many of these
3661
P
challenges are far more acute in policing than in
other occupational and organizational settings.
Police personnel operate in environments with
limited direct supervision, handle a diverse
range of tasks, and make myriad high-discretion
decisions (Allen 1982; Engel 2001; Van Maanen
1983, 1984). Though supervisors and
organizations seek to impose policies, protocols,
structures, and mandates to condition and
direct discretionary behavior, officers enjoy
a considerable degree of latitude in performing
their duties. This totality of circumstances suggests that leaders will have a limited, but not
nonexistent, ability to influence officers. Police
supervisors and leaders are able to exert at least
a limited degree of influence on personnel.
Much of the research studying supervisory
influence has focused on examining traditional
policing outcomes, such as enforcement behavior, use of force, and officer misbehavior (Brown
1988; Huberts et al. 2007). Though such considerations are important, they reflect but a narrow
range of the tasks to which officers attend and the
expectations the public has of their police.
Equally important, they tend to focus on the
ends, rather than the means, of policing. A small
number of studies have attempted to examine
issues such as the nature of police encounters
with the public and self-initiated efforts officer
undertakes to address specific problems. Taken as
a whole, researchers have achieved mixed assessments of whether supervisors can influence the
behavior of subordinate personnel (cf, Allen
1982; Brown 1988; National Research Council
2004). It should be noted that many of these
studies were limited in the types of influence
they were studying and the methods that were
used to measure specified outcomes.
It has conventionally been assumed that bringing about change in police organizations is akin
to “bending granite” (Rowe 2006). In recent
years, increasing experience and empirical evidence have begun to suggest this may not always
be the case. Jermier and Berkes (1979) studied
officers in a Midwestern agency, finding that
officers were supportive of leadership approaches
that emphasized participation. Officers reported
they were less inclined to respond to directive and
P
P
3662
controlling management approaches. Case studies of community policing have highlighted the
role of leadership as a mechanism to facilitate
rapid (and at times, radical) organizational transformation (Wycoff and Skogan 1994). Though
critics have dismissed the roles of William
Bratton and COMPSTAT as causal forces
influencing New York City’s crime decline,
there can be little doubt that his leadership style
in NYPD (and elsewhere) has brought about
rapid change in climate and culture. Thus, leadership can change police culture, though it is still
difficult to determine how and why that outcome
is achieved in some instances and not in others.
Several studies based on data collected for the
Project on Policing Neighborhoods (POPN) tend
to offer some provisional evidence that supervisors can influence front-line personnel (Engel
2001). Importantly, these studies attempted to
assess not just the general idea of supervisory
influence. Instead, supervisors were categorized
based on their general management style. Findings based on the POPN data offer a possible
explanation for the mixed results achieved in
prior studies by suggesting that variation in
supervisory influence is partially a product of
dissimilarity in supervisory styles. Not all supervisors use the same approaches in seeking to
control, influence, and lead. The methods and
styles used by supervisors vary based on the
circumstances, timing, and audience, as well as
the supervisor’s perceptions of various situational exigencies.
Far less is known about the measurement of
leadership efficacy beyond considerations of officer performance, output, and conduct. What other
metrics might be used and who should provide
input on the performance of a given leader?
These efforts can be challenging, as they require
the development of broader definitions and measures of leader efficacy, which tends to invoke
a number of methodological difficulties. Huberts
et al. (2007) studied Dutch police officers. They
found leadership approaches did influence integrity violations by officers. In particular their
results suggested that officers were influenced
by strong role modeling by their leaders. When
leaders took a strong stance against integrity
Police Leadership Styles
violations and were perceived to perform their
duties with a high moral standard, officers were
more inclined to emulate that approach. This is
certainly a favorable representation of what
might be achieved through leadership, at least
based on some important outcomes of relevance
to policing.
As a result of this body of literature, several
important observations and conclusions can be
made about leadership and influence in police
organizations. Supervisors can influence frontline personnel, at least in the production of
some outcomes using some management styles.
Poor leadership practices can also generate real
and negative consequences in the workplace,
including poor productivity, dissatisfaction,
stress, attrition, and absenteeism, among other
concerns (Kelloway et al. 2005). Though traditional police leadership approaches were primarily transactional (stick and carrot) approaches,
transformational strategies seem more promising
in their ability to favorably influence personnel
and behavior. There is evidence that followers
might prefer such participatory approaches and
that leaders are open to their use. To date, however, research has not been able to conclusively
contrast the outcomes of transactional and transformational leadership approaches in policing.
