Investigating Research Integrity Proceedings of the First ORI

Document technical information

Format pdf
Size 1.3 MB
First found Jun 9, 2017

Document content analysis

Language
English
Type
not defined
Concepts
no text concepts found

Persons

Jim Wright
Jim Wright

wikipedia, lookup

Organizations

Places

Transcript

Investigating Research Integrity
Proceedings of the First ORI
Research Conference on Research Integrity
ii
Investigating Research Integrity
Proceedings of the
First ORI Research Conference on Research Integrity
editors
Nicholas H. Steneck, Ph.D.
Mary D. Scheetz, Ph.D.
[HHS Logo]
2002
iii
publication information page
iv
Preface
Over the past twenty years, a consensus has developed that integrity is vitally important to the health
of federally funded research and that the key stakeholders–individual scientists, research institutions,
associations and societies, government sponsors, and the general public–all play important roles in
fostering research integrity. However, there is little consensus about the importance of and a lack of
empirical scientific evidence on specific problems than can and do undermine integrity in research.
Even those of us who are experienced in research integrity issues have in the past based too much of
our thinking on personal experience, personal and philosophical biases, individual case exposes, and
the public, political, and media response thereto. Accordingly, to advance to the next level in our
understanding, it is time for new approaches to the study and discussion of research integrity.
Since its establishment in 1992, the Office of Research Integrity (ORI) has conducted a number
of studies on research misconduct and research integrity, some of which are ongoing. The goal of
these studies has been to develop a knowledge base for addressing important research integrity issues,
including: the impact of misconduct allegations on exonerated scientists, the experience of
whistleblowers in the aftermath of making allegations, the content of research guidelines adopted by
medical schools, and the the incidence of research misconduct. Over time, it became apparent to ORI
that a more comprehensive, coordinated effort in collaboration with extramural research scholars was
needed to develop baseline knowledge for understanding research integrity issues. This recognition
led to the development of the first Research Conference on Research Integrity in November 2000 and
the revised papers published in this volume. ORI has also begun, with support from the National
Institute of Neurological Disorders and Stroke, a related “Research Program on Research Integrity.”
In the background report that begins this volume, Assessing the Integrity of Publicly Funded
Research, Dr. Nicholas Steneck (ORI’s consultant for the November 2000 conference and the related
research program) has summarized the state of the empirical research on research integrity. This
report provided important background information for participants at ORI’s Research Conference on
Research Integrity and for scholars and others in the research community generally.
The research conference background report and the conference papers published in this volume
will hopefully provide an important catalyst for identifying important problems and for improving
our understanding of research integrity issues. Although research integrity has been a high profile
topic for some twenty years and some important preliminary studies have been conducted, the
publications in this volume, while contributing valuable information, make clear how little we really
know about many key issues, such as: how often research misconduct occurs, what situations tend to
encourage or prevent it, how human subjects are best protected, how often conflicts of interest occur
in research and how they affect the integrity of the research, how common questionable research
practices are and what harm they cause to the research process, how students and research trainees
learn the ethics of science, and what career pressures or other factors influence their ability and desire
to follow the most honorable scientific practices.
These unanswered questions provide a significant opportunity for the Public Health Service and
the research community to build a knowledge base for examining research integrity through further
research. Research will permit us to understand in a more thorough and genuine way the influence
v
that research integrity issues have on the careers of scientists, the operation of research laboratories,
the generation of accurate and useful research results, and the confidence of the public and political
community in the research enterprise. It will also provide a science base for making important
decisions—by government, by research institutions, by the community of scientists, and ultimately by
the general public—in response to future research integrity issues and concerns that will inevitably
arise.
Chris B. Pascal, J.D., Director
Office of Research Integrity
vi
Introduction
Researchers and research institutions are universally committed to maintaining high standards for
integrity in research. Precisely what this commitment entails, however, and whether it is being
fulfilled are questions that have not been subject to rigorous critical investigation. What is “research
integrity”? Can it be assessed? Do current research practices meet the high standards individuals and
institutions say they embrace? How are standards for best practices transmitted? Are current
approaches to fostering integrity appropriate and effective? Are all segments of the research
community appropriately contributing to the promotion of high standards for integrity in research?
Many individuals have provided answer to these questions, based on personal experience and
anecdotal evidence. Few scholarly studies have been undertaken to confirm or refute what is
commonly believed to be true about research integrity but is seldom demonstrated.
The papers published in this volume were originally presented at the first ORI Research
Conference on Research Integrity in Bethesda, Maryland, on November 19-20, 2000, and
subsequently reviewed and edited for publication. Abstracts for other papers and posters presented at
the conference but not published in this volume can be accessed at http://ori.dhhs.gov. Together, this
work represents the first comprehensive effort by a group of scholars to take a broad but critical look
at evidence underlying our assumptions about integrity in publicly funded research.
The organization of the Proceedings reflects the collective interests and judgments of the scholars
who responded to the call for abstracts for the Conference. Roughly half of the papers focused on
factors that influence attitudes toward integrity and actual research practices. These factors are
explored in these papers from the perspective of students and mentors, institutions and professions,
medical practice and clinical research, conflict of interest, and, the most-studied subcategory of
integrity, research misconduct. A second group of papers looked specifically at the way research
integrity is taught, either across institutions or in one institution or course. Finally, a significant
number of scholars tackled important methodological issues, looking at specific ways to detect
misconduct, publication practices, and different theoretical perspectives.
To speed dissemination and to facilitate access, all of the papers published in this volume have
previously been made available on the web. This limited-edition, bound copy is intended to create a
more permanent archive of the first Research Conference on Research Integrity. As this volume goes
to press, the call for abstracts for the second Research Conference on Research Integrity is being
transmitted to continue the work begun in November 2000.
Nicholas H. Steneck, Ph.D.
Department of History, University of Michigan
Office of Research Integrity, DHHS
Mary D. Scheetz, Ph.D.
Office of Research Integrity, DHHS
vii
viii
Contents
Preface
v
Introduction
vii
Assessing the Integrity of Publicly Funded Research
1
I. Norms and Environmental Issues
1. Students and Mentors
17
What Would Get You in Trouble: Doctoral Students’ Conceptions of Science and Its
Norms (19)
Data Manipulation in the Undergraduate Laboratory: What are we teaching? (27)
Preliminary Observations on Faculty and Graduate Student Perceptions of Questionable
Research Conduct (35)
Constructing a Personal Model of Research: Academic Culture and the Development of
Professional Identity in the Professorate (41)
Undergraduate Academic Cheating as a Risk Factor for Future Professional
Misconduct (49)
2. Institutions and Professions
55
Comprehensive Guidelines for the Responsible Conduct of Researchers (57)
Research Integrity in Social Work: Status, Issues, and Future Directions (65)
Organizational Influences on Scientific Integrity (73)
3. Medical Practice and Clinical Research
91
Waiving Informed Consent: Long-Term Consequences for the U.S. Military (93)
Promoting Scientific Integrity: The Long Road Ahead–Some Considerations from
Espírito Santo, Brazil (99)
Ethical Research Practice with Human Participants: Problems, Procedures, and Beliefs of
Funded Researchers (105)
Balancing Risks and Benefits of Deception in Assessing Genetic Screening (113)
Research Integrity and the Direct Involvement of Persons with Disabilities (117)
ix
4. Conflict of Interest
125
What is Driving Policies on Faculty Conflict of Interest? Considerations for Policy
Development (127)
The Commercialization of Academic Science: Conflict of Interest Policies and the
Faculty Consultant (133)
5. Understanding Misconduct
141
Preventing Scientific Misconduct: Insights from “Convicted Offenders” (143)
The Relative Efficiency of Research Misconduct Investigations Involving Personal Injury
vs. Injury to the Scientific Record (151)
Ethical Evaluation of Misconduct Cases (159)
Potential Cultural Factors In Scientific Misconduct Allegations (163)
Whistleblowers in Environmental Science, Prevention of Suppression Bias, and the Need
for a Code of Protection (167)
II. Teaching
6. Training in the Responsible Conduct of Research
177
Influencing the Moral Dimensions of Professional Practice:
Implications for Teaching and Assessing for Research Integrity (179)
Research Ethics in US Medical Education: An Analysis of Ethics Course Syllabi (189)
Teaching Ethics in Biomedical Science: Effects on Moral Reasoning Skills (195)
Fostering Research Integrity through Educational Programs: Lessons Learned at the
University of Minnesota (203)
Being a Scientist: Educating for Ethical Conduct (209)
Encouraging Accountability in Research: A Pilot Assessment of Training Efforts (215)
A Plea for Pursuing New Dimensions of Assessment in the Teaching and Learning of
Research Integrity (223)
7. Responsible Conduct of Research Courses
229
The Responsible Conduct of Animal Research (231)
An Effective Short Course on Research Integrity (237)
Resources for Instruction in Responsible Conduct of Research (241)
An Interactive Web Site for Ethics Training: http://storiesandquestions.com (247)
III. Research Theory and Methods
8. Detection Methods
253
The Misuse of Statistics: Concepts, Tools, and a Research Agenda (255)
Images as Evidence: Forensic Examination of Scientific Images (261)
Terminal Digits and the Examination of Questioned Data (269)
x
9. Publication Practices
279
Guidelines on Plagiarism and Paraphrasing in Writing Manuals Across Various Disciplines (281)
Instructions to the Author: An Integrity Issue (285)
Erratum Citation and Accuracy in the Publication Record (291)
10. Theory and Models from other Disciplines
297
An Epistemic Model for Moral Hazards in Scientific Enterprises (299)
Scientific Misconduct as Organizational Deviance (305)
A Market Approach to Research Integrity (315)
Methods for Research on Research Integrity: Doing Research on Sensitive Topics (321)
Research Misconduct: A Multiperspectival Approach (327)
Keyword Index
335
xi
Assessing the Integrity of Publicly Funded Research
Nicholas H. Steneck, Department of History, University of Michigan, USA
Keywords Accuracy, Authorship, Bias, Conflict of interest, Duplicate publication, Misconduct, Other
research practices, Peer review, Research on research integrity, Self correction
Since the early 1980s, when research integrity became a major national concern as a consequence of
reports of misconduct in research, several thousand publications have in one way or another reported
on, analyzed, and/or expressed opinions about the integrity of publicly funded research. Despite
widespread interest in research integrity, however, the integrity of researchers has not been subject to
the same critical study as other professionals. The research articles listed at the end of this paper
account for no more than 3-4% of the total literature on research integrity.
The lack of research on research integrity presents a significant problem for government, research
institutions, and professional societies. If integrity is defined as being honest in your dealings with
others, there is ample evidence to suggest that from time to time publicly funded research falls short
of this mark. As the articles summarized in this Paper confirm, researchers do commit misconduct;
research results are inappropriately influenced by bias, conflicts of interest, and just plain
carelessness; and researchers allow personal ambitions and biases to get in the way of the supposed
objectivity of the research process. Publicly funded research does not always achieve the high
standards that researchers, research institutions, and professional societies commonly set for
themselves. This much is known.
In contrast, too little is known about the causes and significance of, or remedies for, research
practices that fall short of the ideals set for the responsible practice of research.
• Is research misconduct rare or are the cases reported simply the tip of some unmeasured iceberg?
• Are there accepted norms or standards for research and, if so, how are they set, learned, and
monitored?
• Are the regulations that currently govern publicly supported research sufficient and well enough
enforced?
• Which practices that seem to fall short of accepted standards matter most from the standpoint of
protecting the public’s investment in research?
• Are there ways to foster integrity and thereby to prevent misconduct?
• Do research ethics courses make any difference?
• What influence does the research climate have on research integrity?
Each of these questions has at one time or another been raised and answered in the literature on
research integrity. Few of the answers given have been based on critical understandings of research
The information and views presented in this report are those of the author and do not reflect the official views or policies of
the Office of Research Integrity or the co-sponsoring organizations.
Corresponding author: Nicholas H. Steneck, Department of History, University of Michigan, Ann Arbor, MI 48109-1003,
734-647-4868 (voice), 734-647-4881 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
a measure of the degree to which researchers
adhere to the rules or laws, regulations,
guidelines, and commonly accepted professional
codes and norms of their respective research
areas.
Finally, a note of caution needs to be added.
This survey of the RRI literature is of necessity
selective and evolving. It places more emphasis
on the biomedical sciences than the physical or
social sciences. It does not do justice to the rich
literature on peer review. It almost certainly has
missed important articles that need to be included
in the RRI literature. As a result, it will almost
certainly be updated, and therefore comments
and additions are welcomed.
as a profession, largely, as noted, because
research as a profession has not be the subject of
careful observation and controlled study.
The remainder of this Paper presents a brief
analysis and summary of the research literature
on research integrity.
• Section one presents an overview of what is
known about the frequency of research
misconduct (FFP).
• Section two discusses the complex and
growing literature on research practices that
seemingly compromise professional standards but may not constitute outright misconduct.
• Section three surveys the research that has
been done on approaches to providing
instruction on the responsible conduct of
research (RCR).
• Section four explains how the literature cited
in this Paper was selected, some of its
characteristics, and the limitations of this
analysis.
The bibliography at the end provides a complete
list of references cited in the Paper, a summary of
the RRI literature sorted by topics, and a
comprehensive listing, sorted by first author, of
the RRI literature with abstracts.
Throughout this Paper, I have used the terms
“research misconduct,” “scientific misconduct,”
or simply “misconduct” to refer to the three
behaviors outlined in the common government
definition of research misconduct, namely
fabrication, falsification,and plagiarism (FFP) in
proposing, conducting or reporting the results of
research. While none of these behaviors is selfexplanatory, the crucial element in each is a
deliberate intent to deceive or mislead. Deliberate deception is clearly not consistent with
good research practice and is generally agreed to
constitute misconduct.
A second term used throughout this report,
“integrity,” is more difficult to define. Integrity
is a measure of wholeness or completeness.
When applied to professional behavior, it is
essentially a measure of the degree to which
someone’s (or some institution’s) actions accord
with ideal or expected behavior. However, the
ideals or expected behaviors for professional
conduct are complex, not always well defined,
and subject to change or reinterpretation. I have,
therefore, adopted a fairly inclusive definition of
integrity and assumed that it can be thought of as
Misconduct
Opinion about the extent of misconduct (FFP) in
publicly funded research is sharply divided. In
public testimony and editorials, researchers have
commonly argued that research misconduct is
rare. Support for this position is based on the
fact that the documented cases of misconduct are
few in number in comparison with the total
number of individuals engaged in research.
Approximately 200 cases of misconduct have
been confirmed by the federal government over
the last decade. Dividing cases by total
researchers, this works out to a rate of about 1 in
10,000 over 20 years, assuming approximately
2,000,000 active researchers, or 1 in 100,000 per
year. Critics of the way publicly funded research
is conducted and administered counter that the
reported cases represent the tip of a larger but
uncharted iceberg. Support for this view is based
in part on documented and presumed examples of
the reluctance of researchers and research
institutions to pursue cases of misconduct (for
early warnings about possible larger numbers,
see: 1, 2). Which, if either, opinion is correct
remains to be determined.
Direct evidence
Research undertaken to clarify the extent of
scientific misconduct suggests that it may be
more common than the 1 in 10,000 or lower
estimates. Evidence for this position comes from
three direct approaches to measurement:
• It is reasonable to presume, based on research
in other fields, that confirmed cases underestimate actual cases (3). Further research is
needed to determine whether under-reporting
in research is trivial or significant.
2
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
Year
Author
Population
Place
Sample
Size
Responses
(%)
Misconduct
FFP
1976
St. James-Roberts
Readers, New Scientist
En gland
??
199
(?)
92%
?
1987
Tagney
Phys, biol, behav, & soc. scientists
Major research university, US
1100
245
(22%)
–
32%
1992
Kalichman
Biomedical trainees
UC San Diego, US
2010
549
(27%)
36%
–
1993
Swazey
Chem., civil eng., microbiol., sociol.
US survey, faculty/graduate
4000
--/-(72/59%)
44/50%
6/9%
1993
Hals
PIs, biomedical sciences
Health Region IV, Norway
159
119
(70%)
27%
–
1995
Bekkelund
Biomedical researchers
Norway, random survey
274
215
(80%)
22%
3%
1996
Eastwood
Post-doctoral training fellows
US, random national survey
1005
324
(33%)
58%
3-12%
Table 1. Surveys of the Level of Misconduct in Research
troubling discrepancy between public statements
about how “rare” misconduct in research
supposedly is and the more private belief on the
part of many researchers that it is in fact fairly
common. How can these two views be
reconciled?
Second, whatever the actual rate of
misconduct, it is not so much the rate as the
significance of the misconduct that matters most.
Summarizing the results of scientific data audits
of the Cancer and Leukemia Group B’s clinical
trials, Weiss et al. conclude that “scientific
improprieties have occurred very rarely...”
(8, p. 459). “Very rarely, in this case, is based on
a quantitative estimate of 0.28% (p. 462)–28
cases of misconduct for every 10,000 clinical
researchers or one case for every 357 clinical
researchers. On what basis can this rate be
judged as either “rare” or “significant”? Clearly,
understanding the importance of misconduct in
research requires not only better estimates of
numbers but also of significance. How much
does a case of misconduct in research actually
cost the public in terms of wasted research
dollars, of deceptive findings that mislead other
researchers until the misconduct is discovered,
and perhaps of negative impacts on patient
health?
• Surveys of knowledge of misconduct consistently report knowledge rates above 1%
(Table 1). Reported knowledge of misconduct remains above 1% (1 in 100, or 100
times higher than the 1 in 10,000 estimate)
even when researchers are asked about their
own research group and when misconduct is
specifically limited to FFP. One survey
specifically asked researchers whether the
misconduct they were aware of was public
knowledge. Of the roughly one-in-four
researchers who were aware of misconduct
(27%), 47% said that the cases were not
public knowledge (4).
• Audits of research procedures and results have
turned up “significant problems” or “major
deviations” a levels that range at and above
the 10% level (5-8). These results do not
correlate directly with FFP, since they do not
take into account whether discrepancies
result from deliberate actions.
The results of surveys, audits, and estimates of
the rate of under-reporting raise two important
issues for further consideration. First, however
the results of surveys and audits are ultimately
interpreted or clarified, there remains the
3
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
career pressures may make it necessary to engage
in practices that they also know are wrong.
That significant numbers of beginning
researchers may in fact do what they say they
will do has been confirmed in a series of audits
of the research publications listed on residency
fellowship applications. These audits report
significant numbers (15% and higher) of
misrepresentations, from seemingly trivial
offenses such as inflating author rank to listing
articles “in press” when they were not, listing
papers in journals that do not exist, and listing
bogus articles in real publications (Table 3) (2327). Similar practices are generally counted as
FFP when they occur in research grant
applications or resumes submitted for promotion.
One final piece of indirect evidence that
should be noted is the confirmed reluctance of
researchers to report suspected misconduct.
• As noted above, Hals reported that roughly
one-in-four researchers (27%) who knew of
misconduct, said that the cases they knew of
were not public knowledge, which could
mean they were not reported (4).
• In Tagney’s survey conducted at one research
institution, roughly half of those who reported suspecting misconduct took no action
(28).
• Korenman’s
1992
1996
study of the
Kalichman Eastwood attitudes of
15.1%
12%
researchers and
14.8%
institutional
7.3%
15%
representatives
13.5%
-toward miscon1.3%
< 2%
duct found that
14.2%
27%
researchers were
-41%
more likely to
favor informing
Indirect evidence
Gathering information on the likely prevalence of
misconduct in research can be approached
indirectly. For example, many studies have
documented that cheating is common in the
educational system at all levels and in all
programs. The rates vary from well above 50%
for high school and college undergraduates (912) to levels between 10% and 30% for
professional students (13-20). One survey
specifically asked whether misconduct at this
level was indicative of future performance. Of
246 faculty and administrators responding, 216
(86%) felt that it was so indicative (14, p. 34). If
this estimate of the relationship between student
conduct and later professional conduct is true, it
would support the contention that the prevalence
of misconduct in research may be higher than the
small number of confirmed cases suggest.
The prevalence of a willingness to engage in
misconduct has been documented into graduate
and post-doctoral research education.
Kalichman’s and Eastwood’s surveys report that
significant numbers of students (above 10%,
except for fabricating data) will omit or change
evidence and add honorary authors if it will help
get papers published or grants funded (Table 2)
(21, 22). Students who are in the beginning
stages of becoming researchers clearly feel that
Action
Past misconduct (yes/no?)
Future misconduct (yes/no?)
...modify data for paper
...modify data for a grant application
...fabricate date for a paper or grant application
...select or omit data for paper or grant application
...list an undeserving author
Table 2. Self-reported attitudes toward misconduct
Author
Specialty
Total applications
...with citations
...misrepresented
Total citations
...misrepresented
Research experience
...not confirmed
1995
Sekas
Gastroenterology
236
53 (22%)
16 (30%)
--138 (59%)
47 (34%)
1996
Gurudevan
Emergency
Medicine
350
113 (32%)
23 (20%)
276
44 (16%)
---
1997
Panicek
Radiology
1998
Bilge
Pediatrics
201
87 (43%)
14 (16%)
261
39 (15%)
---
404
147 (36%)
29 (20%)
410
41 (10%)
---
Table 3. Misrepresentation in medical resident training program applications
4
1999
Dale
Orthopaedic
Medicine
213
64 (30%)
11 (17%)
76
14 (18%)
---
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
Panel on Scientific Responsibility and the
Conduct of Research specifically set out a
separate category of research behavior called
“Questionable Research Practices.” The Panel
recognized that such practices “...violate
traditional values of the research enterprise and
... may be detrimental to the research process,”
but it was not willing to include them under
“misconduct.” It did concede, however, that
since “...the relationship between these two
categories is not well understood ... [i]t may be
difficult to tell, initially, whether alleged
misconduct constitutes misconduct in science or
a questionable research practice” (30, pp. 5-6,
29).
Whether or not “other questionable
practices” constitute misconduct is irrelevant for
the purposes of this Report. What is relevant is
the fact that any practice that deviates
significantly from the “rules, regulations,
guidelines, and commonly accepted professional
codes or norms for the responsible conduct of
research” (the definition for integrity given in the
Introduction) can compromise and currently are
compromising the integrity of publicly funded
research. However, until more is known about
these practices, it will be difficult to suggest how
seriously they need to be taken.
The remainder of this section summarizes
some of the research on other practices that can
compromise the integrity of research. The
summary is intended to be more illustrative than
exhaustive. Some aspects of research practice,
such as authorship and peer review, have been
the subject of intense study and hundreds of
publications, thanks in large part to the
Congresses on Biomedical Peer Review
organized by JAMA editor, Drummond Rennie
(31). Exhaustive coverage is therefore not
possible. Rather, the goal of this section is to
focus on some areas of potential concern and
illustrate some of the findings that have emerged.
colleagues whereas institutional representatives favored reporting to supervisors and
deans (29).
These findings confirm the suspicions of the “tipof-the iceberg” school, which argues that
reported cases are not an accurate measure of
actual levels of misconduct. No controlled
studies of under-reporting have been undertaken
to assess the rate of under-reporting, making it
difficult to conclude whether it is significant.
Cheating or misconduct on the path toward
becoming a researcher does not, of course,
demonstrate that misconduct continues once
students become researchers. Under-reporting
may not seriously compromise estimates of the
amount of misconduct. Reasons can be given to
suggest that some of the estimates of misconduct
given in the various surveys reported above may
be too high as well as reasons to suggest that they
may be too low. The differences between the
“rare” and “tip-of-the-iceberg” schools can
therefore not be resolved easily. What is
important to note, however, is that in seeking to
refine understandings and resolve the differences
between the two schools, the range of uncertainty
that exists is significant. In terms of decimal
points, the range is not a matter of one or two
orders of magnitude but closer to four or five
orders of magnitude, varying from 1 in 100,000
or less to 1 in 100 or more. And this, in turn,
makes it difficult, if not impossible, to estimate
the public costs of misconduct when determining
what policies are needed to protect the public’s
investment in research.
Other Research Practices
Over the past twenty years or longer, the
discussion of “research integrity” has focused
primarily on “research misconduct,” based on
widespread agreement that misconduct (FFP) is
wrong or fraudulent. While it is true that
research misconduct clearly can undermine the
integrity of publicly supported research and
therefore needs to be taken seriously, so can other
research practices, such as sloppy research,
inappropriate bias, conflict of interest, or poor
mentoring.
The existence of other research practices that
can compromise integrity has been recognized by
the research community, but there has been no
agreement on how to respond to them or how
seriously they should be taken. In its 1992
report, Responsible Science, the NAS/NAE/IOM
Accuracy
Accurate information is vital to research.
Research is a cooperative and cumulative
enterprise. Researchers build on the work of
others, which means the information they have
about other work and the way research is
conveyed must be accurate; however, a number
of studies suggest that research results are not
always conveyed accurately.
• Information presented in abstracts does not
5
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
substantial agreement about whether to
publish (36) and by studies of how fraudulent publications have made it to press (37).
How much effort should be made to improve
peer review requires more information about
how well it is working and the price of its
shortcomings.
• Peer review has been shown to have institutional (38), national (39, 40), methodological
(39, 41), gender (42) and outcome biases
(43-45). Bias, obviously, runs counter to the
value-neutral goal of research.
• Considerable uncertainty exists about the best
ways to improve peer review. Traditional
approaches, such as blinding, issuing clear
instructions, or relying on experienced
researchers, have had different measures of
success (46-53).
• Studies of peer review have raised questions
about whether it helps or hinders innovation
(54, 55).
One review of the rich literature on peer review
concludes: “Because of the central place of peer
review in the scientific community and the
resources it requires, more studies are needed to
define what it does and does not accomplish”
(56). This work will fortunately be fostered by
the future Congresses on Biomedical Peer
Review and similar efforts.
always accurately reflect the information
given in the article itself. One study reported
major discrepancies in abstracts (inconsistencies or information that was not contained in
the body of the article) in 55 of 203 randomly selected articles (32).
• Studies have reported that significant numbers
(above 10%) of published articles misuse
statistics or contain statistical errors (33).
• Random checks on citations and quotations in
published articles have reported error rates
well above 10%. Errors were counted as
“citation errors” when the names, pages, or
other information needed for locating an
article was inaccurate (minor) or when the
referenced article could not be located based
on the information given (major). Errors
were counted as “quotation errors” when the
reference oversimplified or exaggerated
information given in the referenced article
(minor) or when the information given in the
original article did not support or contradicted claims made in the reference (major)
(34, 35).
Inaccuracies in abstracts, the use of statistics, and
references do not necessarily invalidate research
results. Conclusions or pieces of evidence
presented only in an abstract but not in the body
of an article could be true. Research results
bolstered by inflated or deceptive statistics or
inaccurate references to other studies might still
be true. At issue, however, is not whether the
results are ultimately true or accurate but whether
the word (or words in this case) of researchers
can always be trusted. The clear answer to this
question, unfortunately, is that it (they) cannot.
Self-Correction
Researchers constantly read and check each
other’s work. The routine process of using the
work of others in the day-to-day practice of
research provides an additional mechanism for
detecting and correcting errors and other
problems in research, such as research
misconduct. Research is, in other words, selfcorrecting, which further ensures its integrity.
However, research on the effectiveness of selfcorrection in research has shown that this
mechanism is not as vigilant as one might expect.
• Studies of some of the first publicly documented cases of misconduct found that
publication of a retraction reduced the
citation of fraudulent articles but did not
eliminate it (57-59).
• One recent study of articles retracted for a
broad range of reasons, from outright fraud
to acknowledged experimental errors or later
failure to replicate, concluded that retracted
Peer Review
Inaccuracy and other problems in publication are
purportedly reduced, if not eliminated, through
peer review. In general, the peer review system
enjoys considerable support within the research
community and is seen by most as the foundation
on which professional self-regulation rests. This
does not mean, however, that peer review is
above criticism or not in need of further
improvement.
• That peer reviewers miss problems in publications has been documented by the fact that
different reviewers detect different problems
in manuscripts, even when they are in
6
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
for judging authorship found that 19% (65)
and 36.4% (66) of papers did not meet these
criteria.
• Evidence suggests that the rules for authorship are poorly understood, interpreted
differently by different researchers, and not
well communicated from senior to junior
researchers (22, 67, 68).
• Patterns of authorship and the increase in
disputes over authorship suggest that decisions about authorship are significantly
influenced by the research environment (69,
70).
The importance of the truthful reporting of
research contributions through authorship is
widely recognized. The NIH Guidelines for the
Conduct of Research note in particular that:
articles continue to be cited and used as a
significant rate. Of 299 post-retraction
citations listed in the Abridged Index
Medicus, only 19 (6%) mentioned the
retraction; 17 (6%) explicitly and 263 (88%)
implicitly reported the retracted work as
“valid” (60).
• Research on the process by which articles are
retracted and erroneous information withdrawn has show that it is slow (60, 61) and
in some key ways ineffective (60-63).
Findings such as these have important policy
implications. In his study of retraction notices,
Budd agrees that research is self-correcting, but
then he adds: “...there may be a great deal of
time, effort, and money spent in discovering that
some research is not useful. If erroneous or
fraudulent work lives on in the literature, the
amount of time, effort, and money to correct
work may be even greater” (60, p. 297). At
issue, in other words, is not whether research
errors are corrected, but when. Failure to correct
the literature in a timely and responsible manner
is as much a matter of integrity, viewed from the
public’s investment in research, as a failure to
correct at all.
For each individual the privilege of authorship
should be based on significant contribution to
the conceptualization, design, execution, and/
or interpretations of the research study, as well
as a willingness to assume responsibility for
the study. Individuals who do not meet these
criteria but who have assisted the research by
their encouragement and advice or by providing
space, financial support, reagents, occasional
analyses or patient material should be
acknowledged in the text but not be authors.
(71, p. 10)
Authorship
In principle, research results are more important
than researchers. Who publishes an article
should not matter. What matters most are the
results. In practice, however, authorship is
vitally important to, and significantly influences,
the research process. Most research funding
today is dependent on productivity. Review
panels want to know not only what a researcher
is planning to do but what she or he has done.
Advancement in academic research is not
possible without publication. Getting one’s name
on research papers is important–so important that
as many as one in five aspiring researchers
misrepresents publications on résumés in an
attempt to improve his or her standings as a
researcher (see Table 4).
As with the other research practices
discussed in this section, there is considerable
evidence to suggest that the ideal standard for
determining authorship is not followed in
practice and that expected authorship practices in
general are sometimes not clearly defined or
conveyed.
• Two studies that used the ICMJE criteria (64)
Authors who ask or agree to be listed on papers
to which they have not made substantial
contribution compromise the integrity of the
research environment. The same is true of the
41% of graduate students who report a
willingness to list undeserving authors on their
papers (see Table 3, above).
Duplicate Publication
In its advice to intramural researchers, NIH
research Guidelines caution researchers about
duplicate publication:
Timely publication of new and significant
results is important for the progress of science,
but fragmentary publication of the results of a
scientific investigation or multiple publications
of the same or similar data are inappropriate.
(71, p. 8)
Despite widespread agreement that duplicate
publication is inappropriate, the rate of duplicate
publication (publishing the same article twice
without reference) seems to hover at about 10%
(Table 4) (72-76). Based on his study of
publication trends in the British Medical Journal,
Waldron suggested that duplicate publication was
7
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Study
Waldron (1992)
Bernard (1993)
Koen (1994)
Blancett (1995)
Bloemenkamp (1999)
Journal
BMJ
NTvG
NTvG
INJS
NTvG
Articles
354 published
172 published
108 rejected
642 published
148 published
Duplicate %
6-12%
11%
4%
9%
7%
Table 4. Percent duplicate publication
tenets” (p. 138). From the public’s perspective,
however, it makes no difference whether the
duplication is intended or not. If researchers do
not take steps to ensure that a second or third
publication of a body of data is recognized as
such, the public could be harmed and the
integrity of the research process undermined.
increasing (72). Bleomenkamp more recently
reported that the duplicate publication rate for
articles in Nederlands Tijdschrift voor
Geneeskunde has remained constant over the last
10 years and the number of authors referencing
the second publication has increased
significantly, from 22% to 73%.(76).
Duplicate publication adversely effects
research in a number of ways. It can waste time
(editors and reviewers) and resources (library
funds and reprint costs). It also makes it difficult
to evaluate the productivity of researchers. But
perhaps most importantly, in clinical research it
has the potential to inappropriately distort or
bias findings if the duplicate publications are
more prevalent in one treatment regimen.
• In a meta-analysis of post-operative effects of
ondansetron, Tramer and Reynolds reported
that “17% of published studies and 28% of
the patient data were duplicated. Moreover,
duplication was more common in studies that
reported greater treatment effect. This bias,
according to Tramer and Reynolds, “led to a
23% overestimation of ondansetron’s
antiemetic efficacy” (77).
• Jefferson reports that in a Cochrane review of
the effects of Plasma Derived Vaccines, he
and his colleagues suspected that 25% (15 of
60) of the trials identified during the first
phase of review were duplicate publications.
This percentage increased to 43% (3 of 7)
when they progressed to the second phase of
review. Being aware of the problem of
duplicate publication, his group excluded the
duplicate studies, but doing so is not common practice (78).
In the final analysis, Jefferson considers only
“publishing redundant material with the intention
of misleading the public, editors and readers, in
order to make them believe the study is different
from the original” as a “breach of current ethical
Bias and Conflict of Interest
There has been considerable debate about the
role of values and personal interest in research
ever since Merton proposed “disinterestedness”
as one of four key values on which science rests
(79, p. 116). It is now widely recognized that
values influence research (80), but there is also a
common understanding that the influence of
values should be minimized and made public,
particularly when financial interests are involved.
Considerable evidence exists to support the
contention that personal interest does influence
research behavior. Positive-outcomes bias
(favoring publications that report positive results
over those that report negative results or that do
not find results) has been demonstrated in a
number of studies (44, 81, 82). The reverse
effect has also been reported, that is, slower
publication rates for studies that fail to find a
particular result (45). Studies are just beginning
to assess how these interests affect research and
whether they are being properly managed (8385).
In calling controversial publication,
reporting, and other research practices
“questionable,” the NAS report, Responsible
Science, highlights an important problem. (30)
“Integrity” is not an all-or-nothing proposition.
There is a difference between a failure to check
the spelling of every author’s name or to catch
every typo and using improper statistics or
delaying the publication of a manuscript to please
a sponsor. It is not easy to pinpoint where or
when high standards for integrity in research give
way to careless research practices, to
8
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
succeeding.
irresponsible research practices or to misconduct.
The extremes (high standards for integrity and
misconduct) can be defined, but behaviors that
fall between, to one extent or another, are all
subject to interpretation. This, in turn, makes it
imperative that these behaviors are well
understood and their consequences evaluated,
both as part of the process of reassuring the
public that its research funds are being spent
responsibility and as needed background
information for developing responsible conduct
of research training programs.
RCR training
Straightforward efforts to evaluate the impact
RCR training has on attitudes or anticipated
behaviors have not reported any clear positive
results. Studies by Kalichman et al. and
Eastwood et al. compared receiving or not
receiving RCR training with anticipated research
behaviors. A study by Brown compared
receiving or not receiving RCR training with
self-reported perceptions of different ethical
standards. None of the studies found any
significant correlations between attitudes or
anticipated behaviors and RCR training (21, 22,
88). Brown’s study did report that RCR training
increased awareness of options in ambiguous
situations (p. 490). However, Eastwood’s study
reported that fellows who received RCR training
were more willing to grant honorary authorship
than fellows who did not (p. 95). Overall, direct
measures of attitudes and anticipated behavior
have pointed to some possible benefits, perhaps
one puzzling negative, and a great deal of
similarity between those receiving and not
receiving RCR training.
Efforts to refine the study of the impact of
RCR training have led to a difference of views on
appropriate outcome measures. Based on a
three-year effort to develop and assess an RCR
course at Dartmouth College, Elliot and Stern
argue that “if ‘ethical behavior’ is removed as a
basis for the evaluation of teaching ethics,”
effective assessment tools can be developed. In
the place of ethical behavior, they propose using
two familiar measures of success in academic
courses in general: “the skills and content taught
in the course and the learning environment in
which the teaching takes place” (89, p. 348). The
project allowed them to develop and test various
tools for evaluating these ends, which they argue
can be accomplished, “but only if [teaching of
academic research ethics] is treated as an
academic discipline by both faculty and students”
(p. 355).
Others believe that striving for some type of
behavioral or moral reasoning change is
appropriate for professional ethics instruction,
including RCR training, and that such change can
be measured. In a series of studies of medical,
veterinary, and dental education, Self, Baldwin,
Bebeau and colleagues have reported that: a)
traditional professional education programs may
erode and b) the addition of ethics instruction to
Education
It is commonplace for reports on research
misconduct/integrity to emphasize the
importance of education. Professions have an
obligation to society to educate future
generations of professionals, which includes
making future professionals aware of the
standards for responsible practice. Moreover, if
professional ethics education prevents
misconduct, it is in a profession’s best interest to
encourage this education, which most in fact do.
Through the 1980s, research ethics training
was commonly relegated to the laboratory and to
mentoring. This changed in 1989 when NIH and
ADAMHA instituted required “instruction in the
responsible conduct of research” (RCR) for all
training grants (86). The requirement stipulated
that training programs had to have instruction in
RCR, which in turn had to be described in the
training grant application. Although the
requirement technically had no “regulatory
teeth,” coming as it did in the highly competitive
environment of grant-getting, researchers and
research institutions quickly complied and
instituted a wide variety of research ethics or
RCR training programs (87).
The increase in formal RCR training raises
an obvious and researchable question: has it or
will it make any difference? At the present time,
there is no convincing evidence that it does, but
this does not necessarily lead to the conclusion
that RCR training is ineffective, unnecessary, or
unwise. The newness of most programs means
that their impact may not yet be apparent. RCR
training is delivered in different ways and
different settings, making it difficult to isolate the
influence this one factor has on the complex
process of becoming a responsible researcher.
And perhaps most importantly, there is no
agreement on the goals of RCR education,
making it difficult to judge whether it is
9
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
traditional programs improves the ability of
students to engage in moral reasoning (90-97).
Whether changes in the ability to engage in
moral reasoning measured in professional
education settings generally can be applied to
RCR training in particular and whether changes
in moral reason have any lasting professional
consequences remains to be determined.
The research needed to plan effective RCR
programs will clearly need to take into account
more than what goes on in the RCR classroom.
Studies have shown that environment is closely
linked to what students feel they must do as
opposed to what they should do (17, 18, 20, 22).
Although the 1995 survey of the attitudes and
experiences of 2,000 graduate students with
misconduct (Table 2, above) indicates “that
fraud, plagiarism, and related forms of
misconduct are the results of individual
predilections or failures of judgement...” (98, p.
225), Anderson et al. in commenting on these
results still point to important influences exerted
by environment and mentoring relations (p. 226).
Without attention to the full context within which
integrity is learned and decisions made about
right and wrong actions, the goal of ensuring the
responsible conduct of research through RCR
training could well be negated by influences in
the research environment.
Other efforts to educate
In discussions of ways to improve the integrity of
research, surprisingly little attention has been
given to the role of clear rules and routine
monitoring or data audits. If the ultimate goal of
research ethics/integrity policy is simply to
ensure high standards for publicly supported
research, the simplest way to achieve this goal
may be to make the rules as explicit and clear as
possible and then to check to make sure they are
being followed. For each of these approaches to
“educating” researchers, there is interesting
research that suggests what may or may not
work.
Over the last decade, new rules have been
formulated for reporting research. Particular
attention has been paid to two key areas–journal
publication in general and clinical trial reporting.
Studies of the effect of new rules suggested that
they have had mixed results.
• Two studies that looked at the adoption of
specific standards for reporting clinical trials
by several medical journals concluded that
there was room for improvement (99, 100).
Junker suggest that more journals should
require authors to follow the Consolidated
Standards of Reporting Trials (CONSORT)
(101). Clarke and Chalmers conclud that
“there is little evidence that journals have
adequately implemented the CONSORT
recommendation that results of an RCT
[randomized controlled trial] be discussed in
light of the totality of the available evidence”
(p. 280).
• In studies of measures to improve the quality
of abstracts, Pitkin found that instructions to
the authors had little impact (32, 102, 103).
• In a study of the impact of guidelines published in the British Medical Journal for
manuscripts on the economics of health care,
no difference was found in the quality of
manuscripts, although the guidelines were
judged to be useful for editorial purposes
(104).
• In a comparison of systematic reviews and
meta-analyses published following the
procedures of the Cochrane Collaboration
versus the more open-ended general reviews
published in journals, Jadad reported more
methodological rigor in the Cochrane
reviews (41).
• In a study of the impact of professional codes
in physics, Tarnow reported that postdoctoral
students were generally not aware of publication rules and spent little time with advisors discussing publication practices (68).
As a group, this research seems to support the
perhaps not unexpected conclusion that rules
alone will not change behavior and must be
accompanied by efforts to both make them
known and take them seriously. Simply making
information about rules for responsible behavior
available is not an effective way to foster
responsible behavior.
In contrast, data audits seem to have a
significant effect on research behavior. Two
studies of major government data audit programs
both report that serious misconduct declined over
the course of the studies.
• Shapiro and Charrow’s study of FDA audits
conducted between 1977 and 1988 reported
that the rates of specific deficiencies remained about the same throughout but “the
10
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
research integrity is based on some research.
For the purposes of this Report, “research”
has been defined as studies that have some
element of controlled investigation, which means
primarily but not exclusively surveys and
quantitative assessments. Limiting the definition
of research in this way obviously eliminates
many thoughtful articles and books from the
literature review, such as editorials, analytical
writings, historical and cases studies, and
philosophical analyses. The fact that works such
as these are not included in this Report should
not be taken as suggesting they are not important.
They clearly are crucial and in other contexts
certainly need to be considered. However, for
the purposes of the ORI RRI program, the
immediate goal is to gather hard evidence
relating to actual research practices, so that
policy-making can be based on the way research
is conducted as opposed to the way we may think
it is conducted.
Controlled quantitative research plays an
important role in scholarly investigation. Most
significantly, it helps establish reference points
for organizing and evaluating other information.
For example, historians, journalists, and others
have amply documented that misconduct takes
place in research. However, without some
quantitative assessments, it is difficult to know
what to make of individual cases of misconduct
or even of the entire body of confirmed cases.
Are they typical or atypical? Is misconduct
common or rare? Without some controlled
counting or surveys, it is difficult to place
individual events and behaviors into context.
Locating research on research integrity is not
a simple task. Keyword searching for the most
part does not separate scholarly analyses from
empirical studies. References located through
searches for “scientific misconduct,” “research
ethics” and other keywords need to be evaluated
for both relevance and method. The articles
summarized in this Report have been located
through standard keyword searches in several
different databases, checking references listed in
bibliographies, and in some cases by searching
for publications by scholars with known RRI
interests. Major emphasis has been placed on
work relating to the biomedical sciences in
particular and the hard sciences more generally.
Less attention has been paid to research on
integrity in the social sciences. The final RRI
bibliography contains 136 entries, most of which,
but not all, have some empirical or controlled
overall level of seriousness of the problems
... declined” (7, p. 130).
• Weiss et al. in their detailed look at the results
of audits conducted by the Cancer and
Leukemia Group B (CALGB) conclude that:
“The CALGB data audit process has been
successful in uncovering the very rare
instances of scientific misconduct and
pressuring group members to improve
adherence to administrative requirements,
protocol compliance, and data submission. It
has also served to weed out poorly performing institutions” (8, p. 464).
If results matter, then one of the most effective
ways to educate researchers about their
responsibilities may be to check more carefully
the work they produce.
Data audits have been resisted because they
are allegedly expensive, time-consuming, and
perhaps even counter-productive; e.g. too much
concern about the bookkeeping required to pass
audits might slow the progress of science. There
currently are no data to support these concerns.
There is evidence, reviewed by Armstrong, that
peer review can slow innovation in research (54,
pp. 70-71), but no evidence that data audits have
a similar effects. Moreover, Glick’s rough
estimates of the cost of data audits, based on
conservative estimates of the amount of careless
work and misconduct that may be affecting
research results, suggests that over the long term,
they will save public dollars. “Data auditing
would increase research productivity by 2.5-6%
(...), so that each dollar spent on such audits
might eventually benefit the public, 20 years
later, by an amount equivalent to $25-60” (3, p.
81). These results and estimations will no doubt
be challenged, but for now the evidence seems to
suggest that research audits might be an effective
and efficient way to detect misconduct and
reduce the rate of other questionable practices.
Research Literature Overview
As noted in the Introduction, over the last 20
years or longer, several thousand publications
have in one way or another addressed the issue of
integrity and/or misconduct in research. Most of
these publications are based on some research.
Reporters do research for news stories. Journal
editors investigate problems before writing
editorials. Taken to mean simply investigation or
study, most if not all that has been written about
11
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
research component.
That RRI has not yet developed into an
organized research field is more than evident
from the fact that the 136 articles summarized in
this Report appeared in 45 different journals
(Table 5) and two books (105, 106). Most
journals published only one or two articles.
There are, however, three important exceptions.
• Fifty-one of the 136 (37.5%) articles appeared
in JAMA. Most of these articles are on
integrity in publication and are the product
of the three peer review conferences organized by Drummond Rennie.
• Fourteen of the 136 articles (10%) appeared in
Academic Medicine. These articles are
mostly concerned with student conduct, not
research integrity specifically, but have been
included because they provide important
background on the values researchers may
have had as students.
• Eleven of the 136 articles (8%) appeared in
Science and Engineering Ethics. This group
of publications is split nearly evenly between
research ethics training and publication
practices. SEE is unfortunately not indexed
in MedLine®, which limits the knowledge of
this important group of publications.
Together, these three journals account for 76 of
the 136 articles. Three journals had three
research articles; five journals had two, and the
remainder published a single research article on
research integrity.
The fact that research on research integrity is
distributed so broadly through the scholarly
literature almost certainly slows research
progress. At the present time, the standard search
tools simply do not cut across the different
disciplines that contribute to RRI. What is
“discovered” in one field is thus not easily
known in other fields. More importantly,
however, is the fact that the absence of a well
defined literature and corresponding research
community makes interdisciplinary research on
research integrity more difficult. This second
shortcoming is particularly important for the
development of research on research integrity,
which of necessity must be interdisciplinary and
Journal of the American Medical Association (51)
Academic Medicine (14)
Science and Engineering Ethics (11)
British Medical Journal (3)
Journal of Professional Nursing (3)
Nederlands Tijdschrift voor Geneeskunde (3)
Accountability in Research (2)
Bulletin of the Medical Libraries Association (2)
Journal of Dental Education (2)
Lancet (2)
Medical Education (2)
Medical Reference Services Quarterly (2)
New Scientist (2)
Tidsskrift for den Norske lægeforening (2)
AIDS Education and Prevention (1)
American Journal of Medicine (1)1
American Journal of Public Health (1)
American Journal of Roentgenology (1)
American Scientist (1)
Annals of Emergency Medicine (1)
Annals of Internal Medicine (1)
Cambridge Quarterly of Healthcare Ethics (1)
Canadian Medical Association Journal (1)
Cancer Investigation (1)
Cognitive Therapy and Research (1)
Controlled Clinical Trials (1)
Image: The Journal of Nursing Scholarship (1)
Journal of Allied Health (1)
Journal of Bone and Joint Surgery (1)
Journal of Clinical Epidemiology (1)
Journal of General Internal Medicine (1)
Journal of Higher Education (1)
Journal of Information Ethics (1)
Journal of Investigative Medicine (1)
Journal of Medical Education (1)
Journal of Medical Ethics (1)
Journal of the Am. Veterinary Medical Association (1)
Journal of the Royal College of Physicians, London (1)
Minerva (1)
Nature (1)
New England Journal of Medicine (1)
Nordisk Medicin (1)
Nurse Educator (1)
Research in Higher Education (1)
The Psychological Report (1)
Table 5. Journals with RRI articles, listed by number of articles.
12
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
broadly inclusive.
The need for interdisciplinary research raises
one last observation about the RRI literature and
by implication the RRI community. Most of the
literature cited in this Report appears in
biomedical journals. The only major exception
are the eleven articles in Science and Engineering
Ethics, which, it should be noted, are not indexed
in MedLine® but are in BioEthicsLine, without
abstracts. That research on the integrity of
biomedical research (the primary focus of this
report) appears in biomedical journals is certainly
understandable, but the existence of this
publication pattern raises serious questions for
interdisciplinary research.
To be taken seriously in most academic
settings today, researchers must first succeed in
their primary research field. This means that
sociologists must publish in sociology journals,
psychologists in psychology journals, and so on.
In addition, they must pursue research that is
important to their primary fields of research.
Institutional factors such as this unquestionably
make the development of interdisciplinary
research on research integrity more difficult.
When added to the fact that there are few
incentives for researchers who are the subject of
RRI investigations to study their own integrity,
rather than pursuing research in their primary
fields of interest, establishing an interdisciplinary
RRI initiative and RRI community poses a
significant challenge.
8. Weiss RB, Vogelzang NJ, Peterson BA, Panasci LC,
Carpenter JT, Gavigan M, et al. A successful system of
scientific data audits for clinical trials. A report from the
Cancer and Leukemia Group B [see comments]. JAMA
1993;270(4):459-64.
9. Stern EB, Havlicek L. Academic misconduct: results of
faculty and undergraduate student surveys. J Allied
Health 1986;15(2):129-42.
10. Perry AR, Kane KM, Bernesser KJ, Spicker PT. Type A
behavior, competitive achievement-striving, and
cheating among college students. The Psychol Rep
1990;66(2):459-65.
11. Schab F. Schooling without learning: thirty years of
cheating in high school. Adolescence 1991;26(104):83947.
12. McCabe DL. Classroom cheating among natural science
and engineering majors. Sci Eng Ethics 1997;3.
13. Stimmel B, Yens D. Cheating by medical students on
examinations. Am J Med 1982;73(2):160-4.
14. Bailey PA. Cheating among nursing students. Nurse
Educator 1990;15(3):32-5.
15. Rozance CP. Cheating in medical schools: implications
for students and patients. JAMA 1991;266(17):2453, 6.
16. Anderson RE, Obenshain SS. Cheating by students:
findings, reflections, and remedies. Acad Med
1994;69(5):323-32.
17. Daniel LG, Adams BN, Smith NM. Academic
misconduct among nursing students: a multivariate
investigation. J Prof Nurs 1994;10(5):278-88.
18. Baldwin DC, Jr., Daugherty SR, Rowley BD, Schwarz
MR. Cheating in medical school: a survey of secondyear students at 31 schools. Acad Med 1996;71(3):26773.
19. Dans PE. Self-reported cheating by students at one
medical school. Acad Med 1996;71(1 Suppl):S70-2.
20. Satterwhite WM, 3rd, Satterwhite RC, Enarson CE.
Medical students’ perceptions of unethical conduct at
one medical school. Acad Med 1998;73(5):529-31.
21. Kalichman MW, Friedman PJ. A pilot study of
biomedical trainees’ perceptions concerning research
ethics. Acad Med 1992;67(11):769-75.
22. Eastwood S, Derish P, Leash E, Ordway S. Ethical
issues in biomedial research: Perceptions and practices
of postdoctoral research fellows responding to a survey.
Sci Eng Ethics 1996;2(1):89-114.
23. Sekas G, Hutson WR. Misrepresentation of academic
accomplishments by applicants for gastroenterology
fellowships [see comments]. Ann Intern Med
1995;123(1):38-41.
24. Gurudevan SV, Mower WR. Misrepresentation of
research publications among emergency medicine
residency applicants [see comments]. Ann Emerg Med
1996;27(3):327-30.
25. Bilge A, Shugerman RP, Robertson WO.
Misrepresentation of authorship by applicants to
pediatrics training programs. Acad Med 1998;73(5):5323.
26. Panicek DM, Schwartz LH, Dershaw DD, Ercolani MC,
Castellino RA. Misrepresentation of publications by
applicants for radiology fellowships: is it a problem?
Bibliography
1. Woolf P. Fraud in science: How much, how serious?.
Hastings Cent Rep 1981;11:9-14.
2. Broad WJ, Wade N. Betrayers of the truth: Fraud and
deceit in the halls of science. New York: Simon and
Schuster; 1982.
3. Glick JL. On the potential cost effectiveness of scientific
audits. Accountability in Research 1989;1(1):77-83.
4. Hals A, Jacobsen G. [Dishonesty in medical research. A
questionnaire study among project administrators in
Health Region 4 ]. Tidsskrift for den Norske
Lægeforening 1993;113(25):3149-52.
5. Shapiro MF, Charrow RP. Scientific misconduct in
investigational drug trials. N Engl J Med
1985;312(11):731-6.
6. Shapiro MF, Charrow RP. The role of data audits in
detecting scientific misconduct. JAMA 1989;261(May
5):2505-11.
7. Shapiro MF. Data audits in investigational drug trials
and their implications for detection of misconduct in
science. In: Lock S, Wells F, editors. Fraud and
miscondcut in medical research. London: BMJ
Publishing Group; 1993. p. 128-41.
13
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Am J Roentgenol 1998;170(3):577-81.
27. Dale JA, Schmitt CM, Crosby LA. Misrepresentation of
research criteria by orthopaedic residency applicants. J
Bone Joint Surg 1999;81(12):1679-81.
28. Tangney JP. Fraud will out—or will not? New Scientist
1987;115(August 6):62-3.
29. Korenman SG, Berk R, Wenger NS, Lew V. Evaluation
of the research norms of scientists and administrators
responsible for academic research integrity. JAMA
1998;279(1):41-7.
30. Committee on Science Engineering and Public Policy
(U.S.). Panel on Scientific Responsibility and the
Conduct of Research. Responsible science: Ensuring the
integrity of the research process. Washington, D.C.:
Nationl Academy Press; 1992.
31. Rennie D, Flanagin A. Congress on Biomedical Peer
Review: history, ethics, and plans for the future
[editorial]. JAMA 1998;280(3):213.
32. Pitkin RM, Branagan MA. Can the accuracy of abstracts
be improved by providing specific instructions? A
randomized controlled trial [see comments]. JAMA
1998;280(3):267-9.
33. Gardner MJ. An exploratory study of statistical
assessment of papers published in the British Medical
Journal. JAMA 1990;263(March 10):1355-57.
34. Eichorn P, Yankauer A. Do authors check their
references? A survey of accuracy of references in three
public health journals. Am J Public Health
1987;77(8):1011-2.
35. Evans JT, Nadjari HI, Burchell SA. Quotational and
reference accuracy in surgical journals. JAMA
1990;263(10):1353-4.
36. Garfunkel JM, Ulshen MH, Hamrick HJ, Lawson EE.
Problems identified by secondary review of accepted
manuscripts. JAMA 1990;263(10):1369-71.
37. Stewart WW, Feder N. The integrity of the scientific
literature. Nature 1987;325:207-14.
38. Garfunkel JM, Ulshen MH, Hamrick HJ, Lawson EE.
Effect of institutional prestige on reviewers’
recommendations and editorial decisions. JAMA
1994;272(2):137-8.
39. Joyce J, Rabe-Hesketh S, Wessely S. Reviewing the
reviews: the example of chronic fatigue syndrome.
JAMA 1998;280(3):264-6.
40. Link AM. US and non-US submissions: an analysis of
reviewer bias. JAMA 1998;280(3):246-7.
41. Jadad AR, Cook DJ, Jones A, Klassen TP, Tugwell P,
Moher M, et al. Methodology and reports of systematic
reviews and meta-analyses: a comparison of Cochrane
reviews with articles published in paper-based journals.
JAMA 1998;280(3):278-80.
42. Dickersin K, Fredman L, Flegal KM, Scott JD, Crawley
B. Is there a sex bias in choosing editors? Epidemiology
journals as an example. JAMA 1998;280(3):260-4.
43. Dickersin K. The existence of publication bias and risk
factors for its occurrence. JAMA 1990;263(10):1385-9.
44. Callaham ML, Wears RL, Weber EJ, Barton C, Young
G. Positive-outcome bias and other limitations in the
outcome of research abstracts submitted to a scientific
meeting. JAMA 1998;280(3):254-7.
45. Misakian AL, Bero LA. Publication bias and research on
passive smoking: comparison of published and
unpublished studies. JAMA 1998;280(3):250-3.
46. McNutt RA, Evans AT, Fletcher RH, Fletcher SW. The
effects of blinding on the quality of peer review. JAMA
1990;263(10, March 9):1371-76.
47. Black N, van Rooyen S, Godlee F, Smith R, Evans S.
What makes a good reviewer and a good review for a
general medical journal? JAMA 1998;280(3):231-3.
48. Callaham ML, Baxt WG, Waeckerle JF, Wears RL.
Reliability of editors’ subjective quality ratings of peer
reviews of manuscripts. JAMA 1998;280(3):229-31.
49. Cho MK, Justice AC, Winker MA, Berlin JA, Waeckerle
JF, Callaham ML, et al. Masking author identity in peer
review: what factors influence masking success? JAMA
1998;280(3):243-5.
50. Godlee F, Gale CR, Martyn CN. Effect on the quality of
peer review of blinding reviewers and asking them to
sign their reports: a randomized controlled trial. JAMA
1998;280(3):237-40.
51. Justice AC, Cho MK, Winker MA, Berlin JA, Rennie D.
Does masking author identity improve peer review
quality? A randomized controlled trial. JAMA
1998;280(3):240-2.
52. van Rooyen S, Godlee F, Evans S, Smith R, Black N.
Effect of blinding and unmasking on the quality of peer
review: a randomized trial. JAMA 1998;280(3):234-7.
53. van Rooyen S, Black N, Godlee F. Development of the
review quality instrument (RQI) for assessing peer
reviews of manuscripts. J Clin Epidemio
1999;52(7):625-9.
54. Armstrong JS. Peer review for journals: Evidence on
quality control, fairness, and innovation. Sci Eng Ethics
1997;3(1):63-84.
55. Weber EJ, Callaham ML, Wears RL, Barton C, Young
G. Unpublished research from a medical specialty
meeting: why investigators fail to publish. JAMA
1998;280(3):257-9.
56. Fletcher RH, Fletcher SW. Evidence for the
effectiveness of peer review. Sci Eng Ethics
1997;3(1):35-50.
57. Friedman PJ. Correcting the literature following
fraudulent publication. JAMA 1990;263(10, March
9):1416-19.
58. Garfield E, Welljams-Dorof A. The impact of fraudulent
research on the scientific literature. JAMA
1990;263(10):1424-27.
59. Pfeifer MP, Snodgrass GL. The continued use of
retracted, invalid scientific literature. JAMA
1990;263(10):1420-3.
60. Budd JM, Sievert M, Schultz TR. Phenomena of
retraction: reasons for retraction and citations to the
publications. JAMA 1998;280(3):296-7.
61. Parrish DM. Scientific misconduct and correcting the
scientific literature. Acad Med 1999;74(3):221-30.
62. Snodgrass GL, Pfeifer MP. The characteristics of
medical retraction notices. Bull Med Libr Assoc
1992;80(4):328-34.
63. Duggar DC, Christopher KA, Tucker BE, Jones DA,
14
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
Watson M, Puckett M, et al. Promoting an awareness of
retractions: the Louisiana State University Medical
Center in Shreveport experience. Medical Reference
Services Quarterly 1995;14(1):17-32.
Editors ICoMJ. Uniform requirements for manuscripts
submitted to biomedical journals. JAMA
1993;269:2282-6.
Flanagin A, Carey LA, Fontanarosa PB, Phillips SG,
Pace BP, Lundberg GD, et al. Prevalence of articles with
honorary authors and ghost authors in peer-reviewed
medical journals. JAMA 1998;280(3):222-4.
Hoen WP, Walvoort HC, Overbeke AJ. What are the
factors determining authorship and the order of the
authors’ names? A study among authors of the
Nederlands Tijdschrift voor Geneeskunde (Dutch
Journal of Medicine). JAMA 1998;280(3):217-8.
Shapiro DW, Wenger NS, Shapiro MF. The
contributions of authors to multiauthored biomedical
research papers. JAMA 1994;271(6):438-42.
Tarnow E. The authorship list in science: Junior
physicists’ perceptions of who appears and why. Sci Eng
Ethics 1999;5(1):73-88.
Drenth JP. Multiple authorship: the contribution of
senior authors. JAMA 1998;280(3):219-21.
Wilcox LJ. Authorship: the coin of the realm, the source
of complaints. JAMA 1998;280(3):216-7.
Health NIo. Guidelines for the Conduct of Research in
the Intramural Research Program at NIH. [Bethesda,
MD]: [National Institutes of Health]; 1997.
Waldron T. Is duplicate publishing on the increase? Br
Med J 1992;304:1029.
Barnard H, Overbeke AJ. [Duplicate publication of
original manuscripts in and from the Nederlands
Tijdschrift voor Geneeskunde]. Ned Tijdschr Geneeskd
1993;137(12):593-7.
Koene HR, Overbeke AJ. [The ultimate fate of articles
rejected for publication in the Nederlands Tijdschrift
voor Geneeskunde]. Ned Tijdschr Geneeskd
1994;138(49):2443-6.
Blancett SS, Flanagin A, Young RK. Duplicate
publication in the nursing literature. Image: The Journal
of Nursing Scholarship 1995;27(1):51-6.
Bloemenkamp DGM, Walvoort HC, Hart W, Overbeke
AJPM. [Duplicate publication of articles in the Dutch
Journal of Medicine in 1996]. Ned Tijdschr Geneeskd
1999;143(43):2150-3.
Tramer MR, Reynolds DJ, Moore RA, McQuay HJ.
Impact of covert duplicate publication on meta-analysis:
a case study. Br Med J 1997;315(7109):635-40.
Jefferson T. Redundant publication in biomedical
sciences: Scientific misconduct or necessity? Sci Eng
Ethics 1998;4(2):135-40.
Merton RK. A note on science and democrary. Journal
of Legal and Political Sociology 1942;1(1-2):115-26.
Jasanoff SJ. Is science socially constructed: And can it
still inform public policy? Sci Eng Ethics
1996;2(3):263-76.
Mahoney MJ. Publication prejudices: An experimental
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
15
study of confirmatory bias in the peer review system.
Cognitive Therapy and Research 1977;1:161-75.
Dickersin K, Min Y-I, Meinert CL. Factors influencing
publication of research results. Follow-up of
applications submitted to two institutional review
boards. JAMA 1992;267(3):374-8.
Campbell TID. Understanding the potential for
misconduct in university-industry relationships: An
empirical study. In: Braxton JM, editor. Perspectives on
scholarly misconduct in the sciences. Columbus, OH:
Ohio State University Press; 1999. p. 259-82.
Boyd EA, Bero LA. Assessing Faculty Financial
Relationships With Industry : A Case Study. JAMA
2000;284:2209-14.
Cho MK, Shohara R, Schissel A, Rennie D. Policies on
faculty conflicts of interest at US universities. JAMA
2000;284:2203-8.
Health NIo. Reminder and update: Requirement for
instruction in the responsible conduct of research in
national Research Service Award institutional training
grants. NIH Guide for Grants and Contracts 1992;21:23.
Mastroianni A, Kahn JP. Encouraging accountability in
research: A pilot assessment of training efforts.
Accountability in Research 1999;7:85-100.
Brown S, Kalichman MW. Effects of training in the
responsible conduct of research: A survey of graduate
students in experimental sciences. Sci Eng Ethics
1998;4(4):487-98.
Elliott D, Stern JE. Evaluating teaching and students’
learning of academic research ethics. Sci Eng Ethics
1996;2(3):345-66.
Self DJ, Wolinsky FD, Baldwin DC, Jr. The effect of
teaching medical ethics on medical students’ moral
reasoning. Acad Med 1989;64:755-9.
Baldwin DC, Jr., Daugherty SR, Self DJ. Changes in
moral reasoning during medical school. Acad Med
1991;69(9 Suppl):S1-3.
Self DJ, Schrader DE, Baldwin DC, Jr., Wolinsky FD. A
pilot study of the relationship of medical education and
moral development. Acad Med 1991;66(10):629.
Self DJ, Schrader DE, Baldwin DC, Jr., Root SK,
Wolinsky FD, Shadduck JA. Study of the influence of
veterinary medical education on the moral development
of veterinary students. J Am Vet Med
Assoc1991;198(5):782-7.
Self DJ, Baldwin DC, Jr., Wolinsky FD. Evaluation of
teaching medical ethics by an assessment of moral
reasoning. Med Educ 1992;26(3):178-84.
Self DJ, Schrader DE, Baldwin DC, Jr., Wolinsky FD.
The moral development of medical students: a pilot
study of the possible influence of medical education.
Med Educ 1993;27(1):26-34.
Bebeau MJ, Tohma SJ. The impact of a dental ethics
curriculum on moral reasoning. J Dent Educ
1994;58(9):684-92.
Self DJ, Baldwin DC, Jr., Wolinsky FD. Further
exploration of the relationship between medical
education and moral development. Camb Q Healthc
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Ethics 1996;5(3):444-9.
98. Anderson MS, Louis KS, Earle J. Disciplinary and
departmental effects on observations of faculty and
graduate student misconduct. In: Braxton JM, editor.
Perspectives on scholarly misconduct in the sciences.
Columbus, OH: Ohio State University Press; 1999. p.
213-35.
99. Clarke M, Chalmers I. Discussion sections in reports of
controlled trials published in general medical journals:
islands in search of continents? JAMA 1998;280(3):2802.
100. Junker CA. Adherence to published standards of
reporting: a comparison of placebo-controlled trials
published in English or German. JAMA
1998;280(3):247-9.
101. Begg C, Cho MK, Eastwood S, Horton R, Moher D,
Oking I, et al. Improving the quality of reporting of
randomized controlled tirals: the CONSORT statement.
JAMA 1996;276:637-9.
102. Pitkin RM, Branagan MA, Burmeister LF. Accuracy of
data in abstracts of published research articles. JAMA
1999;281:1129-30.
103. Pitkin RM, Branagan MA, Burmeister LF.
Effectiveness of a journal intervention to improve
abstract quality. JAMA 2000;283(4):481.
104. Jefferson T, Smith R, Yee Y, Drummond M, Pratt M,
Gale R. Evaluating the BMJ guidelines for economic
submissions: prospective audit of economic submissions
to BMJ and The Lancet. JAMA 1998;280(3):275-7.
105. Lock S, Wells F. Fraud and misconduct in medical
research. London: BMA Publishing Group; 1993.
106. Braxton JM, ed. Perspectives on Scholarly Misconduct
in the Sciences. Columbus, OH: Ohio State University
Press; 1999.
16
I. Norms and Environmental Issues
1. Students and Mentors
What Would Get You in Trouble: Doctoral Students’ Conceptions of Science
and Its Norms
Melissa S. Anderson, Educational Policy and Administration, University of Minnesota, USA
Key Words: Graduate education, Graduate students, Norms
Undergirding the academic enterprise is a web of assumptions about how the members of the
academic community should conduct their professional lives. These assumptions are expressed in
ways ranging from the most explicit directives (legal, institutional, contractual) to the implicit, takenfor-granted understandings that facilitate everyday interactions among members of the profession.
They constitute the normative underpinnings of the academic profession.
Braxton and Bayer define norms as “shared beliefs within a particular social or professional
group about behavior expected or desired in a given situation or circumstance” (1). In the academic
context, the four norms that Robert Merton (2) identified in his 1942 analysis–universalism,
communality [to use Barber’s (3) term], disinterestedness, and organized skepticism–have framed
much of the subsequent research. They figured prominently in Zuckerman’s seminal analyses of the
social system of science (4, 5). They are also reflected in Mitroff’s (6) “counternorms”, and they
together capture most of the considerable literature that Braxton (7) compiled on the subject of norms.