Generational Effects
There may be generational and experiential influences that shape the approaches used by supervisors. In the 1960s and 1970s, there was a renewed
emphasis on college education for police personnel. Pursley (1974) compared police chiefs based
on education, experience, and involvement in
professional activities (belonging to national professional associations, publishing in professional
periodicals); he analyzed their responses to
a variety of leadership inventories. He found
that educated chiefs had less experience (both as
officers and chiefs, suggesting they were also
likely to be younger), were more connected with
professional groups and activities, and were more
receptive to more transformational leadership
approaches (though he did not use this term).
Police Leadership Styles
Younger leaders were more willing to support
delegating authority within the workforce,
allowing subordinates to participate in decision
making, embracing more democratic approaches,
and supporting less authoritarian and controlling
leadership approaches. Thus, when considering
how generations influence policing and police
leadership, it is not just a matter of different
mentalities across groups of different officers.
Supervisors from different generations might
also have different ideas about how to lead
personnel.
Cohen (1980) studied command personnel
(captains and above) in the New York City Police
Department based on data collected in 1972. He
found two primary management styles. Tradition-oriented leaders generally had no education
beyond high school. Reform-oriented leaders
generally had attended or were attending college.
Tradition-oriented leaders were focused on
power and authority, while reform-oriented
leaders were more participative and emphasized
community relations and service. Cohen found
that reform-oriented leaders had received lower
ratings on internal performance assessments, but
he attributed this to the probability that traditionoriented leaders were evaluating most reformoriented leaders. Thus, it is would not be surprising to see that the disjuncture in their orientations
toward leadership and the broader role of the
police might result in less positive performance
evaluations.
This holds important implications for understanding leadership. First, it reinforces that leadership is situational; at the very least “good”
leadership is a time-bound concept. Cohen’s
work was being done during a time of transition
and reform in NYPD. Second, leadership practices and police personnel might both experience
different generations. Several factors contribute
to this situation. In general, different generations
are distinct from one another based on variation
in life experiences, culture, and broader society
during their formative years. During eras of transition in policing (the movement out of the traditional model in the 1970s, the proliferation of
community policing in the 1980s and 1900s, the
emergence of various data-driven and evidence-
3663
P
based policing practices in the 2000s), there is
likely to be tension throughout all ranks of police
agencies. Personnel who started their career one
(or two) ideological generations earlier might be
less inclined to adapt their beliefs, values, and
practices. Making matters more complicated,
the role of leadership in periods of reform and
transition might be a function both of a leader’s
generation and position within the organization.
Top executives likely are driving forces behind
implementing these reforms and innovations, so
their support might be expected. In contrast, the
drag or tension may be among mid-level supervisors. Sergeants, lieutenants, captains, and
others who feel no ownership over the new idea
have little reason to embrace new perspectives.
They may resist the change or evenly actively
obstruct its emergence because they view the
change as a threat to both their perspective on
leadership and their continued advancement in
the organization.
Andreescu and Vito (2010) found contemporary managers expressed strong support for ideals
consistent with transformational leadership
approaches. These included inclusive and
human-oriented styles of leadership, the articulation of an organizational vision, caring for the
well-being of employees, and setting an example
for employees. Transformational approaches
remain a challenge for police organizations.
Though support for this style of leadership is
apparent among police personnel and leaders
(Witte et al. 1990), there is also a belief that
such circumstances as participatory management
are not actually found in most organizations.
Stamper (1992) found a disjuncture between
how executives perceived themselves and how
members of their executive staff perceived
them. Executives believed they lead their organization in one fashion. This self-perception tended
to differ from how those around the executives
perceived their leadership style and influence.
Intuitively it makes sense that younger officers might be more likely to support transformational, participatory, and democratic approaches
to leadership. These youth are more oriented
toward social interaction, social involvement,
and being allowed a voice in aspects of
P
P
3664
organizational operations. Scholars and professionals continue to speculate that there might be
a generational preference for transformational
leadership approaches, but there is still an
absence of research evidence assessing this
matter in policing contexts. As the labor force
becomes increasingly filled with those from generations desiring more transformational leadership, it will be increasingly important that police
leadership adapt to reflect that situation. If organizations are to maintain long-term viability, it
will be essential to create work environments
that attract and retain newer generations of
officers.
Conclusions
Leadership remains a vital issue in modern police
organizations. Its importance cannot be refuted,
and the need for leadership (not simply strong
management) is quite apparent within the profession. What remains more elusive is a clear understanding of when and how leadership matters as
a force influencing the values, beliefs, attitudes,
and behaviors of police personnel. Though there
is a natural appeal toward transformational leadership styles, the potential benefits of these
approaches remain presumed rather than proven
in police organizations. Issues of leadership will
most likely continue to become increasingly
important in the future as police organizations
need to be more adaptive in responding to everchanging criminal threats and community expectations. This will be compounded by work forces
increasingly composed of officers from more
recent generations, who may well require a very
different style of leadership and a very different
organizational environment. If police organizations are going to remain viable in confronting
crime and disorder, real leadership will be of
increasing importance. Both the research community and the profession itself must support
the development of greater knowledge of the
role leadership will play in these dynamic future
environments.