Others, however, have argued for a more complex understanding of norms. Mulkay, for example,
has claimed that norms are best understood as ideologies or “evaluative repertoires” (8). That is,
norms constitute a kind of standardized narrative that academics use to describe and evaluate
behavior and to prescribe responses to certain behaviors (8). Ajzen and Fishbein have described the
significance of “subjective norms” that reflect what others, who are important to an individual, think
he or she should do (9). From this perspective, neither an abstract normative system or an
individual’s own internalized norms are as important as the individual’s understanding of others’
expectations. Finally, Braxton and Bayer have demonstrated how a combination of inductive and
survey-based strategies could uncover a complex set of norms in collegiate teaching (1).
The present study takes a different approach to the norms of the academic profession, with
corresponding implications for the design of the study. First, it emphasizes the implicit over the
explicit, on the assumption that implicit norms can be particularly powerful in shaping behavior. This
study therefore relies on narrative descriptions of norms, instead of on a particular formulation of the
normative structure of academia. It is rooted in the proposition that more attention needs to be paid to
understanding science and its ethical aspects from the “inside out,” that is through the experiences of
scientists themselves (10-12). It therefore responds to Braxton’s call for study of norms “expressed in
the words of the respondents rather than in a priori definitions of possible norms” (7).
Second, it assumes that norms of a group are particularly salient to newcomers during a
socialization period (13). The data for this study accordingly come from first-year doctoral students,
who are encountering professional norms in intensive ways. Their experiences are likely to produce
Corresponding author: Melissa S. Anderson, Educational Policy and Administration, 330 Wulling Hall, University of
Minnesota, Minneapolis, MN, 55455, 612-624-5717 (voice), 612-624-3377 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
“contrast” in the gestalt sense through the process
of “sense-making”, which highlights the
normative insights they acquire (14).
Third, the study assumes no necessary match
among students’ understanding of the broad
norms of the academic profession, the norms that
they have internalized and view as most salient,
and the behavior of professional colleagues. This
study therefore explores levels of consonance
and dissonance that students perceive among
these three phenomena.
Fourth, this study relies on Durkheim’s
useful proposition that norms are recognized
when they are violated (15). The questions used
in this study to elicit students’ views of norms,
therefore, ask students to contrast their views of
general academic norms, as well as the norms to
which they subscribe, against the behavior of
their colleagues.
Methods
These parameters gave shape to the current study,
which is part of a broader project on doctoral
education, the Academic Life Project, funded by
the National Science Foundation (Grant number
9408S08622). Participants for the current
analysis were 30 first-year doctoral students in
seven science and social science disciplines at a
major research university. (The project will
eventually involve over 100 interviewees and
will be longitudinal.) Semi-structured interviews
of approximately a half-hour yielded narrative
data on norms and related topics.
A series of questions in the interviews asked
students to consider and comment on
relationships between academic norms and
behavior (Do you see any conflicts between what
people think or say you should do and the way
work is actually done?), between their own
perspectives and behavior (Do you see people
around here acting contrary to your advice [to
doctoral students on how to avoid serious
mistakes]?), and between their own normative
perspectives and academic norms (Are there any
ideas or rules about how you should do your
work that you don’t agree with?). These
questions highlighted students’ understandings of
academic research as a social enterprise whose
membership they are entering. Those who
articulated a more complex normative
perspective showed greater awareness of the
social aspects of the scientific enterprise and a
more constructivist approach to knowledge
development in the sciences. They were also less
troubled by dissonance between behaviors and
norms, recognizing the inevitable roles played by
mistakes, errors of fact and of judgment, and
mid-course corrections.
Results
Students’ conceptions of norms that underlie
their work are presented here in terms of the
three contrasts identified above. First, students’
conceptions of general academic norms are
described in light of the behavior of their
colleagues. Then the norms to which they
subscribe are seen in contrast, again, to
colleagues’ behavior. Finally, what they
understand to be academic norms are contrasted
to their own normative orientations.
Correspondence between academic norms
and behavior. The first comparison investigated
is between students’ conceptions of the norms of
their fields and the behaviors of those around
them. The interview question was, “Do you see
any conflicts between what people think or say
you should do and the way work is actually
done?”
Approximately two-thirds of those
interviewed saw no conflict between prescribed
and actual behavior among their colleagues.
Most saw no disjuncture; a few were more
definite: “No, I mean, an emphatic no with the
faculty,” and, “They’re pretty straightforward,
and they’ll pretty much hold true to their word.”
Two students noted that, while they were not
aware of conflict between norms and action, they
did not really know enough about what people
were doing in the department to comment
generally about people’s behavior; as one put it,
“I’m not privy to a lot of the goings on of the
department.”
Five students noted particular areas of
disjuncture between norms and behavior. One
mentioned safety rules:
We all have to go to this safety training before
we are allowed to go in the lab. It’s just kind
of a refresher course every year. And then ...
they always say practically nothing is supposed
to go down the drain. And sometimes stuff
does. But we’re not even supposed to put things
like ... a simple rinsing agent down the drain ...
but it happens all the time.
This student went on to affirm the importance of
the safety rules for two reasons: first, that safety
supports proper procedures (“if you don’t do it
right, it doesn’t work”), and second, not
following these rules is dangerous (“if you don’t
follow the rules in terms of safety, in terms of
20
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Anderson, What Would Get You in Trouble
in the interviews is between the norms to which
students themselves subscribe and the behavior
of their colleagues. Here the question is whether
or not students see people around them acting
contrary to the way the students think they
should act. Employing Durkheim’s view that
norms are best recognized when violated, the
interview protocol invited students to consider
what they would advise incoming doctoral
students to do to stay out of trouble in their work
(15). Responses demonstrate students’
personally held beliefs about how first year
students should act, identified here as subscribed
norms. Students were then asked, as follow-up
questions, “Do you see people around here acting
contrary to your own advice? What are they
doing?”
Responses to these questions fall into three
general categories: tasks, relationships, and
ethics. Most of the responses addressed the work
of graduate students. Several talked about the
need for students to take responsibility for their
own work and progress. As one put it, “I mean,
in our department, it’s a problem both with the
students not taking the initiative to getting all of
their requirements and prelims done and also,
with our department, no one says anything if it
takes you longer.” Others disapproved of student
colleagues’ not getting their work done, taking
too much time to get through their work, or
abandoning the work altogether. All of these
students clearly demonstrated a strong
commitment to hard work and a sense that some
others around them acted contrary to this
subscribed norm.
Not only do students believe in getting the
work done, but several mentioned the need to do
independent work. One science student
complained,
correct procedure, usually that means that the
chemist should not work in the lab”).
A second point of conflict observed by a
student is in admissions procedures for the
graduate students in the department. From the
vantage point of a place on the departmental
graduate admissions committee, the student saw
that, though the department touts a highly
selective admissions policy, the process is
influenced in political ways by influential
individuals on the faculty. The result is that the
department admits less-qualified people than its
policy would suggest.
The third area of dissonance between
prescribed and enacted behaviors is in research.
One psychology student focused on experiments:
We talk a lot about being a very experimental
field and it’s all about experiments, but it’s so
difficult to run experiments now with getting
through the IRB [Institutional Review Board]
and getting subjects.... [I]t’s so much easier to
pass out some sort of survey or some sort of
questionnaire. And so we talk about the
experiment and how wonderful it is, and then
we don’t do it.
Two other students also mentioned research, but
in a different way. They clearly understood the
faculty’s focus on research, but they did not see
faculty providing enough support to students to
get them started on their own research. As one
put it, “I think [it’s] the absence of direction
which is noticeable, which stands out. And I
think some students have felt ... you know,
they’re sort of cast adrift, in some sense, and left
to figure everything out for themselves.” The
other student described her frustration with the
research imperative in light of the same kind of
lack of direction:
There almost seems like there’s kind of pressure
or an expected norm within the department
itself that we get involved with research. Yet,
in our specific discipline, in our area, there
hasn’t been very much guidance or, you know,
pressure to do that.... I have met with my
advisor twice on my own volition — and going
to see her and saying, “Okay. Maybe it’s time
for me to get involved in research,” and each
time she has not had a specific project that
really had any place for me to start.... And I
just kind of walked away from it feeling like,
just thinking that she had just so much going
on already — and really, you know, like almost
I kind of felt like I would be a burden to get
involved at that point.
I think one of the biggest mistakes that they
could make is to do something that is not
independent. I see a lot of people that are
working with their advisors and really, ... I don’t
know the best way to describe this without
sounding mean, but they just have no interest
of their own. They are just a, like a little offshoot of their advisor, like a little worker....
They’re not independent at all.... You know,
what they do is what their advisor says, and I
think that’s a really big mistake, because one
day you can look back and be, like, “Oh. This
isn’t what I wanted to do at all, and if I had the
choice I would have done it completely
differently.”
Correspondence between subscribed norms
and behavior. The second comparison addressed
Taking the initiative for an independent stream of
21
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
inquiry is a step beyond responsibility for getting
work done, a step that some, but not all, first-year
graduate students take.
One student’s story about a graduate-student
peer illustrates her struggle with maintaining her
independence in inquiry. The peer in question is
someone she respects.
But the problem is, he comes from a different
undergraduate background, not an American
system. He’s from a different country, where
being the best in the class was very much
recognized and very much rewarded, and so
he was the best in his class. And so he came
here.... Everyone has been asking him for help,
and so he would do all of his work way in
advance — which was commendable — but
then he would — instead of working and taking
other people’s suggestions and trying to
integrate everything when we were working on
problem sets — he would be, like, “This is
right. I have the answer.” And usually he did.
Usually he was right. But it was annoying to
work with him.... There were times where even
though I knew I would probably get a better
grade if I worked with him, because he would
have the answers, I wouldn’t want to do it. And
also, you don’t want the answers given to you.
Comments about relationships comprise the
next category of responses about the contrast
between subscribed norms and behavior.
Students demonstrate clear ideas about how
people should behave toward each other in the
graduate school setting. Some mentioned the
importance of having an advisor with whom the
student can work. They described examples of
advisors who were not supportive of their
students. This behavior that ran contrary to their
beliefs about how advisors are to act met with
very strong negative reactions.
Other respondents showed a keen sense of
the importance of making a good impression and
expressed dismay that some of their peers did not
appear to understand this point. A science
student said,
I know there’s some people who, whenever
there was an exam, they just didn’t go into the
lab all the time, and I don’t think it left a good
impression on some people who were working
in the lab, working around them.... So if you
don’t seem very serious about your lab work,
then they — someday when you have to go to
them for advice or something — they’re not
necessarily drawn to give you as much time
and make as much of a serious effort.
Another student described impressionmanagement in blunt terms as a quid pro quo:
I guess, just like, you have to do things for
people so they’ll do things for you later. I guess
that doesn’t even sound that bad. But more
like — I can’t think of a particular example —
but just basically doing things that you don’t
want, because you know later it’ll get you
something you do want.
Not only are students aware of the work
imperative, but they are also aware of the need
for others to know that they subscribe to it. As
the quotes illustrate, the norm bears both sanction
and reward. This norm illustrates students’
movement toward full acceptance into the
academic social world.
The third contrast between behavior and
students’ own normative orientations was in the
area of ethics. Those who mentioned ethics said
that they had seen no instances of people acting
contrary to what they themselves understood to
be appropriate behavior. One said, “I’ve never
seen anyone falsifying data, which is very, very
good. And I believe that we don’t have the
second problem, fishing for data. At least in my
group, we don’t have that.” Another noted, “I
haven’t seen, I haven’t heard of anybody lying
about stuff or trying to falsify results.” This
science student went on to describe how
important it is for students to acknowledge
mistakes, so that they are not interpreted as more
serious offenses: “Everybody makes mistakes....
Everyone’s pretty understanding of when your
experiments don’t work or when you did a stupid
mistake or whatever.”
The normative understandings that the
doctoral students reveal through their comments
on the contrast between what peers should do and
what they are actually doing thus center largely
on their work and their relationships with
colleagues. That is, they appear attuned to both
functional and social norms of academic life.
The next step is to contrast their own normative
orientations to what they perceive to be the
general norms of their fields.
Contrast between academic norms and
subscribed norms. Students’ perceptions of
prevalent academic norms may not match their
own ideas about how they should conduct
themselves in the academic world. As both
academic norms and subscribed norms can be
brought into focus by contrasting them against
behavior, so they can be clarified by comparing
them to each other. The relevant question on the
interview protocol was, “Are there any ideas or
rules about how you should do your work that
you don’t agree with?”
22
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Anderson, What Would Get You in Trouble
Beyond students’ attention to task-related
disjunctures between academic and espoused
norms, the most striking pattern in students’
responses is their uncertainty about academic
norms in general. Most of them are keenly aware
that norms vary by discipline or even from one
research group to another. For example, one
noted, “Everyone has such different views about
how to do things.” Another put it this way:
“Each professor sort of has their own research
policy. And that’s academia. They have the
freedom to make up the rules of their group,
within certain bigger boundaries that the school
sets.” Yet another respondent said, “I don’t think
there are very many rules about how we should
conduct our research, other than the whole basic
‘Be ethical and stuff.’ I don’t observe very many
rules about how we should conduct the research.”
This student went on to mention that she might
change her mind as she got further into her
research, when she would have to remember all
the rules about where to put commas — thereby
illustrating just how far she had to stretch to
think of general norms of the field.
Perhaps some of the uncertainty that students
expressed about academic norms is related to the
ways in which such norms are communicated.
The student quoted above who mentioned each
professor having his or her own research policy
went on to say, “Ideally, it should be talked about
as a research group as a whole, but it seems to
me that a lot of stuff is just sort of telephone,
where one person tells another person, and that
person tells the next person.” Another talked
about his reluctance to ask people how things
should be done in the lab:
The task-related points of disjuncture fell
generally in the category of competition and its
attendant work pressures. A student in a social
science department commented,
Everyone’s competing for jobs in academic
environments primarily.... And I guess what that
means for many students is they have to adapt
to a competitive type of atmosphere and in
some cases be more competitive than they
would like to be in order to further their goals
further on. And I think that might be
disheartening for some students.... And I think
all of the students ... try to be good-natured
about the entire thing, but I think the pressure
of continuing to get as many publications as
you can is the reality that dawns on a lot of
students — something they didn’t anticipate,
necessarily, early on.
Another student talked about competitive
pressures to publish in terms of “the whole
production thing” and the “assembly line
production attitude.”
Several students complained about the work
loads they bear in terms of the mismatch between
their professors’ views on how much work they
should do and their own. A science student
talked about peers who never take time off and
“work themselves to death” to live up to what
they perceive as the standards of work in the
field; the student said he would never do that.
Another commented on prevalent norms for the
quality of a dissertation. In this students’
relatively new field in science, it was generally
expected, 10 or 20 years ago, that each
dissertation would open up a completely new
field of inquiry; now, the expansion of the
discipline and the far greater competition due to a
more crowded field make it much harder to have
such an impact through doctoral work. The
student noted, though, that normative
understandings in the field had not changed in
response.
Another point of contrast related to
competition is the matter of independent work.
Several students mentioned that at least some of
their professors require independent, as opposed
to collaborative, work on assignments in
graduate courses. Many of the students were
previously socialized to collaborative norms, and
they found the professors’ insistence on
individual work counterproductive. Here
students’ normative orientations run counter to
the academy’s norms of rewarding people on the
basis of individual achievement and independent
contributions.
The approach towards how you learn your way
around the lab is you just go in there and you
do it. As far as being taught or having anyone
specifically to show you around, you really
don’t, because everyone in there is really, really
busy, because they are doing research. And
they don’t want to take time out of their
research to show you how to work [a machine],
because it’s such a simple thing to them, and
they get really frustrated and impatient with
someone who is just learning how to use it.
And so, generally you just have to go in there
and learn on your own.... I almost felt afraid to
go to other people in the group with my stuff,
because I don’t want to waste their time and I
don’t want to feel stupid either.
Of course, some students were unable to identify
any dissonance between the norms to which they
subscribe and the more general academic norms
23
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
as they see them. One person wryly commented
on the thoroughness of his own socialization to
the general normative structure of the field:
“Maybe I’ve been just so well trained that I don’t
know anything anymore.”
The results in this section show, as did the
earlier results, that students’ normative
conceptions are dominated by functional or taskrelated norms. They also show a general
awareness among students of social norms,
though their conceptions of norms for
interpersonal relations are not as fully developed
as their views on functional norms.
Discussion
The findings presented here contribute to our
understanding of doctoral students’ initial
normative orientations. Students’ conceptions of
normative imperatives are relevant to policy
initiatives that are currently receiving a great deal
of attention. The federal Office of Research
Integrity recently announced a major new
initiative that will focus on the promotion of the
responsible conduct of research. The American
Educational Research Association is currently
preparing to publish a book that will direct
attention to the AERA Code of Ethics and its use.
Dozens of other academic associations are
writing or revising their codes of ethics, and
virtually every major research university has
addressed its institutional policies on ethics and
misconduct in the past five years. The federal
government is seeking to expand its requirements
for formal training in ethics beyond those for
trainees covered by National Institutes of Health
funding. Most of the attention to expanded
training in ethics and related issues focuses on
graduate students and other newcomers to the
academic profession.
Continued self-regulation by the scientific
community depends on the ongoing renewal of
normative conceptualizations that, through their
generational evolution, continue to reflect the
expectations of society for science. Most of the
emerging initiatives are driven, however, by a
sense of urgency or by federal regulations and
directives, without attention to doctoral students’
understanding of science, academic life, and the
norms of their disciplines. Neither do they
reflect ways in which newcomers interact with
and shape the normative bases of their fields
(16).
This study serves as a window onto the
normative assumptions of science, but it
furthermore suggests ways in which those norms
can be communicated within and beyond the
scientific community (17, 18). The doctoral
students interviewed reveal the norms of science
as they understand them, during a period when
they are intensely and reflectively face-to-face
with the way science works. They are the future
membership of the scientific community, but they
are also current participants in the enterprise,
struggling with their own ideas of how they
should behave as scientists.
The results of the interviews demonstrate
intriguing patterns of dissonance among the three
phenomena examined. The interview responses
show that students’ normative conceptualizations
are dominated by functional (task-related) norms,
as we might expect from earlier work on
anticipatory socialization that emphasizes
survival in the graduate or professional-school
setting (16). Augmenting the functionalist
perspective, however, are emergent
conceptualizations of social and ethical norms.
The inchoate nature of first-year students’
personal normative orientations suggests that
approaches to socialization of doctoral students
to academic life, particularly in the areas of
ethics and related issues, may overestimate the
extent of students’ understanding of the academic
system, the nature of research, and the place of
individual academics in the broader context of
research. Students interviewed here showed very
little awareness of their disciplines, beyond their
own work, or of the higher education system,
beyond their own departments. The imperatives
they identified have to do generally with the
work at hand and the people with whom they
interact.
Socialization to the field and to the
normative bases of research in a discipline should
be grounded in the academic world with which
these students are familiar, while at the same
time introduce them to the broader academic
environment. The theme of individual,
independent work that runs through these
interviews suggests that students might not be
subject to as much osmotic group socialization as
many faculty assume. It is also clear that the
channels by which socialization to the normative
aspects of academic life are communicated are
primarily informal. Calls for more formal, more
deliberate approaches to normative socialization
find support in the vagueness with which
students conceptualize the norms that underlie
academic research.
24
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Anderson, What Would Get You in Trouble
Aknowledgements
17. Lapidus JB., Mishkin B. Values and ethics in the
graduate education of scientists. In: Nerad M, June R,
Miller DS, editors. Graduate education in the United
States. New York: Garland; 1997.
18. Bird SJ. Including ethics in graduate education in
scientific research. In: Braxton JM, editor. Perspectives
on scholarly misconduct in the sciences, Columbus
(OH): Ohio State University Press; 1999.
The author gratefully acknowledges the research
assistance of Janet M. Holdsworth and Jennifer
Milleville.
Bibliography
1. Braxton JM, Bayer AE. Faculty misconduct in collegiate
teaching. Baltimore: The Johns Hopkins University
Press; 1999.
2. Merton RK. A note on science and democracy. J of
Legal and Political Sociology 1942; 1(1-2):115-126.
3. Barber B. Science and the social order. New York: Free
Press; 1952.
4. Zuckerman HE. Deviant behavior and social control in
science. In: Sagarin E, editor. Deviance and social
change. Beverly Hills (CA): Sage; 1977. p. 87-138.
5. Zuckerman HE. The sociology of science. In: Smelser
NJ, editor. Handbook of sociology. Newbury Park (CA):
Sage; 1988. p. 511-574.
6. Mitroff I. Norms and counter-norms in a select group of
the Apollo moon scientists: A case study of the
ambivalence of scientists. Amer Sociological Rev 1974;
39, 579-95.
7. Braxton JM. The normative structure of science: Social
control in the academic profession. In: Smart JC, editor,
Higher education: Handbook of theory and research, v.
2. New York: Agathon Press; 1986.
8. Mulkay, M. Norms and ideology in science. Soc Sci
Info 1976; 15(4-5): 637-56.
9. Ajzen I, Fishbein M. Understanding attitudes and
predicting social behavior. Englewood Cliffs (NJ):
Prentice-Hall; 1980.
10. Anderson MS. (1999). Uncovering the covert:
Research on academic misconduct. In: Braxton JM,
editor. Perspectives on scholarly misconduct in the
sciences. Columbus (OH): Ohio State University Press;
1999.
11. Knorr-Cetina KD, Mulkay M, editors. Science observed:
Perspectives on the social study of science. Beverly
Hills (CA): Sage; 1983.
12. Latour B, Woolgar S. Laboratory life: The construction
of scientific facts. Princeton, NJ: Princeton University
Press; 1986.
13. Wanous JP. Organizational entry: Recruitment,
selection, orientation, and socialization of newcomers,
2nd ed. Reading (MA): Addison-Wesley; 1992.
14. Louis MR. Surprise and sense making: What
newcomers experience in entering unfamiliar
organizational settings. Admin Sci Q 1980; 25(2):226251.
15. Durkheim, E. The elementary forms of religious life.
Fields KE, translator. New York: Free Press;
1995[1912].
16. Tierney WG., Rhoads RA. Enhancing promotion,
tenure, and beyond: Faculty socialization as a cultural
process. ASHE-ERIC Higher Education Report, no. 936. Washington, D.C.: The George Washington
University; 1993.
25
Data Manipulation in the Undergraduate Laboratory: What are we
teaching?
Elizabeth W. Davidson, Department of Biology, Arizona State University, USA
Heather E. Cate, College of Liberal Arts and Sciences, Arizona State University, USA
Cecil M. Lewis, Jr., Department of Anthropology, University of New Mexico, USA
Melanie Hunter, Department of Biology, Arizona State University, USA
Keywords: Biology, Cheating, Chemistry, Motivation, Undergraduate research
Arizona State University (ASU) offers a senior-level course entitled “Professional Values in Science”
that addresses a number of topics concerning ethical conduct of research as well as ethical concerns at
the intersection of science and society. The course demands active participation by the students.
Several years ago, on his own initiative a student in the class developed a questionnaire that explored
data manipulation. As most of the students were undergraduates, the questionnaire focused upon
manipulation of data in undergraduate science laboratories. We were startled to discover that over
60% of the students openly admitted to manipulation of data in undergraduate laboratories. These
results led to the development of a more elaborate survey that has been administered to 7
undergraduate Biology and Chemistry courses, enrolling a total of over 700 students. The courses
include both major and nonmajor subjects, at both introductory and upper division level. Arizona
State University has approximately 50,000 students, including (in academic year 2000) ca. 1000
majors in Biology and 250 majors in Chemistry. In the fall semester, 2000, 3137 undergraduates are
enrolled in Biology courses, while 3355 undergraduates are enrolled in Chemistry courses.
Laboratories are therefore limited in available time, are generally supervised and graded by graduate
teaching assistants, and many, but not all, of these courses rely upon traditional laboratory exercises.
Methods:
The survey and instructions to students are presented in at the end of the paper. Students were
advised by the person administering the survey (who was not their course professor or teaching
assistant) that the results would be held anonymous and would not affect their grade. The courses
included Chemistry 115: Introductory, non-majors; Chemistry 335: Organic, non-majors; Biology
201: Anatomy and Physiology, non-majors; Biology 100: Introductory, non-majors; Biology 182:
Introductory, majors; Biology 193: Introductory, majors, critical thinking focus; Biology 385:
Invertebrate Zoology, majors. Seven hundred and two students participated. Institutional Human
Subjects committee approval was obtained. Data were analysed by Spearman correlation.
Corresponding author: Elizabeth W. Davidson, Department of Biology, Arizona State University, Box 871501, Tempe, AZ
85287-1501, 480-965-7560 (voice), 480-965-2519 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
100%
Percent Data Manipulation in Other Courses
Percent Data Manipulation in This Course
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
never
seldom
sometimes
often
almost always
0%
CHM 115 CHM 335 BIO 201
BIO 100
BIO 182 BIO 193
90%
80%
70%
60%
50%
never
seldom
sometimes
often
almost always
40%
30%
20%
10%
0%
CHM 115 CHM 335 BIO 201
BIO 385
Course Prefix and Number
Results
The key question in this survey was Question 5,
“Have you ever manipulated data in this course?”
As shown in Figure 1, between 40.4 and 75% of
students in the surveyed course admitted to
manipulating data “almost always,” and another
20-43.9% admitted to such manipulation “often.”
Students reporting data manipulation “seldom”
represented less than 5% of those surveyed, and
only one student out of over 500 who responded
to this question replied “never.” Using
correlation analysis, we learned that admission of
manipulation in the course surveyed was strongly
correlated to admission of manipulation in other
courses (Spearman Correlation Sig. (2 tailed)
0.355, significant at 0.01 level) (Figure 2).
We asked whether data manipulation was
related to the level (i.e. introductory vs.
advanced) of the course, and whether the course
was designed for majors or non-majors. No
significant difference was found between data
manipulation in Introductory Biology BIO 100
(non-majors) and BIO 182 (majors) or between
these lower division courses and an upper
division course, BIO 385 (Invertebrate Zoology,
majors). We compared responses from BIO 182,
a traditional introductory course, to BIO 193, an
introductory majors course with emphasis on
critical thinking. The smallest percentage of
students reporting data manipulation “almost
BIO 182
BIO 193
BIO 385
Figure 2. Results of survey, Question 10, “Have you ever
manipulated or made up data in any other science course?”
CHM 115, N=87; CHM 335, N=52; BIO 201, N=27; BIO
100, N=81; BIO 182, N=40; BIO 193, N=57; BIO 385,
N=66. N= total number of responses to the specific
question.
always” was in BIO 193, however a large
proportion of the remainder reported
manipulation “often” (Figure 1). Within the two
non-majors chemistry courses surveyed, less data
manipulation was found in CHM 115
(Introductory) than in CHM 335 (Organic), and
indeed the highest overall reported manipulation
(90.5% “almost always” or “often”) was reported
in Organic Chemistry. Conversations with
students in the Professional Values in Science
class and elsewhere confirmed that many have
Percent Data Manipulation observed in this course
Figure 1. Results of survey, Question 5, “Have you ever
manipulated data or made up data in this course?” CHM
115: Introductory, non-majors, N=86; CHM 335: Organic,
non-majors, N=44; BIO 201: Anatomy and Physiology, nonmajors, N=29; BIO 100: Introductory, non-majors, N=200;
BIO 182: Introductory, majors, N=40; BIO 193:
Introductory, majors, critical thinking focus, N=57; BIO
385: Invertebrate Zoology, majors, N=64. N= total number
of responses to the specific question.
BIO 100
Course Prefix and Number
100%
90%
80%
70%
60%
50%
40%
30%
never
seldom
sometimes
often
almost always
20%
10%
0%
CHM 115 CHM 335 BIO 201
BIO 100
BIO 182
BIO 193
BIO 385
Course Prefix and Number
Figure 3. Results of survey, Question 7, “Have you ever
observed anyone else manipulate or make up data in this
course?” CHM 115, N=91; CHM 335, N=67; BIO 201,
N=28; BIO 100, N=237; BIO 182, N=40; BIO 193, N=56;
BIO 385, N=66. N= total number of responses to the
specific question.
28
–––––––––––––––––––––––––––––––––––––––––––––– Davidson, et al., Data Manipulation in the Undergraduate Laboratory
Percent Data Manipulation for Better Grade, This Course
100%
90%
Percent Data Manupulation
80%
70%
60%
50%
40%
30%
20%
never
seldom
sometimes
often
almost always
10%
0%
CHM 115 CHM 335 BIO 201
BIO 100
BIO 182
BIO 193
BIO 385
Course Prefix and Number
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
never
seldom
sometimes
often
almost always
0%
CHM 115 CHM 335 BIO 201
BIO 100
BIO 182
BIO 193
BIO 385
Course Prefix and Number
Figure 4. Results of survey, Question 14, “Have you ever
observed anyone manipulate or make up data in any science
course?” CHM 115, N=94; CHM 335, N=70; BIO 201,
N=30; BIO 100, N=96; BIO 182, N=39; BIO 193, N=55;
BIO 385, N=66. N= total number of responses to the
specific question.
Figure 5. Results of survey, Question 6, “If you have ever
manipulated data or made up data, was it motivated by the
thought of a better grade?” CHM 115, N=69; CHM 335,
N=41; BIO 201, N=17; BIO 100, N=246; BIO 182, N=31;
BIO 193, N=55; BIO 385, N=53. N= total number of
responses to the specific question.
manipulated data for Chemistry laboratory
reports, particularly in Organic. Little difference
in data manipulation (Question 5) was found
when analyzed by academic year or by gender.
Two other key questions were 7 and 14,
which asked whether the student had observed
others manipulating data. The results from these
questions were less consistent than responses
about the students own data manipulation. Two
courses (CHM 115 and BIO 201) received an
“almost always” response rate of 100%,
whereas in other courses a much smaller
proportion of students responded “almost
always” (Figures 3, 4).
We investigated motivation for data
manipulation with questions 6 and 11, which
asked whether the students manipulated data
in order to get a better grade. Up to 100% of
students in some courses replied that
manipulation was almost always performed to
obtain a better grade (Spearman Correlation
Sig. (2-tailed) 0.265, significant at 0.01 level)
(Figure 5; data from Question 11 not shown).
When asked whether this was because they
felt that their grade depended heavily upon the
experimental results (Questions 8 and 15), less
than half of students felt that their grade in the
current course depended on experimental
results “almost always”, and from 3.0 to
13.6% of the students replied to Question 8
that their grade “seldom” depended on results
(Figure 6, data from Question 15 not shown;
Spearman (Figure 7, data from Question 16
not shown; Spearman correlation 0.368,
significant at 0.01 level). Finally we surveyed
student preferences for type of laboratory
experiments (Question 17). In all seven courses
combined, only 1.7% of students preferred lab
experiments which place more emphasis on
results, whereas 53.5% preferred more emphasis
to be placed upon processes, and 44.7% preferred
a balanced combination of both techniques
(N=503).
Percent Data Manipulation because Grade
Depends on Experimental Results
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
never
seldom
sometimes
often
almost always
0%
CHM 115 CHM 335 BIO 201
BIO 100
BIO 182
BIO 193
BIO 385
Course Prefix and Number
Figure 6. Results of survey, Question 8, “How often have
you felt as though your grade in this course depended
heavily on your experimental results?” CHM 115, N=102;
CHM 335, N=66; BIO 201, N=35; BIO 100, N=218; BIO
182, N=40; BIO 193, N=58; BIO 385, N=66. N= total
number of responses to the specific question.
29
Class Places too much Emphasis on Experimental Results
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
100%
90%
80%
70%
60%
50%
40%
never
seldom
sometimes
often
almost always
30%
20%
10%
0%
CHM 115 CHM 335 BIO 201 BIO 100 BIO 182 BIO 193
BIO 385
Course Prefix and Number
Figure 7. Results of survey, Question 9, “Do you believe this
course places too much emphasis on experimental results
rather than on the processes used to get the results?” CHM
115, N=98; CHM 335, N=67; BIO 201, N=27; BIO 100,
N=194; BIO 182, N=40; BIO 193, N=58; BIO 385, N=66.
N= total number of responses to the specific question.
Discussion:
Some precautions should be taken in interpreting
these findings. First, the survey was limited to
only 7 courses at a single University, which in
each case were surveyed only once. We intend to
survey additional courses at ASU, and hope to
include at least one other large university and a
small liberal arts college in our analysis. Second,
the survey relies on self reporting. Some of the
students did not choose to answer all questions in
the survey. The total number responding to each
question in each course is presented in the figure
caption. Approximately 25% of the students
chose not to answer Question 5, for example.
Third, the construction of the questions did not
permit us to investigate motivations other than
that the student felt his/her grade depended upon
the experimental results (Questions 8, 9, 15 - 17).
Finally, even though students were given a clear
definition of “data manipulation” at the
beginning of the survey, it is possible that some
may not have clearly understood the definition of
“data manipulation.”
With the above caveats in mind, our results
show a very strong tendency among
undergraduate science students to manipulate or
make up data when writing laboratory reports. As
high as these percentages are, they are similar to
results observed in surveys of cheating on tests,
which Cizek has described as “remarkably and
uniformly high” (1). In surveys taken from 1970
to present, from 42% to over 90% of students
reported cheating on tests by self or others
(reviewed by Cizek, (1)). Out of 6000 students
in 31 universities surveyed by Meade, 67% of
science majors reported cheating on tests (2).
Most surveys of college test cheating ask only
whether the student has ever cheated. Our survey
expands this question to evaluate how frequently
data manipulation in laboratory reports occurs,
allowing us to differentiate between occasional
events and habitual cheating. Although there are
many studies of cheating on college tests, to our
knowledge our study is unique in examining
manipulation of laboratory data by
undergraduates.
Data manipulation apparently does not
diminish as the students progress to upper
division courses or from non-major to major
courses. Commitment to a major subject,
presumably because the student intends to
continue in this area of science for all or part of
his/her professional career, apparently does not
diminish this practice.
These results raise some important questions,
which include: How can this data manipulation
be reduced or eliminated? What are the
implications of data manipulation in the
undergraduate laboratory to the future careers of
these students? In other words, when do the
students stop manipulating data? In graduate,
professional or medical school? When they
begin doing “real” research? When the research
is published?
In response to the first of these questions, the
faculty and the system itself must take significant
responsibility. Faculty must recognize that this
data manipulation occurs, and not turn a blind
eye to this practice. We must examine the reason
why we require laboratory reports in the first
place, and whether there is another method of
assessing whether the student has learned the
necessary laboratory skills. Numerous laboratory
manuals are structured to provide “cook book”
procedures in which students are expected to
verify known biological, chemical, or physical
laws (3). However, these verification exercises
give students a false notion of the deductive
investigative process. They begin their training
with the preconceived notion that a “right”
answer exists and should be found. They are
therefore willing to adjust their laboratory results
for fear that the “wrong” answer would affect
their grade (4).
30
–––––––––––––––––––––––––––––––––––––––––––––– Davidson, et al., Data Manipulation in the Undergraduate Laboratory
balance between process and results.
The second concern, whether undergraduates
continue data manipulation as professional
scientists, has even greater implications. In the
frequently-cited study by Swazey et al., 72% of
graduate students and 59% of faculty reported to
have observed or had direct evidence of some
form of scientific misconduct (6). Data
falsification, however, was reported by a much
smaller proportion of respondents, ranging from
2% to 20%. Apparently, then, data manipulation
does decrease when the student becomes a
“professional” and becomes dedicated to the
science.
Over the last 5 years approximately 400
undergraduates at ASU have been engaged in
research programs funded by the Howard Hughes
Medical Foundation, the National Institutes of
Health and the National Science Foundation, in
addition to individual faculty grants.
Conversations with these students reveal that
once the research becomes “their own” and
important to them personally, they have far less
motivation to manipulate data, particularly if they
have a respectful relationship with the faculty
mentor. Hands-on undergraduate research
experience may therefore be important in
molding the ethical practices of students who
will go on to become professional scientists.
When we emphasize the importance of
getting the “right” answer, we are teaching
undergraduates that their hypothesis must be
supported. In truth, the function of an
experiment should be to allow for a fair test of
the hypothesis. We recognize that there exists
temptation for graduate students and professional
scientists to manipulate data in order to finish
research before a deadline, to obtain the next
grant, or to have an outstanding publication
record. We must take serious responsibility that
we do not teach data manipulation techniques at
the undergraduate level that will continue to be
used in later professional careers.
We must change the common perception
among undergraduate students that their grade
often depends upon producing the “right” answer
(Figure 6). This change will involve not only the
laboratory experimental design, but also the
training of graduate teaching assistants and
elimination of grading based on achieving a
preconceived result. Although students must still
be evaluated on whether they are using
laboratory instruments accurately, we must
consider whether a given laboratory can be
designed for training in the hypotheticaldeductive process in addition to the specific
laboratory technique (4, 5).
Unfortunately, the number of students
enrolled in science laboratory courses at large
universities in many ways promotes cook-book
laboratory exercises. The limited time allowed
for experiments, inability to repeat an
experiment, and disinterest of many teaching
assistants in spending adequate time to grade
reports all contribute to the perception on the part
of students that making up data is more profitable
than writing up what really happened.
Faculty must rethink the reasons for
requiring laboratory reports. If the reasons
include determining whether the student was
present, completed the tasks, understood the
experiment, and learned the techniques, then the
results presented here suggest that these goals are
not being accomplished by the current
mechanism of laboratory reports graded based
upon achieving the “right” answer. Other
mechanisms for discovering whether students
have learned the important aspects of the exercise
may include laboratory-based questions on
exams, and building later experiments upon
earlier laboratory exercises. Instructors must be
willing to address this problem bluntly with the
students and teaching assistants.
At ASU some laboratories have been
redesigned to emphasize the inquiry approach to
laboratories in Biology and Chemistry. Students
generate alternative hypotheses and design
experiments themselves, and concepts are
introduced after, not before, results are obtained.
Teaching assistants are trained to lead students
into open-ended and thought-provoking
questions (4, 5). In spite of these efforts, our data
suggest that data manipulation occurs in these
laboratories as well. As students commonly
state, “everybody does it.” The students
themselves overwhelmingly prefer laboratory
exercises which emphasize processes or a
Acknowledgements
We are grateful to Dr. Anton Lawson for helpful
suggestions.
Bibliography
1. Cizek, GJ. Cheating on tests: How to do it, detect it, and
prevent it. L Erlbaum Assoc, Mahwah, NJ. 1999.
2. Meade J. Cheating: Is academic dishonesty par for the
course? Prism 1992;1: 30-32.
3. Germann PJ, Haskins S, Auls S. Analysis of nine high
31
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
school biology laboratory manuals: promoting scientific
inquiry. J Res Science Teaching 1996; 33: 475-499.
4. Lawson AE, Lewis CM Jr., Birk JP. Why do students
“cook” data? J College Science Teaching 2000; 29(3):
191-198.
5. Lawson AE, Rissing SW, Faeth SH. An inquiry approach
to nonmajors biology. J. College Science Teaching
1990;19(6): 340-346.
6. Swazey JP, Anderson MS, Lewis KS. Ethical problems in
academic research. Amer Scientist 1993; 81: 342-553.
32
–––––––––––––––––––––––––––––––––––––––––––––– Davidson, et al., Data Manipulation in the Undergraduate Laboratory
Data Manipulation Survey
Instructions to students:
Space shuttles blow up, bridges fall, and planes crash and not all are due to natural disasters. An
undergraduate student at ASU has been conducting a research project for the last year and a half.
During his undergraduate career, he found that in some laboratory settings, there appears to be a great
deal of pressure to get the “right” result rather than an emphasis on the scientific and experimental
process. In one of his labs he found that 80% of the students manipulated data in some way during
the semester. He became concerned: where do students learn scientific ethics? Should we have faith
that human morality will overcome pressures to manipulate data in the hopes of a better grade in our
college career, or a publication in our professional career?
The purpose of this survey is to collect data on the extent to which undergraduates feel pressured to
manipulate, change, or make up data acquired in the laboratory. For example, if you only have a 30%
yield of a particular reaction, have you ever felt pressured to say you had more to get a better grade?
Moreover, how did you respond to that pressure? Alternatively, has the lab concentrated on
experimental process rather than actual results?
Data Manipulation: To change or omit acquired data or to make up data without confession to
those evaluating your performance.
1. What is your TA’s name?
2. What is you major and what year are you (freshman, sophomore, etc.)?
3. Are you:
A. Female
B. Male
4. How many science labs have you taken?
A. 1
B. 2-5 C. 6 or more
5. Have you ever manipulated data or made up data in this course?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
6. If you have ever manipulated data or made up data, was it motivated by the thought of a better
grade?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
7. Have you ever observed anyone else manipulate or make up data in this course?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
8. How often have you felt as though your grade in this course depended heavily on your
experimental results?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
9. Do you believe this course places too much emphasis on experimental results rather than on the
processes used to get the results?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
10. Have you ever manipulated or made up data in any other science course?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
33
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
11. If you have manipulated or made up data, was it motivated by the thought of a better grade?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
12. If you have manipulated or made up data, was (were) the course(s):
A. Lower Division (100-200 level) B. Upper Division (300 or 400 level) C. Both A & B
13. If you have manipulated or made up data, what department was (were) the course(s) in? (Please
circle all that apply.)
A. Biology B. Physics C. Chemistry D. Zoology E. Botany F. Microbiology
14. Have you ever observed anyone manipulate or make up data in any science course?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
15. How often have you felt that your grade in a science course depended heavily on you
experimental results?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
16. Do you believe that science courses place too much emphasis on experimental results rather than
on the processes used to get those results?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never
17. Would you like to see lab experiments:
A. Place more emphasis on results. B. Place more emphasis on processes.
C. Have a balanced combination of both.
34
Preliminary Observations on Faculty and Graduate Student Perceptions of
Questionable Research Conduct
Ravisha Mathur, Department of Psychological Sciences, Purdue University, USA
Stuart I. Offenbach, Department of Psychological Sciences, Purdue University, USA
Keywords: Ethical dilemmas, Mentoring, Perceived misconduct, Research ethics training
When thinking about how graduate students learn the values and standards of science, most
universities and departments utilize an apprentice model. In this model, students learn values and
ethics by observing their mentor and through working with the mentor–learning via a kind of
“osmosis” process. However, the mentoring relationship between faculty advisor and graduate
student is one of the most difficult and complex relationships in academia. This sometimes
professional, sometimes personal relationship is generally beneficial to both individuals. Advisors
usually help students develop their careers and develop professionally, as well as help students
network and give them guidance with advice, support, and knowledge. Graduate students help their
advisors by assisting with projects, increasing research productivity, increasing professional visibility
through the student’s research, and can provide their mentors with personal satisfaction and a sense of
competence (1, 2). Despite this mutually beneficial relationship, vital for a graduate student’s career
in graduate school and beyond, faculty members receive very little, if any, training about mentoring.
In fact, given this lack of formal preparation, some suggest the mentoring relationship can cause as
much potential harm as it does benefits (1).
As a mechanism to transmit ethical codes and standards, the mentoring-apprentice model is,
according to some investigators, not very effective (e.g., 3, 4). In order to provide faculty and
graduate students with more effective methods of training and educating students about the
responsible conduct of research, it would be useful to determine which aspects of the practice of
research are most vulnerable to be misperceived, skewed, or violated. In this study, our definition of
the responsible conduct of research includes (but is not limited to) honesty, reporting all collected
data, using appropriate statistical analyses, and fairly recruiting research participants. Although there
is some research describing the types and frequency of scientific misconduct by faculty members and
by graduate students, there is little research examining both faculty and graduate student perceptions
of violations of the responsible conduct of research. Nor do we know how concordant or discordant
these “pairs” are. One purpose of this study was to assess these faculty and student perceptions. A
second purpose of this study was to examine the training that students receive from their faculty
advisors and departments. We hope to pinpoint how training can be improved and enhanced by
examining faculty members’ and students’ perceptions of training and regulations (at both the
department and university level).
Corresponding author: Ravisha Mathur, Department of Psychological Sciences, 1364 Psychological Sciences Building,
Purdue University, West Lafayette, IN 47907-1364, 765-494-6928 (voice), 765-496-1264 (fax),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
In order to investigate these issues, we sent a
survey to faculty members and to graduate
students in each of 30 Purdue University
departments from the schools of Agriculture,
Consumer and Family Sciences, Engineering,
Liberal Arts, Science, and Veterinary Medicine.
Faculty members were certified to chair students’
doctoral committees and graduate students were
certified by the graduate school as doctoral
students. 733 faculty and 242 graduate students
received copies of the survey, and we received a
total of 241 surveys from faculty (of which 225
contained usable data) and 47 surveys from
students (all of which were usable data).1
Although the participation rate in this survey was
comparable to previous research on similar issues
with mail-in surveys (e.g., 5), we were
disappointed that we did not receive more
responses from students (which limited the
analyses and results reported below). The
distribution of returns by Gender and by
Discipline are in Tables 1 and 2, respectively.
Faculty
Grad. Student
Female
47
16
Which of the following are ways that graduate
students learn about professional values and
ethical standards? (Circle all that apply).
1. Brown bag/colloquium
2. Special courses devoted to this topic
3. Interaction with faculty in research work
4. Codes of ethics and professional standards
provided by professional organizations
5. Informal discussion of ethical problems when
they occur
6. Department policies for teaching and research
7. Discussion of ethics and values in regular
course work
Figure 1: Item 2 from Part 1 of the Survey
addressed how information about the responsible
conduct of research is exchanged (Item 2 of Part
1 is shown in Figure 1). The questions in Part 1
focused on how and where students learned about
the responsible conduct of research and if
students and faculty knew of others who had
been involved in ethical conflicts. The main
section of the survey, Part 2, consisted of 38
hypothetical dilemmas (each included a proposed
action to resolve the dilemma). The dilemmas
were written to cover the following types of
problems (which were supported the
confirmatory factor analysis described below):
1) Information Sharing in the Lab;
2) Truth/ Completeness in Writing up Research
Results;
3) Misleading the Audience (Plagiarism);
4) Seeking Credit for doing the Research; and
5) Consent Issues.
(Examples of the dilemmas for each factor are
shown in Figure 2.) Participants responded by
rating each dilemma on a five point Likert scale
(Strongly Disagree to Strongly Agree). The third
and final section of the survey examined
participant’s perceptions of university and
departmental policies on the responsible conduct
of research and whether the faculty member or
graduate students would feel comfortable
reporting incidents of suspected misconduct.
Male
162
29
Table 1: Number of responses by gender
The percentage of responses from both male and
female faculty members and graduate students
matched the gender distribution for the entire
faculty (faculty: 22% female and 78% male;
graduate student: 35.5% female and 64.5%
male). Equivalent comparisons of responses
from the different disciplines were more difficult
to make since different numbers of departments
from each discipline were asked to participate.
As Table 2 indicates, more responses were
received from the Schools of Agriculture,
Engineering, and Science. Only a few graduate
students from Consumer and Family Sciences
and from Liberal Arts participated. Most of the
student responses were from Agriculture and
from Engineering.
There were three parts of the survey. Part 1
Faculty
Grad. Stud.
Agriculture
CFS
Engineering
Liberal Arts
52
13
23
4
32
10
27
2
Table 2: Number of Responses by School
36
Pharmacy &
Medical Sci.
20
7
Science
38
7
–––––––––––––––––––––– Mathur & Offenbach, Faculty and Graduate Student Perceptions of Questionable Research Conduct
a. Sharing Information: Grant is in his office one day and sees his officemate’s lab notebook
open. While paging through the notebook, he discovers that Anli has found a way to metabolize ABC
enzyme. Grant has been working for two months to discover a way to metabolize this enzyme for his
dissertation. After thinking about it for a few days, Grant decides to use the same process to keep his
dissertation on track. He does not bother to tell Anli because she is in his lab group and probably would
not mind anyway. Do you agree with his decision?
b. Writing: Mei has been collecting data for a long-term study for the past two years. Although she
still is in the middle of the data collection phase, the trends she sees in her data are very exciting. She
decides to write up her results and present them as a complete study and continue to collect data for the
full term of the study. She plans to publish those data in at least two “follow-up” reports. Do you agree
with her decision?
c. Misconduct: Angelo has written an article in which he included a full paragraph from a paper
written by a student for a class Angelo was teaching. Do you agree with Angelo’s decision to include the
paragraph?
d. Seeking Credit: John has written an article in which he included a full paragraph from a
pre-publication version of an article reviewing the research literature in his area of interest. The author of
the article was planning to submit it to a prominent journal that publishes such reviews. Do you agree
with John’s decision to include the paragraph?
e. Consent Issues: Professor Gleeson is conducting a research project concerned with social
customs in a village in rural South Africa. The village consists of members of a single tribe, and is led by
a tribal chief and council of elders who make all decisions for the village. The tribal chief insists that he
will decide if his villagers can participate in Professor Gleeson’s research project, and that he (the Chief)
will distribute the payment to the villagers. Professor Gleeson may not ask the villagers whether they
want to participate because that would be an insult and challenge to the Chief and Elders of the village.
Do you agree that Professor Gleeson can go ahead with the research project if the Chief and Elders
approve?
Figure 2: Sample Hypothetical Dilemmas from Part 2 of the Survey
believed supportive faculty members provided
such information. Sixty-seven percent of faculty
members believed professional organizations
provided such information compared to only
15% of graduate students (t = 28.377; Only tvalues significant at .05 or less are reported).
This difference probably reflected a lack of
contact with such organizations by graduate
students. Graduate students also relied more on
other students as a source of information (51%),
a source not considered by faculty members
(15%, t = 16.97).
Interactions with faculty in research work
and informal discussions of ethical problems
were considered effective settings to learn
professional values by 90% or more of students
and faculty. Brown bag discussions, colloquia,
and courses, on the other hand, were not seen as
effective settings by most respondents
(percentages all less than 30%).
We also asked whether respondents ever
discussed with peers value issues related to
external sources of research funding or the
application of research findings. Eighty percent
(Two of these items are shown in Figure 3.)
Items from both Part 1 and Part 3 were adapted
from Judith Swayze and coworkers’ survey of
faculty and students (6). Items for Part 2 were
written by the authors and were based on real
events and scenarios gleaned from reading and
teaching about the responsible conduct of
research for the past five years.
Participants were given a response sheet to
use as their answer sheet and were asked to
return the response sheet in a self addressed
envelope we provided them. Once we received
the survey, a third party removed any identifying
information. The responses on each survey form
were entered into a computer file separately by
the two authors. All coding errors then were
reconciled by the authors.
Results
Part One. The first questions focused on
settings in which respondents learned some or all
of their professional values. Seventy-two percent
of faculty members and 60% of graduate students
37
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
2. How active a role does your department actually take in preparing graduate students to recognize
and deal with ethical issues in your field as part of their training?
Very
active
Active Somewhat
active active
Not very
active
Not at all
active
3. Could you report cases of suspected misconduct in your department without expecting
retaliation?
Misconduct by a faculty member: Yes definitely, Yes, Probably, Probably not, Definitely not
Misconduct by a graduate student: Yes definitely, Yes, Probably, Probably not, Definitely not
Figure 3: Items #2 and #3 from Part 3 of the Survey
of faculty members and 47% of the graduate
students (t = 18.263) did so. In addition 38% of
faculty members and 11% of graduate students
actually knew someone who had refused to
participate in a research project because of
personal reservations about funding sources.
These faculty-student difference probably
reflects differences in age and experience in the
field.
What is clear from these analyses is that
faculty members and students do have different
views of the best place or way to learn about
professional standards and to learn to recognize
ethical research issues.
Part 2: Hypothetical Dilemmas. A
confirmatory factor analysis of the hypothetical
dilemmas produced five factors: 1) Information
Sharing in the Lab; 2) Truth/ Completeness in
Writing up Research Results; 3) Misleading the
Audience (Plagiarism); 4) Seeking Credit for
doing the Research; and 5) Consent Issues. The
alphas for these variables were moderate, ranging
from .47 - .61. We recognize that not all of the
dilemmas applied equally to all of the disciplines
sampled in this survey, but we were pleased that
some general factors appeared. The nature of
the five factors can be explained in several ways.
First (and probably foremost) is the construction
of the scenarios by the principle investigators.
Construction of these scenarios was not a random
process, and the factors extracted from the
analysis may simply confirm biases and
predispositions that entered into our construction
of the items. On the other hand, the areas
represented by the five factors have been
identified by many investigators as areas of
concern vis-a-vis research ethics. The fact that
these items hang together at all may be a
confirmation of the concerns many investigators
and ethicists have about the process of research.
Although we could not adequately examine
the faculty-student differences on the responses
to the Hypothetical Dilemmas because of the
disparity in the number of responses from each
group, we were able to draw some tentative
conclusions. Faculty members clearly took
“more extreme” views than did students. That is,
faculty members were more likely to indicate
strong disagreement or agreement with the action
taken in a dilemma than were graduate students.
For example, on the 20 dilemmas that
contributed to the five factors, more faculty
members responded “strongly agree” (or
“strongly disagree”) on every dilemma.
Graduate students had more moderate responses.
Actually, there were no faculty-student
differences in the number of combined “strongly
agree” and “agree” (or “strongly disagree” and
“disagree”). Thus for the second item in Figure
2, of the 98% faculty members who disagreed
with the action, 80% checked “strongly
disagree.” All of the graduate students disagreed
with the action, but only 43% expressed strong
disagreement. Perhaps faculty members’ greater
experience with ethical issues has led them to be
more certain of their views (or the students’ lack
of experience led them to be more tentative).
Finally, while the responses to the
hypothetical dilemmas made intuitive sense, the
construction of the dilemmas is more complex
than we thought. Respondents often commented
that they saw some items as dealing with
multiple ethical issues or that there was not
enough information presented to make a
judgement. This may be one reason alpha levels
were low for the five factors. More thought must
go into the development of items that have a
more specific focus (and are less complex) for a
survey of this type.
Two sets of analyses were not computed.
38
–––––––––––––––––––––– Mathur & Offenbach, Faculty and Graduate Student Perceptions of Questionable Research Conduct
misconduct by a faculty member or by a
graduate student without expecting retaliation.
The results in Table 6 show that 89% of faculty
members believed they could report misconduct
by a graduate student “safely.” They would
expect no retaliation. The graduate students also
seemed less concerned about retaliation if they
reported misconduct by another student.
Seventy-three percent thought it was safe to
report misconduct by another graduate student.
Reporting misconduct by faculty members was
another matter. Fewer faculty members were
comfortable about reporting misconduct by a
colleague (73%). Only 55% of students thought
they could report misconduct by a faculty
member “safely.” In contrast, 28% of the faculty
members who responded said they would not
feel safe reporting misconduct by a faculty
colleague. Almost half of the graduate students,
44%, were concerned about retaliation for
reporting a faculty member’s misconduct. These
results seem consistent with anecdotal data. A
cursory review of comments from the electronic
list-serve Sci-Fraud reveals a concern by many
participants that to make a good faith allegation
that a faculty member has engaged in
misconduct is to place one’s career in jeopardy.
Finally, we asked about knowledge of
university and departmental policies on
misconduct. Half of graduate student
respondents did not know that the University has
a research misconduct policy and 72% do not
know if their department has such a policy. The
faculty were more knowledgeable – 63% knew
there was a university policy. However, only
half of them were familiar with the policy’s
contents.
Analyses to compare factor scores for students
with those of faculty were not conducted because
the factor scores have not yet been corrected for
directionality differences. That is, some factors
include items with which most respondents agree
and items with which most respondents disagree.
The point values for these items needs to be on
the same scale or have the same valence in order
to examine factor scores. The other analyses not
yet conducted would have compared student
responses with those of their mentors. These
analyses depended on both the student and his or
her mentor actually submitting a survey, and
having the student identify his or her mentor.
Unfortunately, we were able to identify only five
faculty-student pairs, precluding any analysis of
whether the two are concordant or discordant.
Questions about department and
university policies
The questions in Part 3 focused on respondents
perceptions of the role that departments should
take and actually do take in preparing students to
recognize and deal with ethical issues (see Tables
3 and 4). Significantly more students than
faculty (70% vs. 45%) reported almost no effort
by their departments to train them to recognize
and deal with ethical issues in science (it also is
interesting that 16% of faculty members thought
their departments were active, but only 6% of the
students shared that perspective). Thus both
faculty and students believe academic
departments should take a more significant role
in training graduate students to recognize and
deal with ethical issues (we only asked about
academic departments, faculty members and
students may actually ascribe greater
responsibility to larger academic units — e.g.,
schools, graduate school, etc.).
There is a mismatch here –
Somefaculty and students wanting
Very
what Not very Not at all
Active
departments to take a role and
active
active
active
active
departments not doing that. And
Faculty
37
45
14
03
01
there is no formal structure at
the university level for training
Grad. Stud.
22
52
22
04
00
in the responsible conduct of
Table 3: Role a department should take (percent agreeing)
research. Thus, the student is
left to his or her own devices.
SomeVery
Not very Not at all
The most frequent choice made by
Active
what
active
active
active
students seems to be to ask
active
another student or to ask the
Faculty
02
14
38
34
11
advisor.
Grad. Stud.
02
04
26
51
17
The next two questions asked
Table 4: Role a department does take (percent agreeing)
whether one could report
39
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Respondents
Faculty
Members
Graduate
Students
}
}
Misconduct
by
Definitely
Yes
Probably
Yes
Probably
Not
Definitely
Not
Faculty
Students
Faculty
Students
32
48
04
11
41
41
51
62
23
09
40
23
05
01
04
04
Table 5: Reporting Misconduct Responses (percentage agreeing)
Conclusions:
The hypothesis that graduate students learn to
identify and deal with ethical situations in
research from their mentors without specific
instruction or discussion could not be tested
using the data collected in the first “pass” of our
study. We received too few mentor-student data
pairs to make any analysis. Our failure to obtain
data relating mentor’s values directly to that of
their specific students was disappointing – only
five student - mentor pairs were identified (we
hope to rectify this situation by increasing the
size of the student data pool). However, we
believe the modeling or osmosis hypothesis
probably will not be supported because of the
different perceptions graduate students and
faculty members have of how scientific values
are transmitted. Faculty members and students
do rely on other faculty members, but only the
students rely on their student peers. At the same
time, both faculty and students believed that
interactions in the work or lab settings would be
useful in learning to recognize and deal with
ethical situations. Unfortunately, this expectation
means that people seem to want to learn from
“personal experience,” but no one wants to have
that kind of personal experience.
One thing is certain, things will not continue
in the same way. Actions by the National
Institutes of Health to require specific education
on the responsible conduct of research generally
specifically will require universities to do a better
job. That better job might be facilitated with a
more complete understanding of how students
are learning now and by determining not only
what they are learning , but also by determining
what they are NOT learning.
Notes
1. These numbers differ from the totals in Tables 1 and 2
as some participants did not answer the gender or
discipline questions.
Bibliography
1. Johnson, W. B., & Nelson, N. (1999). Mentor-protégée
relationships in graduate training: Some ethical
concerns. Ethics Behav, 9, 189-210.
2. Wright, C. A., & Wright, S. D. (1987). The role of
mentors in the career development in young
professionals. Fam Relations, 36, 204-208.
3. Brown, S., & Kalichman, M. W. (1998). Effects of
training in the responsible conduct of research: A survey
of graduate students in experimental sciences. Sci Eng
Ethics, 4, 487-498.
4. Folse, K.A. (1991). Ethics and the profession: Graduate
school training. Teach Sociol, 19, 344- 350.
5. Mearns, J., & Allen, G. J. (1991). Graduate students’
experiences in dealing with impaired peers, compared
with faculty predictions: An exploratory study. Ethics
Behav, 1, 191-202.
6. Swazey, J. P., Anderson, M. S., & Louis, K. S. (1993).
Ethical problems in academic research: A survey of
doctoral candidates and faculty raises important
questions about the ethical environment of graduate
education and research. Am Scientist, 81, 542-553.
40
Constructing a Personal Model of Research: Academic Culture and the
Development of Professional Identity in the Professorate
L. Earle Reybold, Adult and Higher Education, The University of Texas at San Antonio, USA
Keywords: Academic culture, Faculty development, Graduate socialization, Professional identity
, Research
ethics, Research training, RCR
Doctoral students pursuing academic careers are educated in awkward and mostly tacit
apprenticeships. As students, they are expected to learn professional knowledge and the technical
skills associated with their program of study. Yet, they must simultaneously absorb the culture of
academe and learn their future roles as faculty members. Because learning and thinking are situated
in a social milieu, socialization is a process initiated and established in contexts that construct
knowledge through activity (1). In other words, academic culture and educational knowledge “act
together to determine the way practitioners see the world” (p. 33).
Generally, socialization studies have investigated academic culture as context for student learning
and development. Many of these studies focus on the social aspects of academic culture, particularly
relationships between students and their colleagues or professors (2, 3, 4, 5). These socialization
studies concentrate on students’ experiences as students in higher education and are centered on
classroom modality.
Likewise, inquiry into new faculty socialization segregates faculty roles and responsibilities into
particular genres of experiences such as teaching success (6) and tenure and promotion processes (7).
Unfortunately, faculty socialization studies fail to address how graduate school experiences,
particularly as they are situated in an academic culture, affect the development of professional
identity and ultimately professional decision-making and activity.
When the concept of professional identity and competency is addressed in the faculty
socialization literature, the discussion surveys the development of the faculty teaching roles but
ignores the complex faculty identity as teacher, researcher, and service provider. This lack of
attention to an integrated identity that begins to emerge during graduate studies portrays faculty
socialization in perfunctory terms. For example, Boice discusses new faculty success in terms of
teaching style and mastery (6). The author notes the characteristics of “quick starters,” but these are
teaching characteristics of new faculty, with no attention to the development of these characteristics.
Pollard, Pollard, & Rojewski also investigate the college teaching experience of new faculty (8).
They argue that doctoral students are academically prepared for their careers in higher education, but
their study concentrates only on the impact of higher education culture on new faculty.
Purpose of Study and Research Focus
The purpose of this study is to describe the role of academic culture in determining a personal model
Corresponding author: L. Earle Reybold, Adult and Higher Education, College of Education and Human Development, The
University of Texas at San Antonio, 6900 North Loop 1604 West, San Antonio, TX 78249-0654, 210-458-5429 (voice), 210458-5848 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
of ethical research practice in the professorate.
While little is known about the construction of
faculty identity and role expectations during
graduate studies, even less is understood about
the impact of student experiences on professorial
activities and decision-making, particularly
research competence and reasoning. Two
questions demand consideration. First, how are
doctoral students socialized into the practice of
academic research? Further, how do these
students construct a model of research standards
and ethics that will inform their future practice as
faculty members?
Two general assumptions guide this inquiry:
• Socialization into the professorate is a
developmental rite of passage rather than
two discrete phases of socialization marked
by graduation and/or faculty appointment..
• Preparation for the professorate is situated in
an academic culture that shapes one’s
personal understanding of the professorate
and professional identity and perceived roles.
This study initiates a two-phase longitudinal
qualitative investigation. Using case study
methods (9), this study focuses on doctoral
students’ perceptions of research ethics in
education. Interview questions concentrated on
emergent definitions of research ethics, training
in research ethics, and experiences of ethical
dilemmas.
Case study research is uniquely geared
toward description and understanding of
institutional culture and its impact on
perspective. Merriam describes case study
research as an ideal design for exploring
participants’ understanding and perspective (9).
Further, she says case study is appropriate when
inquiry is interested in “process rather than
outcomes, in context rather than a specific
variable, in discovery rather than confirmation”
(p. 19).
Sampling for this phase of the study is
network sampling, which locates participants
through recommendations of initial participants
and key informants based on selected criteria
(10). Participants were located at three
universities in Georgia and Texas, including
institutions identified as Research I, Research II,
and Doctoral II. Participants were doctoral
students in education preparing for a faculty
career in academe.
Data were collected through in-depth
interviews with doctoral students and faculty
members at three universities in two states, and
through archival data such as program materials
and reflection journals supplement the interview
data. Interviews were conducted using a semistructured format to allow comparison of data
across participants (11). In general, interview
questions addressed student and professional
identity, academic culture, training in teaching
and research, and ethical decision-making as a
professional. Journaling allowed students to
explore and document their process of decision
making as relevant issues arose, the entries were
guided by the following statement: Describe your
decisions that are most important to your
preparation for the professorate.
Standards for Quality Research
Emphasizing qualitative inquiry as a creative
process, Patton (10) reminds researchers of the
“technical side to analysis that is analytically
rigorous, mentally replicable, and explicitly
systematic” (p. 462). Merriam (9) adds that
qualitative research findings “are trustworthy to
the extent that there has been some accounting”
(p. 198) for quality. In general, the criteria for
trustworthy qualitative research include rigorous
and systematic data collection and analysis
techniques, credibility of the researcher, and
belief in naturalistic inquiry (10). The quality of
this study is enhanced by several factors. First, I
have experience as a qualitative researcher and
have taught qualitative methods at the graduate
level. Further, triangulation of methods and peer
review of data and analysis will enhance the
trustworthiness of the data. Finally, the multi-site
design encourages usability of the findings
beyond the university settings included in the
study.
Situating Faculty Identity Development in
Academic Culture
This study is framed by the concepts of research
ethics and integrity, faculty socialization and
enculturation, and professional identity
development.
Research Ethics and Integrity.
Research is often messy and complicated. Bestcase scenarios of theoretical contributions and
improvement of practice are weighed against
questionable issues of right and wrong research
behavior. In these cases, research decisions may
evolve as uneasy guesses with no obvious
consequence. Confronted with uncertain choices,
42
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Reybold, Constructing a Personal Model of Research
individual’s values and beliefs to fit the “cultural
ethos of the institution” (p. 25). Tierney and
Bensimon continue this emphasis on
socialization in academe, focusing on the tenure
process as the locus of organizational
socialization (7). Although they offer strategies
for anticipatory and organizational socialization,
the authors do not focus their attention on the
transition process.
Bergquist examines academe within the
framework of organizational culture, concluding
that there are four distinct cultures: collegial,
managerial, developmental, and negotiating (13).
Culture, he says, “provides meaning and context
for a specific group of people,” adding “the
culture holds the people together and instills in
them an individual and collective sense of
purpose and continuity” (p. 2). Further,
Bergquist says culture defines the nature of
reality for members of a given culture, providing
the “lenses through which its members interpret
and assign value to the various events and
products of the world” (p. 2). Although there are
four distinct cultures within academe, one will
usually be dominant. Bergquist notes that the
interaction among the four unequal cultures helps
“to produce the often confusing and paradoxical
conditions in which contemporary faculty find
themselves” (p. 7).
Both Bergquist (13) and Tierney & Rhoads
(12) note the influence of academic culture on
faculty perspectives, decisions, and behavior;
also, they agree that cultural differences create a
backdrop of conflict for members within a given
culture. This study extends their conclusions to
graduate education, adding that students also are
influenced by academic culture. Further, the
transition process from doctoral studies to the
professorate adds another layer of possible
conflict between academic cultures.
how do researchers define and respond to ethical
dilemmas?
Ultimately, ethical decision-making reaches
beyond the local boundaries of specific research
projects. Because research is fundamental to
higher education, it could be argued that research
decisions symbolize the moral character of
higher education. Under the guise of exploration
and discovery, research is a noble enterprise. But
research agendas are realized within the
“publish-or-perish” mentality of higher education
in which ethical dilemmas may become
stumbling blocks to promotion and tenure. This
is the context where doctoral students are
socialized toward the professorate; this is the
culture that trains future faculty members as
future researchers.