Police Leadership Styles
Related Entries
▶ Managing Innovation of Policing
▶ Measuring Police Performance
▶ Methodological Issues in Evaluating Police
Performance
Recommended Reading and References
Adlam R, Villiers P (eds) (2003) Police leadership in the
twenty-first century: philosophy, doctrine and developments. Waterside Press, Winchester
Allen DN (1982) Police supervision on the street: an
analysis of supervisor/officer interaction during the
shift. J Crim Justice 10:91–109
Andreescu V, Vito GF (2010) An exploratory study
of ideal leadership behaviour: the opinions of
American police managers. Int J Police Sci Manag
12:567–583
Brown MK (1988) Working the street: police discretion
and the dilemmas of reform. Russell Sage,
New York
Burns JM (1978) Leadership. Harper & Row, New York
Cohen B (1980) Leadership styles of commanders in the
New York City police department. J Police Sci Adm
8:125–138
Engel RS (2001) Supervisory styles of patrol sergeants
and lieutenants. J Crim Justice 29:341–355
Girodo M (1998) Machiavellian, bureaucratic, and transformational leadership styles in police managers: preliminary findings of interpersonal ethics. Percept Mot
Skills 86:419–427
Huberts LWJC, Kaptein M, Lasthuizen K (2007) A study
of the impact of three leadership styles on integrity
violations committed by police officers. Policing Int
J Police Strateg Manag 30:587–607
Jermier JM, Berkes LJ (1979) Leader behavior in a police
command bureaucracy: a closer look at the quasimilitary model. Adm Sci Quart 24:1–23
Johnson RR (2006) Management influences on officer
traffic enforcement productivity. Intl J Police Sci Mgt
8:205–217
Kellerman B (2004) Bad leadership: what it is, how it
happens, why it matters. Harvard Business School
Press, Boston
Kelloway EK, Sivanathan N, Francis L, Barling J (2005)
Poor leadership. In: Barling J, Kelloway EK, Frone
MR (eds) Handbook of work stress. Sage, Thousand
Oaks, pp 89–112
Kuykendall JL, Unsinger PC (1982) The leadership styles
of police managers. J Crim Justice 10:311–321
Maguire ER, Katz DM (2002) Community policing, loose
coupling, and sensemaking in American police agencies. Justice Quart 19:503–536
Police Legitimacy and Police Encounters
Mayo L (1985) Leading blindly: an assessment of chief’s
information about police operations. In: Geller W (ed)
Police leadership in America. American Bar Foundation, Washington, DC
McCall MW Jr, Lombardo MM (1983) Off the track: why
and how successful executives get derailed. Center for
Creative Leadership, Greensboro
McGregor D (1960) The human side of enterprise.
McGraw-Hill, New York
National Research Council (2004) Fairness and
effectiveness in policing: the evidence. Committee to
review research on police policy and practices.
In: Skogan W, Frydl K (eds) Committee on law and
justice, division on behavioral and social sciences
and education. The National Academies Press,
Washington, DC
Pursley RD (1974) Leadership and community identification attitudes among two categories of police chiefs: an
exploratory inquiry. J Police Sci Adm 2:414–422
Rowe M (2006) Following the leader: front-line narratives
on police leadership. Policing Int J Police Strateg
Manag 29:757–767
Schafer JA (2001) Community policing: the challenges of
successful organizational change. LFB Scholarly Publishing, New York
Schafer JA (2012) Effective leadership in policing: successful traits and habits. Carolina Academic Press,
Durham
Stamper NH (1992) Removing managerial barriers to
effective police leadership: a study of executive leadership and executive management in big-city police
departments. Police Executive Research Forum,
Washington, DC
Steinheider B, Wuestewald T (2008) From the bottom-up:
sharing leadership in a police agency. Police Pract Res
9:145–163
Tannenbaum R, Schmidt WH (1958) How to choose
a leadership pattern. Harv Bus Rev 36:95–101
Van Maanen J (1983) The boss: the American police
sergeant. In: Punch M (ed) Control in the police organization. MIT Press, Cambridge, MA, pp 275–317
Van Maanen J (1984) Making rank: becoming an American police sergeant. Urban Life 13:155–176
Witte JH, Travis LF, Langworthy RH (1990) Participatory
management in law enforcement: police officer, supervisor and administrator perceptions. Am J Police
9(4):1–23
Wycoff MA, Skogan WG (1994) The effect of
a community policing management style on officers’
attitudes. Crime Delinq 40:371–383
Police Legitimacy
▶ Organizational Change and Police Legitimacy
3665
P
Police Legitimacy and Police
Encounters
John D. McCluskey1, Robert E. Worden2,3 and
Sarah J. McLean3
1
Department of Criminal Justice, Rochester
Institute of Technology, Rochester, NY, USA
2
School of Criminal Justice, University at
Albany, State University of New York, Albany,
NY, USA
3
John F. Finn Institute for Public Safety, Inc.,
Albany, NY, USA
Overview
Research on police legitimacy is an area of growing interest among researchers and practitioners.