Faculty Socialization and Enculturation.
Tierney & Rhoads (12) remind us that
“organizations exist as social constructions” (p.
1) that revolve around shared understandings.
This organizational culture shapes behavior and
expectations, bounding faculty socialization.
Tierney & Rhoads define faculty socialization as
“the process through which individuals acquire
the values, attitudes, norms, knowledge, and
skills needed to exist in a given society” (p. 6).
Their definition of faculty socialization as
transmission of culture complements this study
of professional identity development.
Tierney & Rhoads (12) describe academic
culture as the nexus of five forces: national,
professional, disciplinary, individual, and
institutional. Although these are conceptualized
as distinct subcultures, these forces are
synergistic and do not operate independently of
one another. Professional identity is an aggregate
sense of self that develops across these
subcultures. This process of socialization occurs
in two overlapping stages: anticipatory
socialization and organizational socialization.
The anticipatory stage “pertains to how nonmembers take on the attitudes, actions, and
values of the group to which they aspire” (p.23).
The organizational stage, on the other hand,
involves initial entry and role continuance.
Noting the importance of the transition process,
Tierney & Rhoads comment that when
anticipatory socialization and organizational
socialization are consistent, the socialization
process is affirming. When socialization
experiences are not consistent, the organization
will attempt to modify or transform the
Developing a Professional Identity.
Marcia defines identity development as a selfconstructed organization of drives, abilities,
beliefs and individual history (14). Bruss &
Kopala (15), building on Marcia’s definition,
define “professional identity “the formation of an
attitude of personal responsibility regarding one’s
role in the profession, a commitment to behave
ethically and morally, and the development of
feelings of pride for the profession” (p. 686).
This definition directly connects professional
identity to professional behavior.
While the identity development literature is
43
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
concerned predominantly with the psychological
aspects of self, identity may be viewed as both
personal and social. Social identities result in
identity relationships within a given culture, and
these identity relationships determine identity
status and role expectations (16). For the
purpose of this study, status and role expectations
will be examined as cultural aspects of
professional identity development, particularly as
they relate to anticipatory socialization during
the graduate school experience (7).
Graduate training is expected to nurture the
development of professional identity. In their
discussion of psychology students, Bruss and
Kopala (15) described graduate school training as
professional infancy and “the training institution .
. . as an environment wherein it is the task of the
faculty and training staff to nurture and promote
growth” (p. 686). However, academic culture is
not always nurturing; structural problems in
graduate education are potentially harmful to
students’ self-esteem (17). Attitudes—good and
bad—about professional responsibility, ethical
behavior, and professional pride are constructed
within the cultural context of graduate training.
These attitudes produce social identities and role
expectations that persist through a graduate
student’s transition into the professorate. In
short, academic culture exerts directive force
over professional decision-making and activities.
Chickering & Reisser, in their study of
college student development, define identity as a
sense of self (18). The process of identity
development results in “a solid sense of self
[that] emerges, and it becomes more apparent
that there is an I who coordinates the facets of
personality, who ‘owns’ the house of self and is
comfortable in all of its rooms” (p. 49).
Findings
To describe the role of academic culture in
determining ethical research practice, data were
analyzed within four concentrations: the
perceived role of research in higher education,
the perceived standards for ethical research, the
actual ethical dilemmas experienced by graduate
student researchers, and the factors that hinder or
support ethical research.
What is the perceived role of research in
higher education? Participants in this study
experience research and subsequent publication
as an institutional priority and a personal badge
of prestige. While one participant views the
professorate as a delicate balance of professorial
roles, most participants emphasized the
preeminence of becoming a researcher, and only
one participant noted a teaching role being more
important than a research role. For example,
Betsy says, “research is painful and boring, but
the doctorate is about what the university
considers important—getting published!”
Echoing this sentiment, Claire says the “doctoral
degree is mainly trying to get us into the research
part of being a professor and much less teaching;
it is indoctrination into the research aspect of
being a professor.”
While some participants came in with
considerable research experience, most are
concerned that they don’t “know what to do with
the research” after the dissertation process. Postdissertation concerns include translation of
theory into educational practice, establishing a
research agenda, and getting published.
What are the perceived standards for ethical
research and who defines ethics in academic
settings? Coursework in research ethics is almost
nonexistent. As students, participants expect
professors to guide them through the process of
learning and implementing ethical research, but
they are relying instead on their own sense of
right and wrong. Julia says she relies on her
“internal gyroscope” to guide her decisions; and
Claire relies on her “personal ethics and personal
morals.” Grace adds that “ethics is about power
differences.” Her professors talked about
collaboration and high quality, but their practice
expressed a disregard for the Institutional Review
Board (IRB), quality research, and research
integrity.
More than a lack of definition of ethical
research, participants are concerned and confused
about “grey” areas of research ethics and believe
they must define ethical research according to
their own experiences and standards.
Interestingly, the two participants with training in
medical ethics find research ethics easier to
define. The other participants have scattered
definitions of research ethics, with most
positioning ethical research as a data collection
and/or analysis issue. However, a couple of
participants have a complex, comprehensive
definition of research ethics, including researcher
attitude and choices throughout the research
process. One participant noted that article
readers have an ethical responsibility to read the
results thoroughly. Another participant, Grace, is
quite concerned with the power issues that
44
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Reybold, Constructing a Personal Model of Research
presentations because funding for student travel
to conferences depends on authorship. Grace did
try to confront the professor, but to no avail. The
professor was on the editorial board of the
journal that published the article, and she
believed the issue would not be taken seriously.
Participants report that even when the research
situation is uncomfortable, they “don’t want to
sacrifice the relationship” by removing
themselves from the project.
Another type of dilemma involves committee
make-up. One participant had approval for a
mixed design dissertation, but her committee
politicized her design and held up her research.
She decided “to write it one way for the
dissertation” and then publish it using her mixed
design approach. Other participants experienced
negative “shaping” of their research based on
professors’ interests. As one participant reports,
“professors stay in their comfort zones” and
won’t head committees outside their personal
interests. This is particularly problematic in
small departments with few faculty members.
What factors hinder or support ethical
research? Several factors hinder ethical
research: institutional/structural, relational/
positional, and technical. First, the culture of
academe encourages ambivalence toward the
issue of ethical research. Institutions reward
research productivity, even at the expense of
other professorial roles, perpetuating the adage,
publish or perish. While some professors
“nudge” their students to commit ethical
violations, others ignore the need for training and
guidance in ethical research practice. Dan,
looking toward a future career in academe,
acknowledges that “tenure is political, so go way
beyond their expectations!”
A second factor hindering ethical research is
the role of hierarchy in academic relationships.
Graduate students are afraid to report ethical
violations; they fear losing their assistantships
and professorial support. As a student, one
participant notes that “it’s important to know
where your allegiances lie; the only way you’ll
get lobbied for is if you are clearly in someone’s
camp.” Only one student, Kelly, says her
professors treat her as a peer. Her major
professor, she says, “got me involved with his
projects, but told me to ‘find your own thing—
academia isn’t just doing other people’s work.’”
Several participants alluded to expecting a
similar role as junior faculty; coercion will
continue to force them to make ethical decisions
impact ethical decision-making: “power issues
come into play, whether we like to admit it or
not…these are times we just have to make a
mental note, ‘this is not right’…. But I’m at a
point where I have no power to address this.”
One participant has a collaborative
relationship with her major professor. Kelly says
her discussions with her major professor about
research methods and ethics have been
invaluable, even to the point where she feels
comfortable mentoring other students with
research problems. Although Betsy claims to
have a collaborative and mentoring relationship
with her major professor, she often finds herself
involved in ethical dilemmas with others in the
same department. For the participants in this
study, the most beneficial contribution to ethics
and methods training is involvement in actual
research projects, particularly pilot studies of
own research and collaborative efforts as
research partners with professors, but only when
that contribution is valued and rewarded as equal.
What types of actual ethical dilemmas do
graduate student researchers experience? While
most participants define ethical dilemmas in
terms of research methods, their experiences of
ethical dilemmas focus more on relationships and
issues of power and coercion. One participant
reports her professor “uses” students to review
his own material prior to publication. Student
assignments in non-related courses revolve
around this professor’s research agenda, and
students are expected align their work to match
that agenda. Several participants report being
forced to manipulate data to yield desired
outcomes; if a student refuses, he or she is no
longer funded as a research assistant. Kelly, a
research assistant on a grant-funded study, voiced
disapproval of research decisions being made by
professors on the grant:
I’ve been vocal, but I wasn’t a threat or
anything. I was unhappy with the way the
professors were doing things . . . . I was just
going along, and it hit me. Did I feel free to
leave? No! To a certain extent, this is part of
being a graduate student. I mostly feel free to
voice my concerns, but in this case, it was an
ultimatum—or I was off the grant! I never want
to do this in my own research.
Another participant, Grace, reports working on
presentations and articles with more than one
professor and negotiating authorship—but the
articles were published without her name or with
a different authorship order than negotiated. This
is particularly troublesome at conference
45
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
that might not be supported by academic
expectations.
A third factor that hinders ethical research is
the lack of training and exposure to guidelines.
Only those participants with medical
backgrounds had any courses in ethics, and those
courses dealt with medical ethics rather than
research ethics. Only one participant reports
research ethics discussed in her doctoral research
classes. None of the participants in this study
knew of any guidelines for education research
practice other than IRB guidelines.
Only one participant, Kelly, reports
preparation for ethical research. Her major
professor and a network of research professors
provide guidance through formal and informal
mentoring and involvement in various research
projects. This particular participant has
published with several professors, and her
contributions are valued as equal to those of the
professors. In fact, this professor reminds the
participant that she is his “primary
responsibility” and that she is to develop her own
line of research separate from his. Another
participant feels secure in her relationship with
her major professor, but says her other
experiences with faculty members in the same
department make her bitter and wary. She notes
there are two levels of culture in the department,
and a “lot of politics gets played below the
surface” even though “we represent ourselves as
a united front.”
Summary and Conclusion
Almost all participants in this study raised
themes of power and censorship. The impact of
coercion and fear on the research process must be
explored. Graduate students believe their
research is censored on several levels: personal,
institutional, and systemic. First, graduate
students expressed fear of retaliation if they
resisted their faculty advisor’s management of
their research. Further, these students believe
they are bound by the dissertation committee
structure and the institutional support of highly
productive faculty members. Finally, censorship
is systemic, according to these students’
experiences, because certain topics are
“favorites” of funding agencies. Likewise, these
students believe journal editors and blind reviews
control the emergence of new knowledge.
The goal of higher education is the
preparation and personal development of
competent, well-trained professionals. While
much of higher education focuses on the
development of technical competence and critical
thinking skills, the transformation from student
to faculty member is too often left to chance.
Future inquiry will explore the development
of professional identity throughout preparation
for the professorate, and how this emerging
identity impacts professional decision-making as
a scholar.
Bibliography
1. Brown JS, Collins A, Duguid P. Situated cognition and
the culture of learning. Educational Researcher 1989;
18(1):32-42.
2. Heinrich, KT. Doctoral advisement relationships
between women. Journal of Higher Education 1995;
66(4):447-469.
3. Martinez Aleman AM. Understanding and investigating
female friendship’s educative value. Journal of Higher
Education 1997; 68(2):119-158.
4. Schroeder DS, Mynatt CR. Female graduate students’
perceptions of their interactions with male and female
major professors. Journal of Higher Education 1993;
64(5):555-573.
5. Stalker J. Athene in academe: Women mentoring women
in the academy. International Journal of Lifelong
Education 1994; 13(5):361-372.
6. Boice R. Quick starters: New faculty who succeed. New
Directions for Teaching and Learning 1991; 48:111-121.
7. Tierney WG, Bensimon EM. Promotion and tenure:
Community and socialization in academe. New York:
State University of New York Press; 1996.
8. Pollard C, Pollard RR, Rojewski JW. The college
teaching experience: A qualitative study. College
Student Journal 1991; 25(2):149-154.
9. Merriam SB. Qualitative Research and Case Study
Applications in Education. San Francisco: Jossey Bass;
1998.
10. Patton MQ. Qualitative evaluation and research
methods. 2nd ed. Newbury Park, CA: Sage; 1980/1990.
11. Bogdan RC, Biklen SK. Qualitative research for
education: An introduction to theory and methods. 2nd
ed. Boston: Allyn and Bacon; 1992.
12. Tierney WG, Rhoads RA. Faculty socialization as a
cultural process: A mirror of institutional commitment.
ASHE-ERIC Higher Education Report No. 93-6.
Washington, D.C.: The George Washington University,
School of Education and Human Development; 1993.
13. Berquist WH. The four cultures of the academy. San
Francisco: Jossey-Bass; 1992.
14. Marcia J. Identity in adolescence. In: Adelson J, editor.
Handbook of adolescent psychology. New York: Wiley;
1980. p. 159-187.
15. Bruss KV, Kopala M. Graduate school training in
psychology: Its impact upon the development of
professional identity. Psychotherapy 1993; 30 (4); 685691.
46
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Reybold, Constructing a Personal Model of Research
16. Goodenough WH. (1969). Rethinking ‘status’ and
‘role’: Toward a general model of the cultural
organization of social relationships. In: Tyler SA,
editor. Cognitive anthropology. New York: Holt,
Rinehart and Winston; 1969. p. 311-30.
17. Egan JM. (1989). Graduate school and the self: A
theoretical view of some negative effects of
professional socialization. Teaching Sociology 1989;
17, 200-208.
18. Chickering AW, Reisser L. Education and identity. 2nd
ed. San Francisco: Jossey-Bass; 1993.
47
Undergraduate Academic Cheating as a Risk Factor for Future Professional
Misconduct
Julio F. Turrens, Department of Biomedical Sciences, University of South Alabama, Mobile,
AL, USA
Irene M. Staik, Department of Behavioral and Social Sciences, University of Montevallo,
Montevallo, AL, USA
D. Kristen Gilbert, Department of Behavioral and Social Sciences, University of Montevallo,
Montevallo, AL, USA
W. Curtis Small, Department of Biomedical Sciences, University of South Alabama, Mobile,
AL, USA
John W. Burling, Department of Behavioral and Social Sciences, University of Montevallo,
Montevallo, AL, USA
Keywords: Academic misconduct, Cheating, Undergraduate education
“T’is education forms the common mind:
Just as the twig is bent the tree’s inclined.”
Alexander Pope (1688-1744)
Scientific misconduct may be more acceptable in the minds of those professionals who grew
accustomed to lower academic standards during their formative undergraduate years. The hypothesis
proposed in this paper is that the recent increase in cheating at the undergraduate level is likely to
result in an increase in the number of future professionals involved in scientific misconduct.
Twenty years ago, academic misconduct at the undergraduate level was considered by the great
majority of both students and faculty as unacceptable and dishonest behavior. Currently, not only are
most undergraduate students aware that misconduct is very common but most of them by their Junior
year have participated or witnessed more than one event. Even those students who do not engage in
academic misconduct have become more skeptical of the need to be personally responsible for their
own academic work and accept this lowering of standards as a fact of life.
Because of these changes in the environment of higher education, the incidence and prevalence of
cheating by college students has been an area of intense concern for educators and researchers since
the 1970s. A vast number of articles in the literature indicate that cheating or academic dishonesty is
at epidemic proportions within academia (1-7). A representative sampling of articles documenting this
Corresponding author: Julio F. Turrens , Department of Biomedical Sciences, University of South Alabama, Mobile, AL
36688, 251-380-2714 (voice), 251-380-2711 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Eric # or Journal
Year
Sample size
Institutions
Reported cheating
ED427355
1998
203
four years
two years
78 %
57%
EJ351071
1986
380
ED334921
1990
232
ED347931
1992
87
EJ449186
1992
6000
31 top-ranked
business: 87%
engineering: 74%
science: 67%
humanities: 63%
EJ489082
1994
480
2 colleges
89%
EJ518822
1995
300
Res. High Ed.
37:487-502, 1996
1984
1994
380
474
> 50%
Rutgers
88%
81%
83%
mid size
liberal arts
54.1%
61.2%
Table 1. Studies showing increased cheating by undergraduate students.
recent increase in cheating by students is shown
in Table 1. Estimates in the literature reveal that
75% to 98% of college students cheat at least
once during their college career (8, 9). Students,
also reported that they are deterred from cheating
only by a fear of getting caught and public
embarrassment (2, 10). High achievers and
students who have too little time to study for
tests are particularly vulnerable to cheating (11,
12).
Students also report that their perception of
faculty reactions to cheating is one of apathy.
Faculty members often do not report a case of
student cheating to the institutional justice
system, either for fear of legal repercussions or to
prevent hurting the reputation of the student.
Instead, many faculty members prefer to handle
each case on an individual basis, sending a signal
to students that the repercussions for cheating are
minimal (6, 13). This signal is tantamount to
acceptance of academic dishonesty as a fact in
higher education by both faculty and students.
An added problem is that faculty and
students often do not agree on what actions
constitute cheating in and out of the classroom
(14-17). The literature recommends that college
teachers should be very specific in their
definition of academic dishonesty, giving
concrete examples, and then following through
on consistent discipline when cheating occurs
(18, 19). In an effort to determine the level of
potential disagreement and/or confusion as to
what constitutes cheating behaviors in and out of
the classroom, the students and faculty of the
University of Montevallo were presented with a
variety of examples of academic misconduct, and
then asked to rank their perceived severity on a
scale from 1 to 5 ( 1 = Not Severe to 5 = Very
Severe) (14). The results of this study are shown
in Table 2. In several cases (see questions 2224) there was almost a full point difference
between the student and faculty perception
indicating a lack of communication between
faculty and students. Some of the most
problematic areas of disagreement (see questions
3, 5, 12, 14, and 15) indicate a educational moral
laxity on the part of the students.
One may interpret these results in two
different ways. One possibility is that the results
reflect stricter standards developed by faculty
members as they moved in their careers. In other
words, their perception reflects a more mature
evaluation of the scenario being considered. If
this interpretation is correct, one also would
expect students to improve their moral standards
as they mature. In other words, the students’
perception of what constitutes misconduct,
should not have any influence in their future
professional conduct. This hypothesis, however,
does not take into consideration that the faculty
members polled in this study already had a
different perception of what constituted cheating
50
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Turrens, et al., Undergraduate academic cheating
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Question
Looking directly at another persons paper to copy
an answer during a test
Using "crib notes" or a "cheat sheet" during a test
or class assignment
Getting a copy of the test prior to taking it
Having/paying someone to do homework or athome projects for you
Copying someone’s homework
Using answer book or keys to get homework
answers
Leaving the test to go to the restroom/or another
place to get answers
Answering "here" or signing someone’s name
when he/she is absent
Copying someone’s paper to work and putting your
name on it
Trying to influence a teacher to give you a better
grade
Using sorority/fraternity test files
Finding someone’s idea and using it as your own
Asking for answers with gestures or sign language
during an in-class assignment
Plagiarism of resource materials or documented
work
Using another’s research for your own benefit
Watching someone cheat without reporting it
Not carrying your weight in a group project for
which everyone gets the same grade
Using sources on homework which the professor
told you not to use
Getting a teacher’s copy of a test to sell
Conducting group sessions to swap or check the
accuracy of answers
Giving answers with gestures or sign language
during an in-class assignment
Lying to a teacher about why you are not prepared
in class
Taking money for doing someone’s work
Glancing at another paper and seeing something to
jar your memory
Working with someone else on a take-home exam
Faculty
4.88 ± 0.67
Student
4.38 ±1.29
P
0.0017
4.83 ± 0.70
4.32 ± 1.31
0.0016
4.80 ± 0.76
4.76 ± 0.81
3.94 ± 1.23
4.06 ± 1.20
0.0001
0.0001
4.63 ± 0.87
3.95 ± 1.24
3.77 ± 1.19
3.10 ± 1.34
0.0001
0.0001
4.77 ± 0.84
4.24 ± 1.33
0.0022
4.71 ± 0.79
3.55 ± 1.32
0.0001
4.82 ± 0.73
4.17 ± 1.30
0.0001
3.46 ± 1.31
2.83 ± 1.30
0.0011
3.56 ± 1.46
4.36 ± 1.00
4.54 ± 1.01
3.05 ± 1.47
3.77 ± 1.32
3.93 ± 1.39
0.0178
0.0009
0.0010
4.76 ± 0.75
4.06 ± 1.39
0.0010
4.31 ± 1.13
3.51 ± 1.23
3.93 ± 1.17
3.67 ± 1.40
2.88 ± 1.26
3.62 ± 1.36
0.0008
0.0007
0.0991
4.15 ± 1.16
3.59 ± 1.26
0.0526
4.62 ± 1.03
2.71 ± 1.35
4.22 ± 1.31
2.15 ± 1.34
0.0072
0.0166
4.50 ± 1.18
3.83 ± 1.30
0.0017
4.22 ± 1.98
3.27 ± 1.31
0.0000
4.58 ± 1.01
4.40 ± 1.15
3.62 ± 1.33
3.49 ± 1.24
0.0001
0.0000
3.92 ± 1.37
3.06 ± 1.37
0.0004
Table 2. Perception by Faculty and Students of Cheating Behavior in College. 140 students and 108 faculty members were
asked to assign a value to the perceived severity of the behavior on a scale of 1 to 5, with 5 being most severe. The results
are presented as average ± SD. The study was carried out at the University of Montevallo during the Fall of 1997.
when they were in college. They grew up with a
different set of standards, in an environment in
which cheating was not as prevalent. Thus,
accepting this hypothesis would imply that
regardless of the predominant moral values
among college students at any given point in
history, they will always develop the correct
moral values as they become professionals.
An alternative hypothesis is that, although
the moral standards of most individuals increase
through life, some of these individuals do not see
any need to change their values. For them the
51
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
concept of “misconduct” disappears. The
concern of those interested in maintaining high
post-secondary educational standards is that the
habits established by some college students will
continue to be their habits in graduate school,
employment and research in the future.
Therefore, an increase in the proportion of an
undergraduate students involved in academic
misconduct is likely lead into an increased
incidence of professional misconduct in the
future.
The current situation is likely to deteriorate
even more. The development of the Internet at
the end of the 20th century has also increased the
number of cheating episodes by providing tools
that were not available even 10 years ago.
Students may now download an enormous
amount of information in seconds, which may be
incorporated into a paper with a couple of
keystrokes. Moreover, several virtual companies
have proliferated offering term papers in all
disciplines on a per page cost (see for example,
www.schoolsucks.com, www.ezwrite.com,
www.academictermpapers.com, etc.). In the last
two years there has been a increase in number of
cases of plagiarism by students who simply
download text from the internet, not just at the
University of South Alabama and the University
of Montevallo but also at many other institutions.
When confronted by the faculty, these students
are dismayed at getting caught, but many will
repeat similar behaviors in the future. The only
tools available to faculty to identify these cases is
to search the web for a specific (unique)
paragraph in the paper or to contract the services
of commercial search engines (for example,
www.plagiarism.org) that can look for the papers
sold to students by Internet companies. The first
procedure is time-consuming and limited. Hiring
the services of a company to track these papers
down still requires someone to enter the text in
the Internet and also the becomes too expensive.
Since the formative years of college are
important in setting many of our standards, as the
students’ academic standards decrease future
professionals may find it easier to engage in
scientific misconduct as they will perceive it to
be less immoral and more expedient. For
example, a study done with 2,459 sophomore
medical students showed that 4.7% admitted to
cheating while 66.5% admitted to having heard
of cheating among their peers (20). About 70%
of the students that admitted having cheated in
medical school also admitted to cheating in high
school and college. Thus we see a moral laxity
beginning at the high school level (or before) and
progressing, probably with more cheating
occurring rather than less, as the level of the
academic workload increases.
One of the established patterns of human
development is the relative stability of
personality traits and behavioral habits over the
life span. Thus, traits of dishonesty in the face of
hard or demanding intellectual work in college,
will, in all likelihood, remain stable
characteristics as these college students grow
older. One cognitive/moral development
theorist, Kohlberg, proposed a universal set of
discrete stages of moral development based on
Piaget’s theory of cognitive development (21,
22). As a child develops more complex and
different modes of thinking and reasoning, the
child should also be able to make more complex
and adaptive moral judgments. Kohlberg
proposed a six-level moral developmental
sequence. At Levels 1 and 2, there is a basic
desire to escape punishment and to win some
level of approval from significant others. At
Levels 3, 4, 5, and 6, the individual may progress
from living up to others’ expectations, to
following rules to maintain the social order and
avoid chaos, to adhering to a social contract only
when it appears to be valid to the individual, and,
finally, to upholding moral judgments and
principles despite potential harm or threat to
oneself because of their intrinsic worthiness.
Kohlberg proposes that rarely do most
individuals progress in moral development past
Level 3 or perhaps 4 (21, 22). We do the “right”
thing in any given situation to garner favor and
approval from others who expect a substantial
effort from us. And, if we perceive the rules that
are in place for us to follow to be unfair or
nonsensical, we may make a judgment to avoid
complying with those rules on what we call
moral grounds.
With Kohlberg’s postulations in mind, it is
then easy to hypothesize that an individual who
learned to cheat in academic situations without
active reprisal from faculty or a school
administration, would tend to repeat those
cheating behaviors in future learning/academic/
research situations as a way to gain approval for
completion of the assignment or project. In
addition, if the adult who participated in
academic dishonesty all the way through
graduate school may view the demands of a
thesis or dissertation committee as non-valid, that
52
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Turrens, et al., Undergraduate academic cheating
individual may engage in academic dishonesty
with an almost-clear conscience. The
requirements of “publish or perish,” then, in the
post-academic world may become “non-valid” in
the professional’s mind, and the individual may
continue to participate in dishonesty in research.
In summary, the correlation between cheating
in high school, college and in medical school
supports our hypothesis that future professional
misconduct will also show a positive correlation
with previous history. Thus, we propose that
part of the efforts to promote integrity among
future professionals should be devoted to curbing
cheating at the undergraduate level since an
increase in one is likely to increase the other.
Journal of College Student Development 1989; 30: 401406.
14. Fleming KD, Keegan DM, Staik IM,, Burling JW.
Faculty perception of cheating in college: Toward a
campus consensus. 1998; Paper presented at the 1998
NCUR (National Conferences on Undergraduate
Research) meeting, Salisbury State University, MD.
15. Livosky M, Tauber RT. Views of cheating among
college students and faculty. Psychology In The
Schools 1994; 31: 72-82.
16. Roig M, Ballew C. Attitudes toward cheating by college
students and professors. 1992; ERIC NO: ED349895.
17. Falleur D. An investigation of academic dishonesty in
Allied Health. Journal of Allied Health 1990; 19:313324.
18. Davis SF, Ludvigson HW. Additional data on academic
dishonesty and a proposal for remediation. Teaching of
Psychology 1995; 22: 119-121.
19. Beemsterboer PL. Academic integrity: What kind of
students are we getting and how do we handle them
once we get them? Journal of Dental Education 1997;
61: 686-688.
20. Baldwin DC Jr., Daugherty SR; Rowley BD, Schwarz
MR. Cheating in medical school: a survey of secondyear students at 31 schools. Acad. Med. 1996; 71: 267273.
21. Kohlberg L Essays on moral development: Vol. II. The
psychology of moral development. San Francisco:
Harper & Row; 1984.
22. Kohlberg L. The development of moral judgment and
moral action. In L. Kohlberg (ed.), Child psychology
and childhood education: A cognitive-developmental
view. New York: Longman; 1995, pp. 259-328.
Bibliography
1. Aaron RM, Georgia RT. Administrator perceptions of
student academic dishonesty in collegiate institutions.
NASPA Journal 1994; 31: 83-91.
2. Diekhoff GM, LaBeff EE, Clark RE, Williams LE,
Francis B, HainesV J. College cheating: Ten years later.
Research in Higher Education 1996; 37: 487-502.
3. Haines VJ., Diekhoff GM, LaBeff EE, Clark RE.
College cheating: Immaturity, lack of commitment, and
the neutralizing attitude. Research in Higher Education
1986; 25: 342-354.
4. Holm TT. Ignorance of the crime is no excuse: Do
public speaking students even know what cheating is?
Paper presented at the 84th Annual Meeting of the
National Communication Association (New York, NY,
November 21-24, 1998). ERIC NO: ED427355.
5. Lord T, Chiodo D. A look at student cheating in college
science classes. Journal of Science Education and
Technology 1995; 4: 317-324.
6. Maramark S, Maline MB. Academic dishonesty among
college students. Issues In Education. 1993; U.S.
Department of Education, Office of Research, p. 5,
ERIC NO: ED360903.
7. Whitley BE Jr. Factors associated with cheating among
college students: A review. Research in Higher
Education 1998; 39: 235-274.
8. Graham MA, Monday J, O’Brien K, Steffen S. Cheating
at small colleges: An examination of student and faculty
attitudes and behaviors. Journal of College Student
Development 1994; 35: 255-260.
9. Genereux RL, McLeod BA Circumstances surrounding
cheating: A questionnaire study of college students.
Research in Higher Education 1995; 36: 687-704.
10. Moffatt M. Undergraduate cheating. 1990; ERIC NO:
ED334921.
11. Greene AS, Saxe L. Everybody (else) does it: Academic
cheating. 1992; Paper presented at the Annual Meeting
of the Eastern Psychological Association (Boston, MA,
April 3-5), ERIC NO: ED347931.
12. Meade J. Cheating: Is academic dishonesty par for the
course? Prism 1992; 1: 30-32.
13. Jendrek MP. Faculty reactions to academic dishonesty.
53
2. Institutions and Professions
Comprehensive Guidelines for the Responsible Conduct of Researchers
Gregory Brock, Department of Family Studies, University of Kentucky, USA
Sandra Sutter, Department of Family Studies, University of Kentucky, USA
Ada Sue Selwitz, Office on Research Integrity, University of Kentucky, USA
Keywords: Conduct guidelines, Research ethics, Researcher ethics
In 1989, the Department of Health and Human Services (DHHS) through the Public Health Service
defined research misconduct and established regulations for reporting scientific misconduct among
awardee and applicant institutions (1). The focus of this regulation was on fabrication, falsification,
and plagiarism. More recently DHHS has shifted emphasis toward preventing misconduct and to the
promotion of Responsible Conduct in Research (RCR).
Success in implementing regulatory initiatives on research integrity has been stymied by several
factors. There is disagreement about the extent of research misconduct. Steneck (2) reported that
fewer than 200 cases of misconduct have been documented by federal government research
investigation offices over the past 20 years. Indirect evidence also cited by Steneck, however,
suggests that misconduct may occur far more frequently.
Additionally, there is a lack of clarity about what amounts to research misconduct. In 1989, the
term focused on, “…fabrication, falsification, plagiarism, or other practices that seriously deviate
from those that are commonly accepted within the scientific community for proposing, conducting, or
reporting research.”(1). Defining deviant practice as well as what is common practice is particularly
challenging in view of the rapid development now occurring within many scientific disciplines—what
was deviant can become common practice. Plus, collaboration among academic disciplines, between
universities and industry, between universities and government, and between international research
teams creates new syntheses that further complicate our understanding of what constitutes common
practice. In an effort to address these issues, regulators have turned to requiring training of
researchers as one means of communicating that the incidence of misconduct is troubling. Training
objectives also clarify what amounts to misconduct.
On December 1, 2000, the DHHS Office of Research Integrity adopted and published the final
PHS Policy on Instruction in the Responsible Conduct of Research that delineates RCR training
requirements to all research investigators applying for or using PHS funds and their institutions (3).
Although nine core areas of instruction are specified, the policy does not establish the exact content in
the form of standards and principles within each area. In complying with this mandate, each
institution will be responsible for its own content.
Much attention in the RCR literature has been directed to standards within specific areas, such as
authorship, peer review, and collaborative practices. Presentations at national conferences and
Corresponding author: Gregory Brock, Ph.D., Department of Family Studies, University of Kentucky, 316 Funkhouser
Building, Lexington, KY 40506-0054, 859-257-7742 (voice), 859-257-321 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
institutional committees have addressed RCR
practice standards. As well, many professional
associations have established standards of
conduct within their ethical codes. Institutional
policies such as Guidelines for the Conduct of
Research at the National Institute of Health have
also incorporated a selection of RCR topics (4).
However, no single set of principles
encompassing all aspects of responsible conduct
of research exists in unified form.
Grinnell (5) pointed out that “…promoting
responsible conduct of science requires a clear
description of what doing science entails.” In
addressing why standards are important, Frankel
(6) discussed the need of the general public for
accountability in science, and how a set of
standards not only meets this need but also
increases trust in the scientific community.
Frankel noted specific benefits to establishing
ethical standards: Standards provide an enabling
document, professional socialization, public
accountability, gain public trust/support, improve
public relations, self-preservation, deterrence,
professional support, and are a source of public
policy. Standards also provide guidance when an
ethical course of action is unclear. Mastroianni
and Kahn (7) point out that training students in
the basics of RCR is crucial to the continued
maintenance of public trust in the scientific
community by cultivating the integrity of
research practices. However, results on the
effectiveness of RCR training thus far are
inconclusive (8, 9). Brown and Kalichman (9)
offer the interpretation that a lack of consensus
on what constitutes misconduct may contribute to
the lack of clarity on the effectiveness of training.
Frankel (10) advocates the development of
research standards as the single most important
step in promoting scientific integrity and
handling misconduct. Faced with the new
training requirements established by the PHS,
this step is particularly important for promoting
and supporting a climate of integrity at the
organizational level that can function in a
reciprocal fashion to influence and be influenced
by individual actions.
Initially, the purpose of the document
presented here was to provide a comprehensive
set of guiding principles to serve as a basis for
RCR training at the University of Kentucky.
Content analysis was applied to an exhaustive list
of behavioral guidelines identified in a thorough
review of the research integrity literature
including ethics codes of professional
associations. Guidelines were then sorted into
discrete thematic categories. These categories
were called principles because they identified
core values of research practice. Three groups of
principles emerged from the analysis: General,
Professional, and Focused. Subprinciples also
were defined that served to elucidate
contemporary issues rather than merely
exemplifying situations in which the principles
might apply. A series of revisions were made
after obtaining feedback from research
colleagues and university administrators.
What emerged was a comprehensive set of
guidelines for the conduct of researchers more
akin to a code of conduct for a profession (see
attached guidelines). These guidelines provide a
broad-based foundation for the safe and effective
practice of research across disciplines, settings,
methods, and questions. Our intent in presenting
them here is to increase the awareness and
sensitivity of institutional policy makers to the
many issues that researchers must attend to in the
conduct of their professional responsibilities. By
presenting the results of our analysis, we wish to
further the discussion about the content of RCR
training.
Acknowledgements:
Support was provided by the Vice President for
Research and Graduate Studies, University of
Kentucky.
References
1. Department of Health and Human Services, Public
Health Service. Responsibilities of Awardee and
Applicant Institutions for Dealing With and Reporting
Possible Misconduct in Science. Federal Register Vol.
54 No. 151 Tuesday, August 8, 1989.
2. Steneck, N. Assessing the Integrity of Publicly Funded
Research: A Background Report. Paper prepared for the
Department of Health and Human Services, Office on
Research Integrity Research Conference on Research
Integrity on November 2000, Bethesda, MD.
3. PHS Policy on Instruction in the Responsible Conduct
of Research, Department of Health and Human
Services, Public Health Service, Office of Research
Integrity. Available at URL: http://ori.dhhs.gov/html/
programs/rcrcontents.asp
4. American Association of Medical Colleges. Developing
a code of ethics in research: A guide for scientific
societies, executive summary. Conference Materials of
the Public Responsibility in Medicine and Research;
1997 May 13-14; Bethesda, Maryland.
5. Grinnell, F. Ambiguity, trust, and the responsible
conduct of research. Science and Engineering Ethics,
1999; 5 (2): 205-214.
58
––––––––––––––––––––––––––––––––––– Brock et al., Comprehensive Guidelines for the Responsible Conduct of Researchers
6. Frankel, M. S. Developing ethical standards for
responsible research: Why? Form? Functions? Process?
Outcomes?, Journal of Dental Research 1996; 75 (2):
832-835.
7. Mastroianni, A. C., & Kahn, J. P. The importance of
expanding current training in responsible conduct of
research. Academic Medicine 1998; 13 (12): 1249-1254.
8. Eastwood, S., Derish, P., Leash, E., & Ordway, S.
Ethical issues in biomedical research: perceptions and
practices of postdoctoral research fellows responding to
a survey. Science and Engineering Ethics 1996; 2: 89114.
9. Brown & Kalichman. Effects of training in RCR: A
survey of graduate students in experimental sciences.
Science and Engineering Ethics 1998; 4: 487-498.
10. Frankel, M. S. Scientific community must set the
standards. Forum for Applied Research and Public
Policy, Spring 1998
Guidelines for the Responsible Conduct of Researchers
Preamble: Advancing the scientific record is the noble task of those who conduct research. In large
part the quality of that record is the product of inquiry. Ranging well beyond the conduct of research
however is the realm of activities constituting the work of researchers that influences the public trust,
that affects global well-being, and that indirectly affects the scientific record. The guidelines
presented here define expectations so that researchers uphold the highest ethical standards by
practicing within the bounds of both effectiveness and safety.
Important, sustaining values that support humankind and global well-being serve as the basis for
three groups of principles and sub-principles. (1) General principles apply to all research contexts.
(2) Professional principles define relations among researchers and practices that constitute the
scientific method. (3) Focused-principles address discrete aspects of research practice for particular
investigations, research contexts, or scientific disciplines. Sub-principles elucidate contemporary
issues rather than identifying the component issues of any principle.
Where governmental laws contradict these guidelines, researchers are cautioned to seek
consultation from appropriate authorities and colleagues. Resolution is not always possible,
consequently, researchers act so as to benefit the greater good even if that path demands personal
sacrifices.
In an effort to create a research climate worthy of the public trust, it is incumbent upon
researchers to report any breech of these guidelines to an appropriate authority. Where there is no
relevant authority, researchers are obliged to focus public media attention on wrong doing.
These guidelines apply to professional and amateur researchers, students, research technicians,
research administrators, as well as private, public, and governmental research agency personnel.
General Principles
General Principle 1: Commitment to Society and to Global Well-being
Researchers protect the interests of society within a broader commitment to global well-being. They
recognize that the public has entrusted them to uphold the integrity of the scientific record.
1.1 Researchers do not obligate themselves to withhold research findings that may jeopardize the
health or well-being of others.
1.2 Researchers take active steps to prevent the misuse of their findings that may jeopardize the
well-being of others.
1.3 Researchers take active steps to correct errors or oversights in proposing, conducting, or
reporting research.
1.4 Researchers present themselves to the public in a competent, sincere, and trustworthy manner.
59
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
General Principle 2: Commitment to Competency
Researchers are aware they are responsible for maintaining professional competency and remaining
knowledgeable within their areas of expertise.
2.1 Researchers conduct their work within the scope of their own training and knowledge base.
2.2 Researchers recognize they are vulnerable to stress and impairment. When stress or impairment interferes with their ability to conduct professional responsibilities, researchers seek
assistance.
2.3 Researchers ensure that all persons who assist in the conduct of their research are adequately
trained and perform their responsibilities competently.
2.4 Researchers inform their work with views, values, and co-workers from diverse sources.
2.5 Researchers foster a scientific community in which discrimination based on gender, race age,
sexual orientation, religious affiliation, ethnic or national origin does not occur.
General Principle 3: Understanding Laws, Regulations, and Policies
Researchers are aware of and stay informed of professional, institutional, and governmental
regulations and policies in proposing, conducting, and reporting research.
3.1 Researchers take active steps to resolve discrepancies when policies or regulations are
unclear or contradict one another.
General Principle 4: Conflicts of Interests
Researchers are cognizant that conflicts of interest occur in the context of professional activities and
they recognize and avoid them.
4.1 When researchers cannot avoid real or perceived conflicts of interest, they seek consultation
and take active steps to minimize bias, flawed judgment, harm, or exploitation.
Professional Principles
Professional Principle 5: Peer Review
Researchers respect others’ rights to have work reviewed in a confidential, timely, and objective
manner.
5.1 Researchers assess and disclose multiple roles or allegiances which may undermine the
confidential and fair review of others’ work.
5.2 Researchers take active steps to protect the integrity of review materials and guard the
intellectual property of others.
Professional Principle 6: Research Management and Data Access
Researchers clearly and authentically record data and methods. They protect the integrity of their
research materials. They make data, methods, and materials available to others for analysis or
replication.
6.1 Researchers select materials appropriate for data acquisition, recording, and storage.
6.2 Researchers stay informed of and implement policies for appropriate storage and disposal of
research materials.
6.3 Researchers take active steps to select methods and materials that protect research participants’ right to privacy.
6.4 Researchers take active steps to safeguard data when using electronic or Internet-based
methods.
6.5 Researchers are cognizant of the ownership of their research data, methods, and findings.
Professional Principle 7: Commitment to Credibility
Researchers engage in practices that are currently accepted within the scientific community to
propose, conduct, and report research.
60
––––––––––––––––––––––––––––––––––– Brock et al., Comprehensive Guidelines for the Responsible Conduct of Researchers
7.1 Researchers practice honest stewardship of their research resources and use recognized
accounting methods.
7.2 Researchers do not conduct their professional responsibilities in a manner that is intentionally
deceitful or with reckless disregard for the truth.
7.3 Researchers who witness or suspect fraud or misconduct follow established procedures to
preserve the integrity of the scientific record.
7.4 Researchers accused of fraud or misconduct do not harass those believed or known to have
made accusations against them.
7.5 Researchers do not misrepresent their work by omitting data that changes the meaning or
significance of their findings.
7.6 Researchers do not fabricate or falsify data.
7.7 Researchers do not present or publish component findings of a larger body of work if misunderstanding may result or to conceal findings.
Professional Principle 8: Mentoring, Training, and Supervisory Relationships
Researchers nurture the intellectual, technical, ethical, and career development of their trainees,
supervisees, and students.
8.1 Researchers recognize that trainees, supervisees, and students have needs unique to their
individual strengths and limitations. Researchers provide guidance, constructive feedback,
and assistance that matches the changing needs of each trainee, supervisee, or student.
8.2 Researchers establish clear and appropriate rules and boundaries in their relationships with
trainees, supervisees, and students.
8.3 Researchers do not engage in sexual harassment, disrespect the character of, or impede the
progress of their trainees, supervisees, and students.
8.4 Researchers recognize that exploitation is a risk in relationships where differences in power
exist. They avoid conflicts of interest and dual relationships. Sexual interaction with subordinates is avoided.
8.5 Researchers take active steps to inform trainees, supervisees and students of supervisors’
responsibilities to avoid dual relationships.
Professional Principle 9: Authorship and Publication Practices
Researchers respect the intellectual property rights of others.
9.1 Researchers attribute credit for others’ words and/or ideas in proposing, conducting, or
reporting their own work.
9.2 Researchers facilitate discussion and set ground rules early in collaborative relationships
regarding authorship assignment.
9.3 Researchers assume responsibility for the accuracy of research reports for which they claim
full or co-authorship.
9.4 Researchers preserve the integrity of the scientific record by taking active steps to correct
errors in the publication of their findings.
9.5 Researchers do not submit or publish previously published materials without appropriate
citation.
9.6 Researchers respect the privacy of others’ unpublished work.
Professional Principle 10: Responsibilities to Colleagues and Peers
Researchers recognize they are members of the scientific community and respect the contributions of
others to the scientific record.
10.1 Researchers clarify early in a collaborative project the expectations and responsibilities
among those involved.
61
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
10.2 Researchers do not impede the progress of others’ work.
10.3 Researchers protect the integrity of intellectual property and research materials when
reviewing others’ work.
10.4 Researchers take active steps to maintain positive relations among team members and to
seek consultation if necessary to resolve interpersonal conflicts.
Focused Principles
Focused Principle 11: Protection of Human Participants
Researchers respect the dignity of human participants and take active steps to protect their wellbeing. They follow institutional, professional association, and governmental ethical and
regulatory guidelines.
11.1 Researchers ensure that each participant gives voluntary and informed consent regardless of
age, race, gender, ethnic or national origin, sexual orientation, mental or physical health
status, or incarceration.
11.2 Researchers take active steps to evaluate and to minimize potential risks to participants.
11.3 Researchers respect each participant’s right to privacy, and they take active steps to protect
confidentiality of data or other disclosures.
11.4 Researchers take active steps to achieve an equitable balance of benefits and risks to each
participant.
11.5 Researchers honor fairness and equity in the selection of research participants.
Focused Principle 12: Care and Use of Animals for Research
Researchers are stewards of animals used for research. They follow institutional, professional
association, and governmental ethical and regulatory guidelines.
12.1 Researchers substitute inanimate materials and processes for animals where appropriate.
When this is not possible, researchers make active efforts to use species that may be less
susceptible to pain and distress.
12.2 Researchers take active steps to use procedures which reduce the incidence and/or severity
of pain and distress experienced by animals.
12.3 Researchers take active steps to reduce the use of animals to the minimum number necessary to yield valid answers to their research questions.
Focused Principle 13: Commitment to Native Populations and Other Identifiable Groups
Researchers respect the rights and protect the interests of Native populations and other
identifiable groups.
13.1 Researchers who work with Native populations and other identifiable groups recognize that
to minimize risks and to maximize benefits to individuals and to populations themselves
there is value in obtaining the advice, participation, and viewpoints of those individuals and
populations in formulating research questions, designing research methods, collecting and
analyzing data, and in reporting results.
13.2 Researchers recognize that consent from or consultation with group authorities or representatives is sometimes necessary before obtaining consent from individuals within Native
populations or other identifiable groups.
13.3 Researchers take active steps to distinguish individual property both tangible and intangible
from collective property owned by Native populations or other identifiable groups.
13.4 Researchers take active steps to reduce the risk to Native populations or other identifiable
groups that result from misuse of their research findings.
62
––––––––––––––––––––––––––––––––––– Brock et al., Comprehensive Guidelines for the Responsible Conduct of Researchers
Focused Principle 14: Genetic Research and Technology
Researchers strive to preserve and protect global well-being from the unintended consequences of
genetic research.
14.1 Researchers involved in genetic research take active steps to identify potential risks and
benefits to research participants. They inform participants of the possibility that risks may
not yet be identified.
14.2 Researchers take active steps to protect the confidentiality of genetic materials collected
from human participants and do not allow the use of these materials for purposes which
may discriminate against or harm an individual or group of individuals.
14.3 Researchers are sensitive to social, physical, psychological and environmental factors that
may influence individuals’ consent to participate in genetic research.
14.4 Researchers inform individuals, their families, and Native
and other identifiable populations of the disruptive influence that genetic research may have on
their lives. They take active steps to minimize disruptions.
14.5 Researchers are cognizant of the increasing complexity of the ethical concerns about genetic
research. They stay informed of the developing research guidelines as well as the public
discourse about genetic research.
14.6 Researchers actively participate in the development and refinement of ethical standards in
this area.
63
Research Integrity in Social Work: Status, Issues, and Future Directions
Margaret Gibelman, Wurzweiler School of Social Work, Yeshiva University, USA
Keywords: Scientific integrity, social work, code of ethics, scientific misconduct, institutional review boards
This paper explores the issue of scientific integrity in social work and its implications for the training
of social work researchers. Data concerning a growing body of cases in which allegations have been
made and/or violation of legal and ethical research standards have been substantiated illustrate that
the integrity of research in social work and related fields is a growing concern. However,
mechanisms to review and monitor social work research are under-developed compared to other
disciplines. A research agenda is offered to assess the status of institutional systems to review and
monitor research in social work and, concurrently, determine social workers’ familiarity with the
profession’s ethical code as it relates to research integrity. Implications for faculty and practitioner
education and training and the development and enforcement of systems to review the integrity of
research protocols are explored.
Scientific misconduct or, more positively, appropriate conduct in the realm of research inquiry, is
a topic that has received very little attention in the social work literature. Unfortunately, this is
because social workers have not, historically, been strong contenders in the successful competition for
federal research grants, particularly large-scale research protocols (1, 2, 3, 4). Social work research is
still in its infancy compared to research in other disciplines. However, there is a professional
commitment to increase the capacity and productivity of social work research, as evidenced by the
burgeoning number of social work research centers and a growing empirical social work literature
base. This expansion of social work research is not without risks. Although the majority of publicized
cases of scientific misconduct have centered largely on bio-medical research and the applied sciences,
the circumstances associated with these cases have strong implications for the preparation of students
and the standards to which social work researchers will be held. The growing number of cases in
fields related to social work, as discussed below, highlight areas of potential vulnerability.
The Status of Social Work Research
Unlike most of the social and behavioral sciences, social work is a practice-based profession rather
than an academic discipline or field. Social work has been defined as the “applied science of helping
people achieve an effective level of psychosocial functioning and effecting societal changes to
enhance the well-being of all people” (5). Historically, its knowledge base has been predicated upon a
liberal arts perspective and has drawn from psychology, psychiatry, sociology, political science,
economics, and other disciplines to formulate applied practice principles. However, within the past
two decades, social work has striven to define its own unique body of knowledge, an effort
incorporated into the purposes of social work itself, one of which is “the development and testing of
professional knowledge and skills...” (6).
Corresponding author: Margaret Gibelman, Yeshiva University, Wurzweiler School of Social Work, 2495 Amsterdam
Avenue, New York, NY 10033, 212-960-0840 (voice), 212-960-0822 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Although research has always had a place
within the purposes of the profession, the larger
socio-political environment has, in recent years,
profoundly affected the priority afforded to
research. There is a growing mandate for all
social workers to incorporate research into their
practice, a phenomenon underscored by the
demands of funding bodies, oversight agencies,
and consumer choice movements for hard data
documenting that programs of service lead to
tangible results. A leading force has been that of
managed care, which has brought with it
heightened demands for accountability, with
particular emphasis on documenting the
successful outcomes of service (7).
At the same time that external demands to
provide empirical evidence of the impact and
outcomes of services grow, social workers, to
better protect the interests and well-being of the
people they serve, are seeking to empirically
examine the consequences of the managed care
movement, itself. This has translated to a concern
about documenting the effects of managed care
(e.g., short-term hospitalization; short-term
treatment; limited provider choice). These
developments have led to the need for a new or
enhanced repetoire of research skills on the part
of not only academics and researchers, but
among the totality of social workers directly
providing, supervising, or managing the delivery
of human services.
The long and ongoing admonishment that the
profession must develop an internal research
capacity has borne fruit. In fact, a notable
number of studies have been conducted on the
status of research productivity and the scholarly
contributions of social workers (8, 9, 10, 11, 12,
13). Perhaps the most significant influence,
however, on the growing social work research
enterprise has been the shift in criteria for tenure
and promotion within academia, which remains
the richest source of social work research (14,
15). Longevity of academic careers now rests
firmly on scholarly productivity and standards
related to both quality and quantity continue to
rise as social work is increasingly held to the
same standards as other academic and
professional units within the university (4). A
related factor in the emphasis on research
productivity is the growing sophistication of
faculty in identifying funding sources and
competing successfully for publicly supported
research dollars.
The emergence of schools of social work as
major research centers and the increased
productivity of social work researchers has been
long in coming. The mandate to create a
coordinated research infrastructure had been
echoed for two decades (16, 17, 18, 19).
National Institute of Mental Health funding has
been a major impetus to establish social work
research centers at academic institutions. In this
process, however, the profession faces a host of
issues and challenges, foremost among them the
preparation of future researchers, including
socialization to the ethos of scientific integrity.
Ethical Guidelines
The latest revision of the Code of Ethics of the
National Association of Social Workers
(NASW)(20) emphasizes the central role of
research: “social workers should contribute to
the knowledge base of social work and share
with colleagues their knowledge related to
practice, research, and ethics. Social workers
should seek to contribute to the profession’s
literature and to share their knowledge at
professional meetings and conferences” (Section
5.01(d), p. 24). Section 5.02 (b) of the Code
(1996) encourages social workers to “promote
and facilitate evaluation and research to
contribute to the development of knowledge”
(p. 25).
The Code of Ethics not only seeks to
establish an obligation on the part of social
workers to engage in knowledge building
through empirical research, but also provides the
basic guidelines for how such research is to be
conducted. Specific provisions pertain to riskbenefit analysis, voluntary and written informed
consent, protection from harm, confidentiality,
and accurate reporting of findings. Further, the
Code sets forth the obligation of social workers
to educate themselves and for programs of social
work education to provide relevant education
concerning responsible research practices.
An important caveat about ethical guidelines
exists that is idiosyncratic to the profession —
the limited application of the Code to social
workers. The Bureau of Labor Statistics (21)
estimates that there are approximately 542,000
professional educated social workers in the
United States (at the bachelor’s, master’s and
doctoral levels). At the same time, current
membership of the National Association of
Social Workers is approximately 155,000. The
Code of Ethics is a product of the National
Association of Social Workers and, upon joining,
66
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gibelman, Research Integrity in Social Work
ethics and that there had not been an explicit
violation of the Code of Ethics (24). However,
as social workers increasingly compete
successfully for federal research funds, they
become subject to the same level of scrutiny as
researchers in other disciplines. Similarly, as
IRBs extend their purview to include privately
supported research, more diligent reviews of
social work research protocols can be expected.
As the social work profession seeks to
enhance its research capability in a credible and
responsible manner, there is much to be learned
from the experience of related disciplines and
professions. In recent years there has been a
growing number of cases of scientific
misconduct among allied health-related
industries (e.g., nursing, psychology, and
psychiatry), the predominant theme of which
concerns plagiarism and/or falsification or
fabrication of data (25, 26, 27, 28, 29). Eight
cases from the helping professions over the last
decade were identified from media reports, out of
an unknown universe of substantiated cases of
misconduct. Unlike many cases of misconduct
substantiated in the bio-medical fields, these
cases were absent allegations of human subjects
violations. However, findings of misconduct
highlight the diligent reviews to which research
reports are subject and the serious penalties that
are levied when ideas are appropriated or results
falsified. Sanctions include forced resignations,
criminal prosecution, ineligibility from receiving
publicly supported grants or serving on review
panels, and remedial courses in ethics. These
sanctions have widespread and serious
implications for how research is conducted and
highlight the potential consequences that may
ensue when procedural and ethical breaches are
uncovered.
members must pledge to abide by the Code. But
what about the more than 387,000 social workers
who are not members of NASW and not
committed to abiding by the provisions of the
Code? These social workers may belong to other
professional associations which have their own
ethical guidelines, but data to support this
contention are lacking (22). Social work
researchers based in institutions of higher
education may have their own review and
oversight procedures, separate from universitywide IRBs, but again there is an absence of
substantiating empirical data. An unknown, but
impressionistically high proportion of social
work research is outside the purview of federal
funding, which may mean that IRB review
procedures are not applied. (It should be noted,
however, that such research is now selectively
being reviewed by IRBs to conform with their
own internal procedures, partially reflecting the
prevalence and influence of the growing number
of studies sponsored by private sources,
including pharmaceutical companies, in areas
such as genetic testing (23).)
Finally, in some instances, social work
research may be absent of any oversight by any
source. This latter scenario is most likely to
prevail among those working in service
organizations which have not yet established
review and oversight procedures and may,
indeed, not even recognize the need to do so. Of
particular concern is the mandate for practice
agencies to engage in research without
assurances of appropriate procedures and absent
collaborations with educational institutions from
which such protocols may be borrowed.
Learning from the Mistakes of Others
To date, public disclosure of cases of scientific
misconduct within the social work research
community have been absent. Over a 10 year
period of vigilant reporting of scientific
misconduct, the Chronicle of Higher Education
referenced only one situation involving a social
worker. This case concerned a researcher who
submitted bogus articles to professional journals
as part of an experiment to test peer-review
practices (24). Because the research did not
involve the use of Federal funds, review of
allegations of ethical misconduct remained
within the purview of the adjudication process of
the NASW. Ultimately, NASW dismissed the
complaint, arguing that the issue involved a
disagreement over research methods rather than
Emerging Issues
The mistakes of researchers of allied disciplines
suggest the scope and magnitude of potential
areas of scientific misconduct that may similarly
affect social work. Further, the record on
misconduct shows that attention to the initial
review of protocols is only a beginning step in an
ongoing process necessary to ensure scientific
integrity. Although a systematic process for
reviewing research proposals, including attention
to scientific validity of the study design, can
alleviate many potential problems, it is in the
reporting of research findings, at least to date,
that the allegations of scientific misconduct are
67
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
most likely to occur. Reports of research are, in
fact, reviewed; how research is carried out and
findings reported are subject to scrutiny, and,
sometimes, reprisals. This fact presents a
formidable problem in balancing the traditional
academic freedom associated with the pursuit of
research and institutional responsibility to ensure
accountability of the outcomes of such research.
The extent to which a school of social work can
monitor the work of its faculty and students is
inherently limited.
While only about 30% of the cases of
scientific misconduct are eventually determined
to be founded, the impact of the allegations is
profound (30). The investigation of allegations
consumes significant institutional resources and
can ruin careers, even if the allegations are
unfounded. If allegations are confirmed, it is
lethal to a researcher’s career (see, for example,
31), causes reputational damage to the university,
and may affect public perceptions of the integrity
of all research. Worse, human lives and wellbeing may be compromised (4).
Internal systems to prevent and, when
necessary, address scientific misconduct are not
without their critics. There are enormous
workload implications, particularly for senior
faculty who may not have the time or desire to
spend their time monitoring junior faculty. There
are also those who argue that when schools/
universities serve as the “scientific validity
police” of their own colleagues, they will either
join ranks in defense, or, to the other extreme,
find against their colleagues for fear of
accusations of institutional bias (32, 33).
Current Review Mechanisms
Since allegations and, in some cases, findings of
scientific misconduct are, by definition, after-thefact of the activity, the most significant lesson
from these cases is the importance of ensuring
that research review and monitoring procedures
are uniformly followed. The integrity of
scientific research is monitored by two main and
distinct sources: professional associations and
their applicable ethical codes and institutional
review boards (IRBs). In social work, these
mechanisms for ensuring research integrity are
less firmly entrenched. As discussed earlier, there
is no one body with the authority or jurisdiction
to oversee the entirety of the social work research
enterprise. The guidelines detailed in the
profession’s Code of Ethics about ethical
research conduct are, however, limited by their
lack of applicability to a large proportion of
social workers. Social work educators, who are
the major producers of research, are illrepresented among the membership of NASW
and are thus outside of its monitoring and
adjudication provisions. Thus, the question of
what mechanisms govern academic social work
research remains unanswered.
The majority of schools of social work are
housed in research universities which have their
own IRBs and the logical source of research
review and oversight lies with IRBs. However,
the focus of many, if not most, IRBs on biomedical research, with the composition of IRBs
reflecting this emphasis, has limited the informed
review of social work protocols. Social and
behavioral science research protocols, including
those of social work, are often “expedited” and/
or are reviewed by researchers who are
unfamiliar with the nature of such scientific
inquiries. (An analogy holds when social and
behavioral scientists are asked to participate on
IRBs in the review of bio-medical research.)
Without the procedures in place and a cadre of
trained researchers available and able to review
social work research protocols, social work may
well be vulnerable to some of the questionable
research practices that have been unearthed in
related fields.
The expanding boundaries of what
constitutes scientific integrity are of particular
relevance to social work researchers. The
research conducted by social workers, both
students and faculty and agency-based
practitioners, involves interaction with
populations that are often classified as vulnerable
and confidentiality of data is often an issue.
Direct observations, the administration of
questionnaires, review of existing case records,
or the introduction of therapeutic interventions
and the use of control groups that do not receive
interventions may be innocuous or, alternatively,
may pose risks to the emotional, social, or
economic well being of participants (4).
Deception, invasion of privacy, lack of informed
consent, mandatory reporting requirements (such
as cases in which potential child abuse is
identified), or the loss of economic benefits (as
may apply, for example, to the disabled or
welfare recipients) are all examples of harm that
may result from faulty research designs or
misconduct in the implementation of research
protocols (4). Although substantiated cases to
date fall outside of these human protection areas,
68
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gibelman, Research Integrity in Social Work
the nature of the research conducted within the
helping professions suggests the potential of such
misconduct.
A Call for Research on Social Work
Research
Given the relatively undeveloped, but now
rapidly expanding research enterprise in social
work, there is a clear need for information about
how research is monitored and reviewed. The
number of publicized cases of wrongdoing in
fields closely allied with social work suggest that
programs of social work education need to
formulate or revise their procedures for research
review and oversight. Institutional mechanisms
are needed to ensure that: (1) researchers are
cognizant of the ethical issues involved; (2) the
protocols meet university and Federal standards;
and (3) findings are based on systematic and
valid research. The question then becomes
whose responsibility it is to monitor such
protocols and review the research conducted and
how mechanisms can be established which
significantly reduce the potentiality of scientific
misconduct.
Some schools have assembled their own
committees to review and pass judgment about
compliance with university and/or federal
research requirements. However, such reviews
usually focus on issues of methodology and/or
informed consent. This is not sufficient given the
broadened definition of scientific misconduct,
which has been extended beyond the initial focus
on informed consent, risk levels, and coercion
(34). The definition of misconduct now includes
fabrication, falsification, or plagiarism in
proposing, performing, or reviewing research, or
in reporting research results” (35, p. 4). The
extent to which social work education programs
maintain their own review and oversight
procedures is also unknown. Anecdotal evidence
suggests that such internal program mechanisms
are the exception. Given the limited applicable
of the professional Code of Ethics, the unknown
degree of inclusion of social work within the
purview of IRBs, and the similarly unknown
degree of school-specific procedures, the need
for “research on the status of social work
research” is suggested. Possible areas of inquiry
include:
•
•
•
•
•
provisions regarding appropriate scientific conduct,
through such means as: (1) an “exit” test of graduating
BSW, MSW and doctoral students; (2) a sample
survey of agency practitioners; and (3) a sample
survey of agency administrators charged with
responsibility to collect, analyze, and report on clientsensitive data.
An analysis, perhaps through the use of focus groups, of
issues and obstacles to the conduct of ethical research
which result from the demands of external
accountability bodies.
An investigation of the procedures used by schools of
social work to review and monitor faculty and student
research, including the scope of such reviews and the
extent to which the validity of the science itself is
considered.
A survey of social work faculty concerning their level of
participation in university-wide institutional review
boards.
A survey of deans and directors of social work
education programs to identify the frequency, nature,
and types of issues and problems that have arisen in
regard to studies, once approved and implemented.
A content analysis of material covered in federally
prescribed training of researchers and an assessment
of the applicability of such training to the social and
behavioral sciences.
The data emanating from such studies would
provide a basis for an informed assessment of the
extent to which mechanisms for research review
and monitoring are in place and how well they
operate. Such information could form the basis
for developing or revising review procedures
through university IRBs, through separate IRBs
potentially established for the social and behavior
sciences, or through social work educationspecific structures. Further, such information
could be used to develop targeted educational
programs about research integrity to the social
work community.
Conclusion
Research about social work research has tended
to be descriptive, often focused on admonishments about the under-developed state of the art
or analyses of what content areas have been
researched and what gaps exist. Ethical research
conduct has, by and large, been ignored, in part
because of the early stage of development of the
research enterprise. However, the issue of
research integrity takes on increasing importance
as social work gains a legitimate role in the
conduct of scientific inquiry. The profession is
likely to experience a stronger imperative to
engage in research as demands for accountability
and documentation of the outcomes of human
services continue to grow.
• An analysis of the social work education curriculum to
ascertain the degree to which ethical conduct is a
component of research courses.
• An assessment of social workers’ familiarity with ethical
69
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Strategies to ensure research integrity
depend, first, on a clearly formulated agenda
based on an assessment of the current status of
review and monitoring systems. Based on hard
data, the professional schools of social work and
their universities can assume the task of
modifying and strengthening procedures in a
manner that is reflective of the burgeoning social
work research enterprise. Means of prevention
as well as amelioration need to be developed,
codified, and enforced. In this process, there is a
need to define the parameters of both appropriate
scientific conduct and what constitutes
misconduct as it relates to social work research
and to elaborate on its meaning with some degree
of precision. Clear university and school
standards, widely publicized, and ongoing
education regarding appropriate scientific
conduct would help alleviate actual or potential
problems as social work secures a more extensive
and important role in the production of research.
Bibliography
1. Austin, D. Research and social work: educational
paradoxes and possibilities. J Soc Sci Res 1979; 2: 15976.
2. Fanshel, D. National Conference on the future of social
work research: Final report. Washington, DC: National
Association of Social Workers, 1979. Contract # 27877-0079[SM]. Sponsored by the National Institute for
Mental Health.
3. Rubin, A., Gibelman, M. Social work research in
mental health: the state of the art. Rockville, MD:
Division of Human Resources/Division of Prevention
and Special Mental Health Programs, National Institute
of Mental Health; 1984. Report no.: 83M055981901D
& 83M052517801D.
4. Gibelman, M., Gelman, S.R. Learning from the
mistakes of others: a look at scientific misconduct
within academia. J Soc Work Educ 2001; 37(2): 241254..
5. Barker, R.L. Social work dictionary. 4th ed.
Washington, DC: NASW Press; 1999.
6. Council on Social Work Education. Handbook of
accreditation standards and procedures. Alexandria
(VA): CSWE; 1994.
7. Gibelman, M., Whiting, L. Negotiating and contracting
in a managed care environment: considerations for
practitioners. Health Soc Work 1999; 24: 180-90.
8. Baker, D.R., Wilson, M.V.K. An evaluation of the
scholarly productivity of doctoral graduates. J Soc
Work Educ 1992; 28: 204-12.
9. Bloom, M., Klein, W.C. Publications and citations: a
study of faculty at leading schools of social work. J Soc
Work Edu 1995; 31: 377-87.
10. Gelman, S.R., Gibelman, M. A quest for citations?: An
analysis and commentary on the trend toward multiple
authorship. J Soc Work Educ 1999; 35: 203-13.
11. Green, R.G., Baskind, F.R., Best, A.M., Boyd, A.
Getting beyond the productivity gap: assessing variation
in social work scholarship. J Soc Work Educ 1997; 33:
541-53.
12. Ligon, J., Thyer, B.A., Dixon, D. Academic affiliations
of those published in social work journals: a
productivity analysis, 1989-1993. J Soc Work Educ
1995; 31: 369-76.
13. Smith, S., Baker, D., Campbell, M., Cunningham, M.
An exploration of the factors shaping the scholarly
productivity of social work academicians. J Soc Serv
Res, 1985; 8: 81-99.
14. Marsh, J.C. Should scholarly productivity be the
primary criterion for tenure decisions? Yes! J Soc Work
Educ 1992a; 28: 132-4.
15. Marsh, J.C Response to Dean Weick. J Soc Work
Educ, 1992b; 28: 138-9.
16. Austin, D. Findings of the NIMH task force on social
work research. Res on Soc Work Pract 1992; 2: 311322.
17. Council on Social Work Education. Research utilization
in social work education. New York: CSWE; 1981.
18. Council on Social Work Education. A planning proposal
to establish a National Institute for Social Work
Research. New York: CSWE; 1983.
19. Estes, R.J. Social work research centers. Soc Work
Res Abstr 1979; 15: 3-16.
20. National Association of Social Workers. Code of ethics.
Washington, DC: NASW; 1996.
21. U.S. Department of Labor, Bureau of Labor Statistics.
Tabulations from the Current Population Survey, 1998
annual averages. Washington, DC: BLS; 1998.
22. Gibelman, M., Schervish, P. Who we are: a second
look. Washington, DC: NASW Press; 1997.
23. IRBs and human stem cell research protocols. Hum Res
Rep 2000 Jan 15; 3-4.
24. Social workers’ group plans no disciplinary action
against researcher who submitted bogus articles. Chron
High Educ 1989 Apr 15; Sect. A: 5.