The concept of legitimacy ties together issues
of compliance, obedience, satisfaction, trust,
and order. The fundamental recognition of
its importance rests on the idea that police will
more easily maintain order when civilians view
the police as legitimate. Regimes, organizations,
and individual officers lacking legitimacy,
conversely, face crises based on undermined public support for their authority. The discussion that
follows considers how legitimacy is influenced
by individual encounters with authorities and
in a reciprocal sense, provides a context for
understanding how citizens relate to police in
terms of future cooperation and compliance.
Fundamentals
Definition of Legitimacy
Legitimacy is an outlook, with cognitive and
affective components as well as behavioral
predispositions. People who believe that the
police are legitimate trust the police to exercise
their authority appropriately, identify with
the police (Tyler and Huo 2002), have confidence
in the police to do the right thing, and feel an
obligation to obey the police. The sense of
P
P
3666
obligation to obey has been central to discussions
of legitimacy, because insofar as legitimacy
has behavioral implications, it points the way
toward broader and more economical ways for
the police to cultivate public support, cooperation, and compliance. More generally, we might
say that when police are perceived as legitimate,
they are thought to be entitled to have their
authority reciprocated with obedience.
From the observation that legitimacy encompasses a sense of obligation to obey, it does not
follow that obedience ensues reliably from
legitimacy; the association between legitimacy –
an outlook – and obedience (a behavior) is
imperfect. Many studies of attitude-behavior
consistency have reported only rather modest
associations, and against that empirical backdrop,
the reported relationships between legitimacy
and obedience appear quite robust, as we
summarize below, but they are far from perfect,
and we might surmise that the strength of the
relationship depends in part on the nature of
the behavior.
Legitimacy, to the extent it is an attitude
toward police, likely shares a high correlation
with other attitudes, such as satisfaction with
the police, which have been studied more
extensively. Furthermore, a variety of research
points to the likelihood that legitimacy is conceptually multidimensional, comprised of trust in
motives, obligation to obey, and identification
with the authority. This suggests that legitimacy
is a composite of both affect toward authorities
and their motives (e.g., trust and identification)
and perhaps what might be considered an
“action”-oriented component based on whether
the authority’s directives should be followed
(obligation to obey) or can form the basis for
citizen authorization (empowerment).
We would add that legitimacy is a property of
different types of authority – of political systems;
discrete political institutions, such as Congress or
the Supreme Court; legal systems; specific legal
institutions or agencies; and even private
employers. We focus here on the legitimacy of
the police, but we caution that inasmuch as the
police exercise a mandate based, arguably, in law
enforcement, police legitimacy is difficult to
Police Legitimacy and Police Encounters
separate from the legitimacy of the legal order
in which police are embedded. The police are,
thus, tied directly to the concept of legitimacy
as they represent the legal order and simultaneously are reliant on deference to that order
as a reservoir of public cooperation, obedience
to their authority, and trust. The relationship
between these concepts and legitimacy is
explored below.
Sources of Legitimacy
Wilbur Miller’s (1977) comparative history of
the New York City police and the London
“bobbies” illustrates the tie between legal order
and legitimacy. In England, for example, the
police were viewed as impersonal representations
of the established legal order, which was primarily divided along the lines of social class in the
relatively homogeneous English society. As such
the police had impersonal authority that stemmed
from a general civilian deference to the larger
political structure in 1800s England. Conversely,
in New York, during the 1800s, there was no such
easily agreed upon legal order, given the diversity
of religious, ethnic, and social classes occupying
the city. Thus, in this environment, police
authority was believed to emanate from personal
authority that was essentially earned respect and
deference from the community which accumulated over the course of face-to-face encounters.
Legitimacy of one’s role and action that ties
directly to the law would appear to be advantageous to police as the bobbies enjoyed a reluctant
respect even among the lower classes with
whom they had a more adversarial relationship.
However, this simplification glosses over several
differences between the US police system and
those of much of the remainder of the world.
First, centralized authority in England allowed
for uniform lawmaking and centralized police
forces. The USA, conversely, has a proliferation
of thousands of agencies, each which has to
build its own reservoir of authority from a local
(or State) body of law. Thus, fragmentation and
decentralization push legal legitimacy to local
levels in the US political system. While one
might argue this is a weakness and causes police
to diverge from a strictly legal mandate, there is
Police Legitimacy an