25. Burd, S. Nursing professor found guilty of scientific
misconduct. Chron High Educ 1995 July 7; Sect. A:23.
26. Burd, S. Research powerhouse under the microscope.
Chron High Educ 1994 Jun 15; Sect. A: 24.
27. Campbell, P.W. Researcher found guilty of misconduct.
Chron High Educ 1999 Jan 15; Sect. A:34.
28. Walker, P.V. Public health service finds 3 scientists
guilty of misconduct. Chron High Educ 1996 Jan 12;
Sect. A: 26
29. Wheeler, D.L. Suggestions offered for investigating
reports of biomedical-research fraud. Chron High Educ
1991 Nov 27; Sect. A: 6.
30. Goodman, B. Scientists exonerated by ORI report
lingering wounds. Scientist 1997, June 9; 11: 124.
31. Schmidt, P. U. of Arizona fires prominent researcher
over charges of scientific misconduct. Chron High Educ
1998 Aug 14; Sect. A: 12.
70
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gibelman, Research Integrity in Social Work
32. Block, L.E. Science’s credibility depends on effective
peer review. Chron High Educ 1991 July 31; Sect. B:2.
33. Silbergeld, E.K. Annotation: Protection of the public
interest, allegations of scientific misconduct, and the
Needleman case. Am J Public Health 1995; 85: 165-6.
34. Broad, W.J. Data tying cancer to electric power found
to be false. New York Times 1999 July 24; Sect. A:1,
A10.
35. New definition proposed for research misconduct. Hum
Res Rep 1999 Dec 14; 4.
71
Organizational Influences on Scientific Integrity
Michael D. Mumford, Department of Psychology, University of Oklahoma, USA
Whitney B. Helton, University of Oklahoma, University of Oklahoma, USA
Keywords: Ethical behavior, Integrity, Organizational influences, Scientific integrity, Situational influences
Our image of the working scientist remains inherently romantic (1). We envision an individual,
working alone, pursuing knowledge in an area solely for its intrinsic interest. As attractive as the
image may be, it has little to do with the realities of current work in the sciences (2, 3, 4). Scientists
work in a distinctly social setting, conducting their work in both collaboration and competition with
others (5, 6). This work, moreover, occurs in organizational settings, including business, government
and academia. Thus, the pressures that face people working in any organization – pressures of time,
conformity, resources, and production – also confront scientists.
Although one might argue that scientists, by virtue of their work, are granted more autonomy and
are carefully buffered from the more “ugly” demands of organizational life, the conditions currently
confronting most scientific endeavors are such that we can expect organizational pressures to become
a progressively more important influence on scientific work. The emerging forces of the new
economy, where innovation is the true competitive edge, move scientists from the periphery of the
business world to the heart of the industrial enterprise (7). Academia, moreover, under the financial
pressures imposed by funding cutbacks, has placed a new emphasis on responding to the needs of the
business community (8). Finally, academia has begun a slow process, for good or ill, of learning how
to manage itself differently, and manage itself like a business.
Given these pressures, there is a need to understand how organizational variables influence
scientific integrity. Unfortunately, systematic studies of scientific integrity are virtually nonexistent.
However, a number of scholars have sought to understand the variables that influence integrity in
organizational settings as a general phenomenon. Accordingly, our intent in the present study is to
examine prior studies of integrity with respect to their implications for understanding organizational
influences on scientific integrity. We will begin by considering the findings obtained in one line of
research concerned with the individual and situational factors that influence integrity in
organizational settings. Subsequently, we will examine the kind of organizationally-based situational
variables that might influence scientific integrity using a multi-level perspective that considers
situational variables operating at the individual, group, and organizational levels of analysis (9).
Studies of Integrity
Psychological studies of integrity have typically employed one of two broad approaches (10). The
first approach holds that integrity, or the lack thereof, is primarily a function of certain characteristics
of the situation in which people find themselves. Thus, studies along these lines examine the
Corresponding author: Michael D. Mumford, Department of Psychology, University of Oklahoma, 455 W. Lindsey, #705
Dale Hall Tower, Norman, OK 73019, 405-325-5583 (voice), 405-325-4737 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
opportunities provided for dishonest behavior
(11), the reinforcements and punishments
associated with unethical acts (12), perceptions
of procedural justice (13), and stress and
authority norms (14). The second approach
holds that a lack of integrity is primarily a
function of certain characteristics of the
individual. Scholars applying this second
approach have sought to develop global measures
of integrity (15, 16), and identify certain unique
characteristics of people that are associated with
a lack of integrity (17, 18).
to the emergence of object beliefs, or the view
that others can be used as tools for personal gain
(14, 22). In harming others, unless such effects
are inhibited by self-regulation, people are likely
to acquire negative images of others and their
relationships with others. Thus, object beliefs,
along with fear, may lead to the emergence of
negative life themes. Negative life themes, along
with object beliefs, power motives, selfregulation and outcome uncertainty reflect beliefs
and motives held to exert direct effects on
people’s willingness to engage in destructive
unethical acts. Figure 1 provides a summary of
the key structural relationships specified in this
model.
In an initial test of the plausibility of this
model, O’Connor, Mumford, Clifton, Gessner,
and Connelly obtained biographies for 82 notable
historic leaders (21). They content-coded the
“rise to power” chapters included in each
biography for leaders’ expression of behaviors
indicative of the seven characteristics included in
this model (e.g., object beliefs, narcissism, etc.),
and obtained indices of the harm done to society
by leaders’ policies. In a subsequent causal
modeling effort, not only was support obtained
for the ability of these variables to predict harm
done by leaders’ policies, it was found that the a
priori structural model presented in Figure 1
provided adequate fit to the observed data. The
resulting model is shown in Figure 2.
In the second set of investigations, Mumford,
Connelly, Helton, Mowry, and Osburn sought to
determine whether the variables included in this
model could account for scores on standard
measures of integrity (34). Here 292 subjects
were asked to complete two overt measures of
integrity, the Reid Report (35) and the London
House PSI or Personnel Selection Inventory (36).
Individual Variables
In one series of studies along these lines,
Mumford and his colleagues (19-21) sought to
develop a general model of the individual
characteristics likely to promote destructive or
unethical acts. To identify the characteristics of
individuals related to the propensity for unethical
acts, Mumford and his colleagues reviewed
relevant studies in the clinical (22-24),
management ethics (12, 18, 25), socialpersonality (26-28), and criminology (29-31)
disciplines. This review resulted in the
identification of seven individual characteristics
that might plausibly be related to socially
destructive unethical behavior: 1) narcissism,
2) fear, 3) outcome uncertainty, 4) power
motives, 5) object beliefs, 6) negative life
themes, and 7) lack of self-regulation.
These differential characteristics were held to
operate as a dynamic syndrome in shaping
unethical acts. It was held that narcissism, or
extreme self-absorption and overevaluation of the
self leads to a motivated defense of a weak selfsystem (22, 32). This perception of threat, in
turn, induces outcome uncertainty and activates
power motives as a defensive strategy. Fear, or
anxiety, is also held to lead to
+
perceptions of threat, thereby
leading to outcome uncertainty
Self -Regulation
(33). When people are uncertain
about their capacity to attain
desired outcomes, self-protective
+
Object Beliefs
Negative Life Themes
tendencies will activate power
+
Lack of
motives, although the activation of
+
+
Integrity
power motives may be somewhat
Fear
Power Motives
+
inhibited by the tendency of
+
fearful individuals to withdraw.
+
+
+
Once activated, power motives
induce a tendency to harm or
Narcissism
+
Outcome Uncertainty
exploit others which, with the
Figure 1. General structural model for individual influences on integrity.
resulting desensitization, may lead
74
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
to complete the background data
scales measuring the beliefs and
Self-Regulation
motives relevant to integrity (e.g.,
.02
object beliefs, power motives,
- .10
etc.). Additionally, manipulations
.97
Negative Life Themes
Object Beliefs
were made in the conditions of
.06
Leader
task performance, specifically
- .04
.97
Destructiveness
authority norms, psychological
.02
Fear
Power Motives
.24
distance, and feelings of self.55
efficacy. It was found that MBA
.81
.26
.83
.13
students who expressed individual
Narcissism
.40
Outcome Uncertainty
characteristics held to influence
the occurrence of unethical acts
Figures 2. Test of structural model for individual influences with respect to
would take unethical actions when
leader destructiveness.
feelings of self-efficacy were low.
However, they would not
Both these measures examine theft, dishonesty,
necessarily make unethical decisions unless they
and punitive attitudes as direct markers of
integrity. In addition, 400 subjects were asked to had reason to believe that the actions taken
would be supported by people in authority. Thus,
complete two commonly used personality based
measures of integrity (37) – the Socialization and it appears that situational variables might
influence ethical decisions potentially interacting
Delinquency scales of the California
Psychological Inventory (CPI). Here background with individual predispositions in conditioning
the occurrence of unethical behavior or,
data scales were developed to measure each of
alternatively, by creating unique effects on
the characteristics included in this model using
the procedures suggested by Mumford, Costanza, unethical behavior.
Connelly, and Johnson (38). Again, it was found
Situational Variables
that the structure of the a priori model was
In fact, beginning with the work of Hartshorne
confirmed. However, here it was found that
and May (11), many scholars have argued that
although scores of these differential variables
situational variables might exert strong effects on
yielded effective prediction of integrity test
unethical behavior. In an initial investigation
scores (r = .32), the obtained prediction was not
intended to identify the kind of situational
of overwhelming power. Figure 3 illustrates the
variables that might influence the occurrence of
nature of the results obtained in this study, while
Table 1 describes the items used to measure these unethical acts, Gessner, O’Connor, Mumford,
Clifton, and Smith developed a set of life history
variables.
A potential explanation for the limited, albeit items intended to capture exposure to situations
significant, impact of these variables on integrity likely to influence development, or expression of,
test scores may be found in a study conducted by the various individual characteristics held to
Mumford, Gessner, Connelly,
.07
O’Connor, and Clifton (20). In
this study, 152 Masters of
Self-Regulation
Business Administration (MBA)
students were asked to work on an
- .19
.02
in-basket exercise which presented
.37
Object Beliefs
Negative Life Themes
32 decisions that might be made
.05
by regional sales managers. On
.27
.11
Dishonesty
- .15
half of the items included in this
- .24
.10
Fear
Power Motives
in-basket exercise, the MBA
.46
students were presented with
.51
.52
ethical decisions where the actions
.43
selected might result in harm to
Narcissism
.30
Outcome Uncertainty
others or harm to the organization.
Figures 3. Test of structural model for individual influences with respect to
Prior to starting work on this
integrity.
task, the MBA students were asked
.35
75
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Individual Scales
Object Beliefs
Example Items
Surprised by how much people invest in friendships; did not do favors for people who
could not return them; told white lies to get own way; viewed dealing with people as a
game; has not gotten emotionally involved when dealing with people.
Power Motives
Frustrated when could not convince friends to adopt one’s view; was important to be
on the winning side; was willing to make a scene to get compliance from others;
enjoyed making others do things; liked to have the last word.
Negative Life Themes
Enjoyed parties where people were really out of control; was not upset by media
violence; spending time with family was not important; has not reflected upon one’s
purpose in life as much as others.
Outcome Uncertainty
Often planned for things that never happened; wished things would slow down or
remain the same; worried about the future; annoyed by people who claimed something
was a sure thing; wished there were more guarantees in life.
Fear
Friends thought they worried too much; often agonized over decisions; often woke up
at night for no apparent reason; was bothered by things that could go wrong when
things were going well; had difficulty making decisions about the f uture.
Narcissism
Tried to make self look good; was important to receive praise from others; spend a lot
of time worrying about appearance; did not talk about things not of interest to them;
did not spend time with others whose opinions were different.
Lack Of Self-Regulation
Not hard on one’s self; rarely said the right thing at the right time; not important to
identify own limitations; took long to fit in with an unfamiliar crowd; did not express
opinions according to the situation at hand.
Table 1: Examples of Items Included in the Individual Scales
influence unethical behavior (e.g., object beliefs,
outcome uncertainty, etc.) (39). A subsequent
factoring of these items after they had been
administered to 285 undergraduates, lead to the
identification of seven situational factors:
1) alienation, 2) non-supportive family, 3) negative role models, 4) life stressors, 5) competitive
pressure, 6) exposure to negative peer groups,
and 7) financial need. Table 2 illustrates the
nature of the items used to measure these
variables.
To examine the impact of these variables on
integrity, Mumford, Connelly, Helton, Mowry,
and Osburn, administered the life history items
measuring exposure to these situational factors to
the 292 subjects asked to complete the two overt
integrity tests, the Reid Report and the PSI, and
the 400 subjects asked to complete the two
personality-based tests, the CPI socialization and
delinquency scales (34). In this study, scores on
the overt and personality based measures of
integrity were both correlated with, and regressed
on, the seven situational scales.
The first major finding to emerge from these
analyses was that the situational scales were
correlated with scores on the measures of
individual characteristics held to influence
unethical behavior (e.g., negative life themes,
object beliefs, etc.) yielding bivariate correlations
in the .40s. The second major finding indicated,
however, that the situational variables were
strongly related to integrity test scores producing
relationships in the mid-.20s to low-.50s. Of
these variables, exposure to negative peer groups,
alienation, and financial need appeared to
produce the strongest relationships across the
four measures of integrity. The third major
finding to emerge in these analyses indicated that
the situational variables yielded better prediction
of scores on the four integrity tests than the
individual variables while yielding significant
gains in prediction when added to the individual
variables. The results obtained in this third
analysis are summarized in Figure 4 which
indicates that the situational variables accounted
for far more variance in integrity test scores than
the individual variables.
Although these findings underscore the
fundamental importance of understanding
situational influences in attempts to understand
and control unethical acts. These findings leave
two crucial questions unanswered. First, they do
76
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
Situational Scales
Alienation
Example Items
Had worked in a setting where they saw discrimination; had superiors who were
condescending; worked with people and withheld information; had belonged to
organizations in legal trouble; lost something because others took advantage of status or
position; often worked in situations where they could not keep up with demand.
Non-Supportive Family
Parents were not consistent in praise of punishment; parents did not explain why they
punished; parents and teachers did not praise work; did not have input into important
family decisions; parents and siblings did not help with schoolwork.
Negative Role Models
Parents broke promises; parents openly criticized others; often witnessed violent
arguments among adults in household; parents gave harsh punishments; parents lost
temper for no apparent reason; family had different standards than other families.
Life Stressors
Unable to go to school due to health; had to cope with large unexpected expenses;
teachers made unrealistic work demands; had serious illness; schoolwork effected by
problems of family members; was in situations where they could not keep up with work.
Competitive Pressure
Often experienced competition among coworkers; concerned about finding a good job
after graduation; frequently sought recognition for work; had to be competitive to get
ahead at work or school; selected people for membership in clubs; was involved in team
projects.
Negative Peer Group
Friends had a cynical attitude towards society; high school and college friends had
trouble with law; friends and family were heavy users of drug and alcohol; observed
people breaking rules while growing up; saw people taken advantage of; witnessed
verbal/physical violence.
Financial Need
Many families in neighborhood they grew up in received some type of public assistance;
lost mother or father; regular schedule was not emphasized in family; members of family
had been in trouble with law; people could take things away from them because of family
position.
Table 2: Examples of Items Included in the Situational Scales
deadlines, the need to acquire resources, and
uncertainty about project outcomes (40). When
these occupational demands are combined with
the intense focus characteristic of those engaged
in scientific work (41), it seems plausible to
argue that stress represents an endemic feature of
life in the sciences. Although, up to a point,
stress may contribute to productivity, high levels
of stress may not only prove debilitating, but,
more centrally, may contribute to incidents of
unethical conduct through two distinct
mechanisms (42). First, high levels of stress may
lead people to take more risky actions than they
might under other conditions due to the negative
effects of stress on self-regulation (27). Second,
stress reduces the cognitive resources available
for reasoning and analytical problem solving
(43). This loss in cognitive capacity is
noteworthy because effective moral reasoning
inhibits the occurrence of unethical acts (18, 44,
45). These observations, in turn, lead to our first
not tell us exactly how unethical acts are
influenced by situational variables. For example,
situational variables might constrain unethical
behavior, interact with individual variables or,
alternatively, compel unethical behavior in their
own right. Second, these findings do not tell us
about the specific kinds of situational variables
that act to influence unethical behavior in the
kind of organizational settings in which scientists
are likely to work. Accordingly, in the following
sections, we will examine the specific kinds of
situational variables operating at the individual,
group, and organizational levels that might
influence scientific integrity.
Individual Level
Of the situational variables found to be related to
integrity, stress seems to be the variable most
likely to be linked to integrity in research work.
Scientific work is known to be demanding and
stressful resulting from multiple commitments,
77
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Personality Based Tests
CPI
CPI
PSI
Socialization Delinquency
Honesty
INDIVIDUAL SCALES
Multiple Correlations
Cross Validated Multiple Correlation
SITUATIONAL SCALES
Multiple Correlation
Cross-Validated Multiple Correlation
SITUATIONAL SCALES ADDED TO
INDIVIDUAL SCALES
Multiple Correlation
Cross-Validated Multiple Correlation
Change in R Square
Overt Tests
Reid
PSI
Honesty
Theft
Reid
Theft
.42
.36
.38
.31
.36
.29
.27
.20
.25
.07
.30
.17
.62
.49
.58
.51
.57
.40
.43
.38
.35
.26
.28
.12
.67
.62
.26**
.61
.50
.23**
.61
.58
.24**
.47
.38
.17**
.41
.27
.11**
.40
.17
.07*
Figure 4: Comparison of Individual and Situational Variables with Respect to the Prediction of Integrity Test Scores.
*P < .05 ** P < .01
two propositions.
• Proposition One: Incidents of unethical behavior
will be more frequent when individuals experience stress and overload.
• Proposition Two: Attempts by organizations to
reduce stress by minimizing time pressure,
managing overload, clarifying goals, and
providing requisite resources will reduce incidents of unethical behavior.
Actions taken to reduce work demands, of
course, are not the only steps that might be taken
to reduce stress and unethical behavior in
organizational settings. Both stress and
uncertainty about outcomes are influenced by
people’s feelings of competence and their ability
to exert positive, effective control over their
work environment. In keeping with this
observation, Weeks, Moore, McKenney, and
Longnecker administered vignettes calling for
ethical decisions to managers with greater and
lesser experience (46). They found that
experienced managers were more likely than
their less experienced counterparts to make
ethical decisions. Other studies by Arlow and
Uhlrich (47), Chonko and Hunt (48), Kidwell,
Stevens, and Bethke (49), and Teal and Carroll
(50) also indicate that more experienced
successful workers, workers with greater
expertise, are less likely to engage in unethical
activities or make unethical decisions. As noted
above, one potential explanation for these
findings is the ability of experienced, competent
workers to handle stress and uncertainty.
Experienced, competent workers, however, may
also feel less need to take shortcuts. Regardless
of the explanation used to account for these
effects, however, it is clear that organizations
may take a number of steps to build competence
and expertise through educational and mentoring
programs, careful selection of employees, and
providing people with time to pursue continuing
education projects (2).
Competence and expertise, of course, also
allow people to induce effective control over
their work environment. Given the impact of
stress, outcome uncertainty, and fear on unethical
acts, one would expect that control beliefs would
be related to unethical behavior in organizational
settings. In fact, studies by Hegarty and Sims
(12), Trevino and Youngblood (18), and Reiss
and Mitra (51) all indicate that people who have
a strong internal locus of control are less likely to
engage in unethical acts than people who believe
their actions are controlled by external forces.
What is important to recognize here, however, is
that organizations can build feelings of control by
assigning people to tasks commensurate with
their capabilities, allowing input to critical
decisions, and buffering people from
uncontrollable events. Taken as a whole, these
observations imply the following three
propositions.
• Proposition Three: Less skilled or less experienced scientists will be more likely to engage in
unethical acts and will be more sensitive to
organizational pressures that promote unethical
acts.
78
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
the potential contribution of its’ worth to society
as a whole (55, 56). As Bowie points out, this
intrinsic motivation buffers individuals from
situational pressures likely to promote unethical
acts (57) . He notes, furthermore, that a variety of
organizational policies might influence alienation
and intrinsic motivation including explicit
recognition of social contributions as well as
contributions to the “bottom line”, allowing
individuals to pursue personally interesting work,
and maximizing autonomy in decision-making.
These observations suggest the following
proposition.
• Proposition Eight: Attempts by the organization
• Proposition Four: Organizational actions
intended to develop expertise and maximize
feelings of competence will inhibit unethical acts.
• Proposition Five: Organizational actions intended to maximize people’s control of their
environment will inhibit unethical acts.
As important as competence and control may
be to the management of stress and the
minimization of unethical behavior, some
consideration should be given to family and
social relationships. Family and social
relationships, specifically supportive
relationships, help people cope with stress while
the implied commitment to others embedded in
these relationships promotes a prosocial outlook.
Accordingly, Mumford, Connelly, Helton,
Mowry, and Osburn (34) found that exposure to a
non-supportive family environment was related
to a lack of integrity. Unfortunately scientists, in
part due to their introversion (52) and, in part due
to their work commitments (53), appear to have
some difficulty in establishing viable family and
social relationships. By the same token,
however, scientists do appear to establish viable,
long-term collaborative relationships and create
social connections through their network of
enterprise (5, 54). These observations, in turn,
suggest that incidents of unethical behavior will
occur less frequently among scientists who have
a rich extensive network of supportive
professional colleagues. Moreover, by colocating scholars with similar interests,
encouraging collaborative work, recognizing the
value of multiple-authored publications, and
providing time for collegial interactions,
organizations can reduce incidents of scientific
misconduct. Thus:
• Proposition Six: Individuals lacking collabora-
to recognize and reward social contributions and
allow individuals to pursue their unique interests
will reduce incidents of scientific misconduct.
Eisenberger and Cammeron, however,
remind us that creative work, including scientific
work, is not simply a matter of intrinsic
motivation (58). People’s work as scientists is
also motivated by extrinsic factors such as pay,
recognition, and status. At first glance, it might
seem plausible to argue that extrinsic rewards
lead to unethical behavior. However, the
relationship between the pursuit of extrinsic
rewards and unethical behavior appears
somewhat more complex with the pursuit of
extrinsic rewards contributing to unethical acts
only when people expect that the unethical
behavior will be rewarded, the unethical act will
not be detected, and the act, if detected, will not
be sanctioned by the organization (12, 18, 59).
One implication of this expectancy model is that
high performers will sometimes engage in
unethical acts because they believe they are less
likely to be sanctioned by the organization (60,
61)–potentially resulting in a culture that seems
to condone such acts. Another implication of this
expectancy model is that ethical behavior will
decrease when extrinsic rewards such as pay and
promotions are based on immediate short-term
production demands rather than long-term
contributions to others (62).
In considering the impact of production
demands, however, it is necessary to bear in mind
a unique characteristic of scientific work.
Scientists’ rewards are often explicitly tied to
production such as journal publications, patents,
and fielding new software (63, 64). By expressly
tying extrinsic rewards to production counts,
however, one can expect that misconduct will
increase whenever ambitious, extrinsically
motivated individuals, individuals motivated by
tive networks will be more likely to be involved
in incidents of scientific misconduct.
• Proposition Seven: Organizational actions
intended to facilitate and recognize the value of
collaborative activities will minimize incidents of
scientific misconduct.
Our foregoing observations with regard to
collaboration point to another factor likely to be
involved in incidents of scientific misconduct –
alienation. Alienation among scientists is not a
strictly social phenomenon. Alienation from the
work, and the work’s potential contributions to
society, appear particularly significant with
regard to scientific misconduct because scientific
work is often motivated by intrinsic interest in
the work for its own sake and an abiding belief in
79
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
financial needs, status concerns, and recognition,
encounter significant reverses in the production
process. Thus, organizations might minimize
misconduct by rewarding progress towards goals
as well as production output, recognizing
alternative indices of performance such as impact
and innovation, and providing a minimal degree
of security and visibility for all group members
based on their unique strengths.(65) Taken as a
whole, our preceding observations about
extrinsic motivation suggest the following four
propositions.
• Proposition Nine: Organizational reward systems
that stress long-term innovation and impact will
tend to minimize incidents of unethical behavior.
• Proposition Ten: Organizational reward systems
that recognize progress as well as output will
tend to minimize incidents of unethical behavior.
• Proposition Eleven: Scientific misconduct will
occur more frequently when extrinsic rewards are
based on production and people are treated
harshly for production setbacks.
• Proposition Twelve: Scientific misconduct will
occur less frequently in organizations where all
incidents of misconduct are treated similarly,
regardless of the past performance of the people
involved.
Groups
The Mumford, Connelly, Helton, Mowry, and
Osburn study not only points to the influence of
individual level situational influences on
integrity, such as stress, relational support,
alienation, and financial need, it also underscores
the importance of certain group level influences
(34). In this study, three variables operating at
the group level, role models, exposure to
negative peer groups, and competitive pressure,
were found to influence integrity. Again, all
three of these situation variables appear to
represent important influences on integrity in
organizational settings.
In organizations, role modeling is commonly
subsumed under this broader area of leadership
(66), and there is, in fact, reason to believe that
the behavior of people assigned to formal
organizational leadership roles will influence the
manifest integrity of their “followers”. In one
study along these lines, Schminke and Wells had
81 business students participate in a four-month
long strategic planning simulation (67). During
the course of this simulation, measures of ethical
decision-making were obtained along with
measures of group process variables and
leadership styles, specifically consideration and
initiating structure. They found that the leaders’
emphasis on initiating structure contributed to
ethical decision-making, presumably because the
initiation of structure led group members to focus
on task accomplishment rather than personal
concerns. In another study along these lines,
Zabid and Alasgoff found that the behavior of
people’s immediate superior exerted stronger
effects on the occurrence of unethical acts than
other putative organizational influences such as
climate and codes of conduct (68).
Leaders appear to influence ethical behavior
through a variety of different mechanisms, some
of which may inhibit unethical acts and some of
which may promote such acts. Sims, in a study
of leadership in financial services firms,
identified four ways leadership behavior
contributes to or promotes integrity (69). He
argues that leaders promote ethical behavior by
a) focusing the attention of people on ethical
issues, b) responding to crises based on ethical,
productive concerns rather than self-protection,
c) allocating rewards based on long-term
contributions rather aggressive self-promotion,
and d) applying sanctions for incidents of
unethical behavior. Along similar lines, Minkes,
Small, and Chatterjee have argued that leaders’
articulation and communication of personal,
ethical, and moral values will promote integrity
on the part of group members (70). Contrawise,
it appears that leaders who articulate poor values
or exhibit self-serving, narcissistic behavior
implicitly encourage unethical behavior on the
part of subordinates (71, 72). Vredenburgh and
Brender point out, moreover, that leaders who
consistently abuse power through arbitrary
actions, a focus on personal control, and
inequitable decisions, induce stress, fear, and
outcome uncertainty while activating the power
motive linked to unethical acts (73).
Although it seems clear that leaders have an
impact on ethical behavior in general, the
question remains as to whether leaders have a
similar impact on the ethical behavior of
scientists. One might argue that, due to their
greater autonomy and specialized professional
expertise, scientists are less susceptible to leader
influence (66, 74). Although this argument
seems plausible, the available evidence indicates
that leaders exert notable effects on people’s
behavior in research settings (75). A case in
point may be found in Hounshell’s analysis of
research on synthetic fabrics in Dupont’s Pioneer
80
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
acts. Scientists have been found to be highly
competitive evidencing not just competitive
intensity but some degree of hostility and
arrogance (79)–all dispositional factors likely to
make scientists particularly susceptible to the
negative effects of competitive pressure.
Competitive pressure, however, may not always
be destructive provided it is managed effectively
by the organization (80). More specifically,
when competition is accompanied by respect for
competitors, people feel that they have sufficient
technical competence to compete effectively, and
competition is viewed as a depersonalized,
professional challenge, then competition may
contribute to performance and ethical behavior
(81, 82). These observations, in turn, suggest the
following three propositions.
• Proposition Sixteen: Unethical acts are more
research laboratories where the vision defined by
founders in the 1920s continued to shape the
laboratories’ research programs well into the
1990s (76). Nonetheless, the autonomy and
expertise of scientists suggest that leader
influences on ethical issues will be less evident in
day-to-day direction and more evident in the
leaders’
a) definition of a coherent constructive
research vision, b) focus on production as
opposed to status relationships, and c) articulation of ethical values in interactions with staff.
When these observations are considered with
respect to the findings sketched out above, they
suggest the following three propositions:
• Proposition Thirteen: Scientific misconduct will
be less common in groups where leaders have the
expertise needed to define a coherent vision for
the work.
• Proposition Fourteen: Scientific misconduct will
be less common in groups where the leader
actively articulates ethical values, potential social
contributions of the work, and enhancement of
the work rather than career status.
• Proposition Fifteen: Scientific misconduct will
be less common in groups where the leader
focuses on effective direction of production
activities rather than personal professional
recognition, maintenance of control, or social
acceptance.
likely to be observed when ambitious, highly
competitive people are placed in competitive
settings where they lack requisite skills.
• Proposition Seventeen: Organizations that take
actions to reduce personalized competitive
pressure by evaluating performance on an
absolute rather than relative basis or by encouraging collaborative work among potential competitors are less likely to experience incidents of
unethical behavior.
• Proposition Eighteen: Unethical behavior is less
likely to occur when leaders, or organizational
practices, encourage people to analyze and
identify the merits in competitors’ work.
Leadership, of course, is not the only group
level variable that might influence integrity in
organizational settings. For example, Mumford,
Connelly, Helton, Mowry, and Osburn found that
competitive pressure was related to a lack of
integrity (34). The effects of competition on
ethical behavior, however, appear to be quite
complex in organizational settings. One way
competition appears to influence ethical behavior
may be found in the tendency of people to
discount the relevance of moral considerations to
decision-making in competitive situations (77).
Another way competition influences ethical
behavior is that negative perceptions of
competitors’ intentions provide a justification of
unethical acts (78). Still another way
competition influences ethical behavior is by
inducing feelings of stress and uncertainty (39).
These varied mechanisms by which
competition influences ethical behavior are all
clearly applicable to scientists. In the case of
scientists, however, it is quite possible that these
negative aspects of competition represent
particularly important influences on unethical
Personalized competition within-groups, of
course, may result in conflict and a lack of
cohesiveness. In this regard, the Schminke and
Wells study cited earlier is noteworthy. In
addition to examining leadership styles and their
influence on ethical decision-making, they also
examined the effects of group cohesiveness (67).
Here it was found that cohesiveness influenced
ethical decision-making both directly with more
cohesive groups making more ethical decisions
and indirectly with cohesive groups evidencing
higher performance which, in turn, led to more
ethical decision-making. These findings suggest
that actions taken to induce cohesiveness through
development and articulation of a shared,
common vision, use of group as well as
individual rewards, integration of members work
activities, and encouragement of within-group
collaborative efforts will all contribute to ethical
behavior. Thus, the following three propositions
seem indicated.
81
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
• Proposition Nineteen: Unethical acts are more
likely to occur in non-cohesive conflict-laden
groups.
• Proposition Twenty: Cohesiveness within a
group will reduce scientific misconduct both by
enhancing performance and minimizing the
negative effects of within-group competition.
• Proposition Twenty-One: Organizational actions
that lead to higher cohesiveness, such as development of a shared vision on the allocation of
group, as well as individual, rewards, will reduce
incidents of scientific misconduct.
Although it appears that cohesiveness may
contribute to integrity, a cautionary note seems in
order. Many prior studies of groups, including
destructive behavior on the part of groups,
indicate that conformity pressures can induce
destructive, unethical behavior when the primary
concern is maintenance of harmonious group
relations and the goals being pursued by the
group are likely to result in destructive, unethical
behavior (24, 83). Hence:
• Proposition Twenty-Two: When high levels of
ethical behavior among group members will
contribute to integrity. As might be expected, the
bulk of the available evidence does indicate that
ethical norms within a group lead to ethical
behavior. For example, studies by Barnett (86),
Kawathatzopoulos (87), Verbke, Ouwerkerk, and
Peelen (88), and Weaver and Farrell (89) indicate
that when groups communicate expectations for
ethical behavior, and sanction violations by
group members, ethical decision-making
improves and unethical acts become less
frequent. In this regard, however, it is important
to bear in mind a point made by Fritz, Arnett, and
Conkel (90), Grimalda (91), and Schokkaert and
Sweeney (92). More specifically, the effects of
group norms on ethical behavior will vary with
people’s commitment to the group. Accordingly,
the following three propositions seem indicated.
• Proposition Twenty-Four: Ethical behavior will
be more common in groups that have, and
actively apply, positive normative standards in
group decision-making and the application of
sanctions.
• Proposition Twenty-Five: The effects of ethical
norms on integrity depend on building feelings of
commitment to the group, the organization, or the
profession.
• Proposition Twenty-Six: the creation and
articulation of normative ethical standards by
leaders on professional organizations will prove
less effective when groups are experiencing rapid
change and commitment is low.
cohesiveness prohibit questioning of group
actions, cohesiveness may be related to unethical
acts.
As implied by our foregoing proposition,
exposure to the behaviors of, and expectations
imposed by, other group members may influence
ethical behavior in organizational settings (34).
Exposure to peer groups is commonly held to
influence integrity through the models for
appropriate behavior provided by other group
members and the normative expectations
imposed on people by other members of the
group (39, 84) . Accordingly, Murphy has argued
that anomie, or normlessness, will engender
unethical behavior because group members lack
models for appropriate behavior and sanctions
are not imposed for unethical acts (10). In
keeping with argument, Leede, Nijhof, &
Fisscher, note that when groups are experiencing
conditions of rapid change the resulting
breakdown in extant normative structures may
lead to an increase in the frequency of unethical
acts (85). Thus,
• Proposition Twenty-Three: When groups are
experiencing rapid changes in personnel, technology, or productions processes, incidents of
unethical behavior will increase.
The notion that normlessness will contribute
to the occurrence of unethical acts also implies
that the presence of normative expectations for
Organizations
The Mumford, Connelly, Helton, Mowry, and
Osburn study focused primarily on situational
factors operating at the individual or group level
(34). As a result, this study does not directly
address the various organizational level variables
that might be related to integrity. Nonetheless,
the nature of the individual and group based
situational influences on integrity do suggest that
certain organizational level variables will also
influence integrity. One set of organizational
level influences suggested by our foregoing
observations is the organization’s operating
environment – specifically three features of the
organization’s operating environment turbulence,
munificence, and interdependence.
Environmental turbulence refers to rapid
changes in technology, business processes,
product markets, and competitors (93). Of
course, turbulence will lead to normlessness as
well as uncertainty about the requirements for
82
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
partners, or government agencies. As might be
expected, high interdependence appears to
promote ethical behavior (99, 100, 101).
Although it is unclear exactly what mechanisms
shape the influence of interdependence on ethical
behavior the following proposition does seem
indicated:
effective performance, both conditions that can
be expected to promote unethical acts.
Accordingly, Morris, Marks, Allen, and Perry
found that ethical values were less evident among
people working for organizations operating in a
turbulent environment (94). Along similar lines,
Rossouw has argued that the turbulence induced
by social disruption can lead to unethical acts on
the part of organizations (95). Among scientists,
however, it seems likely that turbulence will
exert larger effects when its impact is evident in
their immediate technical environment or in
employment practices. These observations, in
turn, lead to the following two propositions.
• Proposition Thirty: Unethical behavior occurs
less frequently in organizations where performance depends on the support, or goodwill, of
other entities.
The organization’s operating environment is,
of course, one influence on the structure of the
organization. Structure, or the manifest division
of labor in an organization, has not commonly
been studied as an influence on integrity.
However, the available evidence indicates that
unethical acts are less likely to occur in small
organizations (102, 103) and in organizations
where roles and responsibilities are clearly
defined (85, 104). One explanation for this
pattern of findings may be found in diffusion of
responsibility and its derivative effects or
alienation. In keeping with this alienation and
diffusion of responsibility notion, Dooley and
Fryxell found that diversification was related to
corporate pollution levels (105). These
observations imply the following proposition:
• Proposition Twenty-Seven: As turbulence
increases in the organization’s operating environment the frequency of unethical acts will increase.
• Proposition Twenty-Eight: Scientific misconduct
will increase in periods of rapid change in
technological paradigms and employment
practices.
In contrast to turbulence, munificence refers
to the availability of resources and the low
degree of competitive pressure evident in the
organizations’ operating environment. In fact,
the available evidence indicates that munificence
is related to ethical conduct in organizational
settings. For example, Verschoor (96), in a study
of Fortune 500 companies, found that ethical
conduct with regard to organizational
shareholders increased with financial
performance while Judge (97), in a study of
hospitals, found that scarcity of financial
resources was negatively related to social
contributions. In still another study along these
lines, Zarkada-Fraser found that collusion in
government project bids was related to project
desirability and competition (98). Among
scientists, where resources are critical to
conducting requisite research work, nonmunificent environments may encourage
unethical acts as a way of insuring resource
availability. Thus,
• Proposition Thirty-One: As organizational
structures become more complex, and roles and
role accountability are less clearly defined for
individuals, unethical acts will become more
frequent.
While structure refers to the organization of
the work, climate refers to people’s perceptions
of social interactional expectations with their
work environment (106). Relative to structure,
climate has received substantially more attention
as a potential influence on ethical behavior in
organizational settings. In one study along these
lines, Sims and Keon administered five business
scenarios calling for an ethical decision to 245
business students who were also asked to
complete a survey describing the company for
which they were currently working (107). It was
found that perceptions of their work environment
were related to ethical decision-making. Similar
findings have been obtained by Baumhart (59).
Although there is reason to believe that
organizational climate influences ethical
behavior, more debate surrounds the nature of the
specific climate dimensions involved. Agarwal
and Malloy identify five climate dimensions
related to ethical behavior:
• Proposition Twenty-Nine: As the munificence of
the organizations operating environment decreases, unethical behavior and incidents of
scientific misconduct will increase.
A third, and final, environmental variable
commonly linked to ethical behavior in
organizational settings is interdependence, or the
extent to which organizational success depends
on maintaining viable relationships with other
organizations including suppliers, alliance
83
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1) individual caring 2) social caring, 3) independence, 4) Machiavellianism, and 5) law and
code (108). Vidaver-Cohen proposes a different
model of ethical climate which stresses the
importance of 1) social responsibility, 2) social
support, 3) avoiding harm of others, 4) task
support, and 5) equity of reward procedures
(109). Still another model, one proposed by Key,
views climate as a function of: 1) day-to-day
reinforcement of ethical conduct, 2) punishment
of unethical conduct, and 3) management role
modeling (110). Finally, Argadona and Hartman,
Yrle, and Galle argue that trust and perceptions
of distributive and procedural justice represent
key organizational climate dimensions
influencing ethical behavior on organizations
(111,112).
While a variety of models of ethical climate
are available, it seems likely that some of these
dimensions will prove more important than
others in shaping the ethical behavior of
scientists. Given the hostility and
competitiveness characteristic of scientists (79),
it seem plausible to argue that climates stressing
trust and social support while maintaining
perceptions of procedural and distributive justice
will prove particularly important in minimizing
misconduct (7). The demands of creative work,
moreover, suggest that climates reinforcing
autonomy, openness, and minimization of
premature criticism will also prove useful in
enhancing ethical behavior (75, 113). Thus, the
following two propositions seem indicated.
• Proposition Thirty-Two: Organizational climates
that promote perceptions of trust and fairness will
minimize incidents of scientific misconduct.
• Proposition Thirty-Three: Organizational
climates that are open and not overly critical of
new ideas will minimize incidents of scientific
misconduct.
The climate literature, however, also
underscores the importance of day-to-day
reinforcement on ethical conduct. In the case of
scientists, the importance of ethical standards
implies that professional codes, as well as their
acceptance and embodiment by the organization,
will also influence incidents of scientific
misconduct. In fact, studies by Weaver and
Farrell (89) of American Marketing Association
members, and Gotterbarn (114) of software
engineers, indicate that professional codes are
viewed as important influences on ethical
behavior in the sciences and may lead to
improvements in ethical decision-making.
On the other hand, however, there is no
assurance that professional ethical codes will be
adopted by organizations in their day-to-day
practices. This point is nicely illustrated in a
study by Etheredge who examined attitudes
toward ethical behavior in business managers and
identified two dimensions: a) the importance of
ethics and social responsibility, and
b) subordination of ethics and social
responsibility to organizational effectiveness
(115). Thus, organizations in their quest for
efficiency and control, may reject professional
ethical standards that conflict with organizational
needs. When organizations reject these
professional standards, however, it can be
expected that the resulting organizationalprofessional conflict will induce some stress as
people are forced to choose between these
competing expectations. Although a number of
considerations will influence how this conflict is
resolved, it appears that investment in the
organization, as opposed to the profession, is of
critical importance (116). Accordingly, the
following three propositions seem indicated.
• Proposition Thirty-Four: Incidents of scientific
misconduct will be less common among individuals who are more invested in the profession
rather than the organization they are working.
• Proposition Thirty-Five: Incidents of scientific
misconduct will be less common in organizations
that rely on their professional technical reputation
for market advantage and view organizational
needs as consistent with professional ethical
codes.
• Proposition Thirty-Six: Professional ethical
codes will prove most effective in reducing
scientific misconduct when codes are actively
supported by the organization.
Conclusions and Directions
Figure 5 summarizes the various propositions we
have proposed with respect to the situational
variables influencing ethical behavior at the
individual, group, and organizational levels. In
reviewing these propositions, however, an
important caveat seems in order. More
specifically, although all of the propositions were
formulated based on a review of the
organizational literature as it relates to the
situational variables influencing integrity. Few, if
any, studies have directly examined the influence
of organizational, situational variables on
research integrity. Thus, these propositions
should not be viewed as well established
84
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
Individual Level
Group Level
Organizational Level
1) Incidents of unethical behavior will
be more frequent when individuals
experience stress and overload
2) Attempts by organizations to reduce
stress by minimizing time pressure,
managing overload, clarifying
goals, and providing requisite
resources will reduce incidents of
unethical behavior
3) Less skilled or less experienced
scientists will be more likely to
engage in unethical acts and will be
more sensitive to organizational
pressures that promote unethical
acts
4) Organizational actions intended to
develop expertise and maximize
feelings of competence will inhibit
unethical acts
5) Organizational actions intended to
maximize people’s control of their
environment will inhibit unethical
acts
6) Individuals lacking collaborative
networks will be more likely to be
involved in incidents of scientific
misconduct
7) Organizational actions intended to
facilitate and recognize the value of
collaborative activities will
minimize incidents of scientific
misconduct
8) Attempts by organizations to
recognize and reward social
contributions and allow individuals
to pursue their unique interests will
reduce incidents of scientific
misconduct
9) Organizational reward systems that
stress long-term innovation and
impact will tend to minimize
incidents of unethical behavior
10) Organizational rewards that
recognize progress as well as output
will tend to minimize incidents of
unethical behavior
11) Scientific misconduct will occur
more frequently when extrinsic
rewards are based on production
and people are treated harshly for
setbacks
12) Scientific misconduct will occur
less frequently in organizations
where all incidents of misconduct
are treated similarly regardless of
past performance
13) Scientific misconduct will be less common in
groups where leaders have the expertise needed
to define a coherent vision for the work
14) Scientific misconduct will be less common in
groups where the leader actively articulates
ethical values, potential social contributions of
the work and enhancement of the work rather
than career status
15) Scientific misconduct will be less common in
groups where the leader focuses on effective
direction of production activities rather than
personal professional recognition, maintenance
of control, or social acceptance
16) Unethical acts are more likely to be observed
when ambitious, highly competitive people are
placed in competitive settings where they lack
requisite skills
17) Organizations that take actions to reduce
personalized competitive pressure by evaluating
performance on an absolute rather than relative
basis or by encouraging collaborative work
among potential competitors are less likely to
experience incidents of unethical behavior
18) Unethical behavior is less likely to occur when
leaders, or organizational practices, encourage
people to analyze and identify the merits in
competitors’ work
19) Unethical acts are more likely to occur in noncohesive, conflict-laden groups
20) Cohesiveness within a group will reduce
scientific misconduct both by enhancing
performance and minimizing the negative
effects of within group competition
21) Organizational actions that lead to higher
cohesiveness such as development of a shared
vision or the allocation of group as well as
individual rewards will reduce incidents of
scientific misconduct
22) When high levels of cohesiveness prohibit
questioning of group actions, cohesiveness may
be related to unethical acts
23) When groups are experiencing rapid changes in
personnel, technology, or production progress,
incidents of unethical behavior will increase
24) Ethical behavior will be more common in
groups that have, and actively apply, positive
normative standards in group decision-making
and the application of standards
25) The effects of ethical norms on integrity may
depend on building feelings of commitment to
the group, organization or profession
26) The creation and articulation of normative
ethical standards by leaders in professional
organizations will prove less effective when
groups are experiencing rapid change and
commitment is low
27) As turbulence increases in the
organization’s operating
environment, the frequency of
unethical acts will increase
28) Scientific misconduct will
increase in periods of rapid
change in technological
paradigms and employment
practices
29) As the munificence of the
organization’s operating
environment decreases, unethical
behavior and incidents of
scientific misconduct will
increase
30) Unethical behavior will occur
less frequently in organizations
where performance depends on
the support, or goodwill, of other
entities
31) As organizational structures
become more complex, and roles
and role accountability are less
clearly defined for individuals’
unethical acts will become more
frequent
32) Organizational climates that
promote perceptions of trust and
fairness will minimize incidents
of scientific misconduct
33) Organizational climates that are
open and not overly critical of
new ideas will minimize
incidents of scientific
misconduct
34) Incidents of scientific
misconduct will be less common
among individuals who are more
invested in the profession rather
than the organization for which
they are working
35) Incidents of scientific
misconduct will be less common
in organizations that rely on their
professional or technical
reputation for market advantage
and view organizational needs as
consistent with professional
ethical codes
36) Professional ethical codes will
prove most effective in reducing
scientific misconduct when
codes are actively supported by
the organization
Figure 5. Summary of Propositions at Individual, Group, and Organizational Levels
conclusions but, instead, as a set of hypotheses
that might be used to guide further research.
The need for further research along these
lines becomes even more salient when one takes
two other considerations into account. First,
although the propositions presented in the present
effort all seem plausible, evidence is not
available examining the relative importance of
these various situational variables on scientific
misconduct and research integrity. For example,
given the known dispositional characteristics of
scientists (79), it seems attractive to argue that
85
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
competition, conflict, and a lack of cohesiveness
will have a greater impact on misconduct than
the direction provided by a leader. Unfortunately,
however, evidence allowing us to evaluate the
relative importance of various situational
influences within and across three levels of
analysis is, at this juncture, simply not available.
Second, in formulating these propositions we
have examined organizations as a general
phenomenon drawing heavily from past research
in the “for profit” business arena (18, 107). What
must be recognized here, however, is that
scientists’ work occurs in a variety of settings
aside from the business arena including
universities, government agencies, and non-profit
research institutes. As a result, the unique
characteristics of these non-business settings may
influence the relative importance of the various
situational variables identified in the present
effort. A case in point can be found in our
observations about organizational conflicts with
professional codes of ethics since such conflicts
maybe less pronounced outside the business
setting. Thus, there is a need to assess the
generality of these propositions across work
settings.
Even bearing these caveats in mind,
however, we believe that the present study does
lead to some noteworthy conclusions about
research integrity. To begin, we tend to attribute
incidents of misconduct to characteristics of the
individual. Although the importance of the
scientist’s character is not to be underestimated,
the results obtained in the present effort suggest
that situational variables have a large, perhaps a
larger, impact on integrity than individual
variables. Although this argument is by no
means unique (11), it does suggest that future
studies of research integrity should give as much
attention to situational and individual influences.
The present effort, moreover, has served to
identify an initial set of situational variables that
should be examined in studies of research
integrity. The Mumford, Connelly, Helton,
Mowry, and Osburn study underscores the
importance of stress, alienation, support, need,
role models, peer groups, and competitive
pressure (34). In this paper we have provided
some evidence that these same situational
pressures might also be operating in
organizational settings. For example, stress
appears to be a potentially significant influence
on incidents of misconduct at the individual level
while competitive pressure appears to influence
integrity at the group level. These individual and
group level situational influences, moreover,
appear to be associated with a coherent set of
organizational level influences such as turbulence
and munificence.
In identifying the situational variables
operating at the individual, group, and
organizational levels, moreover, it becomes
possible to draw inferences about the conditions
under which incidents of misconduct are most
likely to be observed and the actions that might
be taken by organizations to reduce incidents of
misconduct. For example, support appears to be
related to misconduct with individuals lacking
collaborative networks and broader social
support being more vulnerable to misconduct.
Organizations, however, by encouraging people
to collaborate and build a strong network of
professional connections, may do much to
minimize misconduct. Similarly, while
competitive pressure apparently plays a notable
role in scientific misconduct, such simple
strategies as avoiding person-to-person
comparisons and insuring adequate resources are
available may do much to minimize the
occurrence of misconduct. Hopefully, the
present effort will serve not only as a framework
for further research examining the impact of
situational variables on scientific misconduct but
will provide a basis for formulating new policies
that will help insure the integrity of the research
process. In fact, given the changes occurring in
many scientific fields, there may well in the
future be an even more pressing need for
practical guidelines along these lines as the
rarefied world of science comes into ever closer
contact with the manifold demands and pressures
of the modern organization.
Acknowledgements
We would like to thank Shane Connelly, Ted
Gessner, Jennifer O’Connor, and Howard Timm
for their contributions to the present effort. Parts
of this effort were supported by a series of grants
from the United States Office of Naval Research,
Michael D. Mumford, Principal Investigator.
Correspondence concerning this article should be
addressed to Dr. Michael D. Mumford,
Department of Psychology, The University of
Oklahoma, Norman, OK 73019, or
[email protected] .
86
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
Bibliography
19. Holt RW, Clifton TC, O’Connor JA, Smith TC, Gessner
TC, Mumford MD. Influences of destructive
personality information on juror decision making. J
Appl Social Psychol. 1997;27:781-99.
20. Mumford MD, Gessner TE, Connelly MS, O’Connor
JA, Clifton TC. Leadership and destructive acts:
Individual and situational influences. Leadership Q.
1993;4:115-48.
21. O’Connor JA, Mumford MD, Clifton TC, Gessner TE,
Connelly MS. Charismatic leaders and destructiveness:
A historiometric study. Leadership Q. 1995;6:529-55.
22. Fromm E. The Anatomy of Human Destructiveness.
New York, NY: Holt, Rhinehart, & Winston; 1973.
23. Gough HG. A sociological theory of psychopathy. Am
J Sociology. 1998;53:359-66.
24. Staub E. The Roots of Evil: The Origins of Genocide
and Other Group Violence. Cambridge, MA:
Cambridge University Press; 1989.
25. Robin DP, Reidenbach RE, Forrest PJ. The perceived
importance of an ethical issue on the ethical decision
making of ad managers. J Business Ethics. 1996;35:1728.
26. Blass T. Understanding behavior in the Milgram
Obedience experiment. The role of personality,
situations, and interactions. J Pers Soc Psychol.
1991;60:348-413.
27. Leith KP, Baumeister RF. Why do bad moods increase
self-defeating behavior? Emotion, risk-taking, and selfregulation. J Pers Soc Psychol. 1996;71:1250-67.
28. Oyserman D, Markus HR. Possible selves and
delinquency. J Pers Soc Psychol. 1990;59:112-25.
29. DiLalla LF, Gottesman II. Heterogeneity of causes for
delinquency and criminality: Lifespan perspectives.
Dev Psychopathology. 1990;1:339-49.
30. Reid JB, Patterson GR. The development of antisocial
behavior patterns in childhood and adolescence. Eur J
Pers. 1989;3:107-19.
31. Walters GD. The Criminal Lifestyle: Patterns of Serious
Criminal Conduct. Newberry Park, CA:Sage; 1990.
32. Emmons RA. Relationship between narcissism and
sensation seeking. Psychol Rep. 1981;48:247-50.
33. Riskind JH, Kelly K, Harmen W, Moore R, Gaines HS.
The loomingness of danger: Does it discriminate all
phobia and general anxiety from depressing? Cognitive
Ther Res. 1992;16:602-22.
34. Mumford MD, Connelly MS, Helton WB, Mowry JR,
Osburn HK. On the construct validity of integrity tests:
Individual and situational factors as predictors of test
performance. Int J Selection Assessment. [In press].
35. Ash P. Predicting dishonesty with the Reid Report. J
Am Polygraph Assoc. 1985;5:139-45.
36. Harris MM, Sackett PR. A factor analysis of an
employee honesty test. J Business Psychol.
1987;2:122-35.
37. Gough HG. The California Psychological Inventory.
In: Newmark CS, ed. Major Psychological Assessment
Instruments: Volume II. Boston, MA: Allyn and Bacon;
1989:67-98.
1. Mumford MD, Baughman WA, Sager CE. Picking the
right material: Cognitive
processing skills and their
role in creative thought. In: Runco MA, ed. Critical
Creative Thought. Cresskill, NJ: Hampton; 2000.
2. Mumford MD, Gustafson SB. Creativity syndrome:
Integration, application, and innovation. Psychol Bull.
1985;103:27-43.
3. Policastro E, Gardner H. From case studies to robust
generalizations: An approach to the study of creativity.
In: Sternberg RJ, ed. Handbook of Creativity.
Cambridge, England: Cambridge University Press;
1999.
4. Weber RJ, Perkins DN. Inventive Minds: Creativity in
Technology. Cambridge, England: Cambridge
University Press; 1992.
5. Abra J. Collaboration in creative work: An initiative for
investigation. Creativity Res J. 1994;7:1-20.
6. Dunbar K. How scientists really reason: Scientific
reasoning in real world laboratories. In: Sternberg RJ,
Davidson JE, eds. The Nature of Insight. Cambridge,
MA: MIT Press; 1995:365-96.
7. Ruppel CP, Harrington SJ. The relationship of
communication, ethical work climate, and trust to
commitment and innovation. J Business Ethics.
2000;17:1563-71.
8. Press E, Washburn J. The Kept University. Atlantic.
2000;285:39-55.
9. Dansereau F, Yammarino FJ. Introduction and
overview. In: Dansereau F, Yammarino FJ, eds.
Leadership: the Multiple Level Approaches. Stanford,
CT: JAI Press; 1998:xxv-li.
10. Murphy KA. Honesty in the Work Place. Pacific
Grove, CA: Brooks/Cole; 1993.
11. Hartshore H, May MA. Studies in Deceit. New York:
MacMillian; 1928.
12. Hegarty WH, Sims H. Some determinants of unethical
decision behavior: An experiment. J Appl Psychol.
1978;58:111-21.
13. Victor BC, Trevino K, Shapiro DL. The influence of
justice evaluations and social context factors. J
Business Ethics. 1993;12:253-63.
14. Darley JM. Social organization for the production of
evil. Psychol Inquiry. 1992;3:199-218.
15. Ones DS, Viswesvaran C, Schmidt FL. Comprehensive
meta-analysis of integrity test validities: Findings and
implications for personnel selection and theories of job
performance. J Appl Psychol Monograph.
1993;78:674-703.
16. Sackett PR, Waneck JE. New developments in the use
of measures of honesty, integrity, conscientious,
dependability, trustworthiness, and reliability for
personnel selection. Personnel Psychol. 1996;49:787829.
17. Hogan J, Brinkmeyer K. Bridging the gap between
overt and personality-based integrity tests. Personnel
Psychol. 1997;50:587-600.
18. Trevino L, Youngblood SA. Bad apples in bad barrels:
A causal analysis of ethical decision making behavior. J
Appl Psychol. 1990;75:378-85.
87
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
38. Mumford MD, Costanza DP, Connelly MS, Johnson JF.
Item generation procedures and background data scales:
Implications for construct and criterion-related validity.
Personnel Psychol. 1996;49:361-98.
39. Gessner TE, O’Connor JA, Mumford MD, Clifton TC,
Smith JA. Situational variables influencing the
propensity for destructive acts: Taxonomy development
and validation. Curr Psychol. 1995;13:303-25.
40. Wild JT. The origin of soft-tissue ultrasonic echoing
and early instrumental application to clinical medicine.
In: Weber RJ, Perkins DN, eds. Inventive Minds:
Creativity in Technology. Cambridge, England:
Cambridge University Press; 1992.
41. Gruber HE, Wallace DB. The case study method and
evolving systems approach for understanding unique
creative people at work. In: Sternberg RJ, ed.
Handbook of Creativity. Cambridge, England:
Cambridge University Press; 1999:93-115.
42. Vosberg SK. The effects of positive and negative affect
on divergent thinking performances. Creativity Res J.
1998;11:165-72.
43. Fiedler FE, Garcia JE. New Approaches to Effective
Leadership: Cognitive Resources and Organizational
Performance. New York: Wiley; 1987.
44. Chang MK. Predicting unethical behavior: A
comparison of the theory of reasoned action and the
theory of planned behavior. J Business Ethics.
1998;17:1825-34.
45. Goolsby JR, Hunt SD. Cognitive moral development
and marketing. J Marketing. 1992;56:55-68.
46. Weeks WA, Moore CW, McKinney JA, Longenecker
JG. The effects of gender and career stage on ethical
judgement. J Business Ethics. 1999;20:301-13.
47. Arlow P, Ulrich TA. Business ethics and social
responsibility of business students: An empirical
comparison of Clark’s study. Akron Business Economic
Rev. 1980;11:17-23.
48. Chonko LB, Hunt SB. Ethics and marketing
management: An empirical examination. J Business
Res. 1985;13:339-59.
49. Kidwell JM, Stevens RC, Bethke RL. Differences in the
ethical perceptions between male and female managers:
Myth or reality. J Business Ethics. 1987;6:489-93.
50. Teal EJ, Carroll AB. Moral reasoning skills: Are
entrepeneurs different. J Business Ethics. 1999;19:22940.
51. Reiss MC, Mitra K. The effects of individual difference
factors on the acceptability of ethical and unethical
workplace behaviors. J Business Ethics. 1998;17:158193.
52. Feist GJ, Gorman ME. The psychology of science:
Review and integration of a nascent discipline. Rev
Gen Psychol. 1998;2:3-47.
53. Feldman DH. The development of creativity. In:
Sternberg RJ, ed. Handbook of Creativity. Cambridge,
England: Cambridge University Press; 1999:169-88.
54. Csikszentmihalyi M. Implications of a system’s
perspective for the study of creativity. In: Sternberg RJ,
ed. Handbook of Creativity. Cambridge, England:
Cambridge University Press; 1996:313-38.
55. Amabile TM. Motivational synergy: Toward new
conceptualizations of intrinsic and extrinsic motivation
in the work place. Hum Resource Management Rev.
1993:3;185-201.
56. Collins MA, Amabile TM. Motivation and creativity.
In: Sternberg RJ, ed. Handbook of Creativity.
Cambridge, England: Cambridge University Press;
1999:297-312.
57. Bowie NE. A Kantian theory of meaningful work. J
Business Ethics. 1998;17:1083-92.
58. Eisenberger R, Cammeron J. Detrimental effects of
reward: Reality or myth? Am Psychol. 1996;51:1153-6.
59. Baumhart R. How ethical are businessman. Harvard
Business Rev. 1961;39:6-19.
60. Bellizi JA, Hite RE. Supervising unethical sales force
behavior. J Marketing. 1989;53:36-47.
61. Deconinck JB. How sales managers control unethical
sales force behavior. J Business Ethics. 1992;11:78998.
62. Harrison JS, Fiet JO. New CEOs pursue their own selfinterests by sacrificing stakeholder value. J Business
Ethics. 1999;19:301-8.
63. Pelz DC, Andrews FM. Scientists in Organizations:
Productive Climates for Research and Development.
New York: Wiley; 1966.
64. Simonton DK. Historiometric studies of creative
genius. In: Runco MA, ed. The Creativity Research
Handbook: Volume One. Cresskill, NJ: Hampton Press;
1997:3-28.
65. Mullins RF, Sherman R. Creativity and performance
appraisal: Shall never the twain meet. Creativity Res J.
1993;6:425-34.
66. Yukl G. Leadership in Organizations. Englewood
Cliffs, NJ: Prentice-Hall; 1998.
67. Schminke M, Wells D. Group processes and
performance and their effects on individual’s ethical
frameworks. J Business Ethics. 1999;18:367-81.
68. Zabid ARM, Alasgoff, SK. Perceived ethical values of
Malaysian managers. J Business Ethics. 1993;12:3317.
69. Sims RR. Changing an organization’s culture under
new leadership. J Business Ethics. 2000;25:65-78.
70. Minkes AC, Small MW, Chatterjee SR. Leadership and
business: Does it matter? Implications for management.
J Business Ethics. 1999;20:327-35.
71. House RJ, Howell JM. Personality and charismatic
leadership. Leadership Q. 1992;3:81-108.
72. Meyer AG, Brown T, Browne MN, Kubasek N. Do we
really want more leaders in business? J Business Ethics.
1998;17:1727.
73. Vredenburgh D, Brender Y. The hierarchical abuse of
power in work organizations. J Business Ethics.
1998;17:1337-1347.
74. Baron F, Harrington D. Creativity, intelligence, and
personality. Annu Rev Psychol. 1981;32:439-76.
75. Mumford MD, Whetzel DC, Reiter-Palmon R.
Thinking creatively at work: Organizational influences
on creative problem solving. J Creative Behav.
1997;31:7-17.
88
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
76. Houndshell DA. Invention in the industrial research
laboratory: Individual act on collective process? In:
Weber RN, Perkins DN, eds. Inventive Minds:
Creativity in Technology. Cambridge, England:
Cambridge University Press; 1992:273-390.
77. Reall M, Bailey JJ, Stoll SK. Moral reasoning on hold
during a competitive game. J Business Ethics.
1998;17:1205-10.
78. Schwepker CH. Understanding salespeople’s intention
to behave unethically: The effects of perceived
competitive intensity, cognitive moral developments,
and model judgement. J Business Ethics. 1999;21:30316.
79. Feist GJ. The influence of personality on artistic and
scientific creativity. In: Sternberg RJ, ed. Handbook of
Creativity. Cambridge, England: Cambridge University
Press; 1999:273-96.
80. Zuckerman H, Cole JR. Research strategies in science:
A preliminary inquiry. Creativity Res J. 1994;7:391406.
81. Pelton LE, Chowdhury J, Vitell SJ. A framework for the
examination of relational ethics: An interactionist
perspective. J Business Ethics. 1999;19:241-53.
82. Rallapalli KC, Vitell SJ,, Szeinbach S. Marketers’
norms and personal values: An empirical study of
marketing professionals. J Business Ethics.
2000;24:65-75.
83. Cote J, Goodstein J. A breed apart? Aecurity analysis
and herding behavior. J Business Ethics. 1999;18:30514.
84. Gessner TE, O’Connor JA, Clifton TC, Connelly MS,
Mumford MD. The development of moral beliefs: A
retrospective study. Curr Psychol. 1993;11:236-59.
85. Leede JD, Nijhof AHJ, Fisscher OAM. The myth of
self-managing teams: A reflection on the allocation of
responsibilities between individuals, teams, and the
organization. J Business Ethics. 1999;21:203-15.
86. Barnett T. A preliminary investigation of the
relationship between selected organizational
characteristics and external whistle blowing by
employees. J Business Ethics. 1992;11:949-59.
87. Kawathatzopoulos I. Development of cognitive skill in
solving business ethics problems: The effect of
instruction. J Business Ethics. 1993;12:379-86.
88. Verbke WC, Ouwerkerk C, Peelen E. Exploring
contextual and individual factors on ethical decisionmaking of sales people. J Business Ethics.
1996;15:1175-87.
89. Weaver KM, Farrell OC. The impact of corporate
policy in reported ethical beliefs and behavior of
marketing practicioners. AMA Proc. 1977;27:477-81.
90. Fritz JMH, Arnett RC, Conkel M. Organizational
ethical standards and ethical commitment. J Business
Ethics. 1999;20:289-92.
91. Grimalda G. Participation versus social exclusion. J
Business Ethics. 1999;21:269-79.
92. Schokkaert E, Sweeney J. Social exclusion and ethical
responsibility: Solidarity with the least skilled. J
Business Ethics. 1994;21:251-67.
93. Castrogiovani GT. Pre-start-up planning and the
survival of new small businesses: Theoretical linkages.
J Management. 1996;22:801-22.
94. Morris MH, Marks AS, Allen JA, Perry NS. Modeling
ethical latitudes and behaviors under conditions of
environmental turbulence: The case of South Africa. J
Business Ethics. 1996;15:1119-30.
95. Roussouw GT. Establishing moral business culture in
newly formed democracies. J Business Ethics.
1998;17:1563-71.
96. Verschoor CC. A study of the link between a
corporation’s financial performance and its commitment
to ethics. J Business Ethics. 1998;17:1509-16.
97. Judge WO. Correlates of organizational effectiveness: A
multilevel analysis of a multi-dimensional outcome. J
Business Ethics. 1994;13:1-10.
98. Zarkada-Fraser A. A classification of factors
influencing participating in collusive tendering
agreements. J Business Ethics. 2000;23:269-82.
99. Enderle G, Tavis LA. A balanced concept of the firm
and the measurement of its
long-term
planning and performance. J Business Ethics.
1998;17:1124-44.
100.Koehn D. Can and should businesses be friends with
one another and their stakeholders. J Business Ethics.
1998;17:1755-63.
101.Valenzuela JLD, Villaconta FS. The relationship
between companies and their suppliers. J Business
Ethics. 1999;22:273-80.
102.Jones MT. The institutional determinants of social
responsibility. J Business Ethics. 1999;20:163-79.
103.Vitell S, Festervand TA. Business ethics: Conflicts,
practices, and beliefs of industrial executives. J
Business Ethics. 1987;6:111-22.
104.Dubinsky AJ, Ingram TN. Correlates of salespeople’s
ethical conflict: An exploratory investigation. J
Business Ethics. 1984;3:343-53.
105.Dooley RS, Fryxell GE. Are conglomerates less
environmentally responsible? An empirical examination
of diversification strategy and subsidiary pollution in the
U.S. chemical industry. J Business Ethics. 1999;21:114.
106.Reichers AE, Schneider B. Climate and culture: An
evolution of constructs. In: Schneider B, ed.
Organizational Climate and Culture. San Francisco,
CA: Josey-Bass; 1990:5-39.
107.Sims RL, Keon TL. Determinants of ethical decision
making: The relationship of the perceived organizational
environment. J Business Ethics. 1999;20:393-404.
108.Agarwal J, Malloy DC. Ethical work climate decisions
in a not-for-profit organization: An empirical study. J
Business Ethics. 1999;20:1-14.
109.Vidader-Cohen D. Moral climate in business firms: A
conceptual framework for the analysis of change. J
Business Ethics. 1998;17:1211-26.
110.Key S. Organizational ethical culture: Real or
imagined? J Business Ethics. 1999;20:217-25.
111.Argandona A. Sharing out in alliances: Trust and ethics.
J Business Ethics. 1999;21:217-28.
89
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
112.Hartman SJ, Yrle AC, Galle WP. Procedural and
distributive justice: Examining equity in a university
setting. J Business Ethics. 1999;20:337-51.
113.Amabile TM, Griskiewicz N. The creative environment
scale: The work environment inventory. Creativity Res
J. 1989;2:231-54.
114.Gotterbarn D. Not all codes are created equal: The
software engineering code of ethics, a success story. J
Business Ethics. 1999;22:81-9.
115.Etheredge JM. The perceived role of ethics and social
responsibility: An alternative scale structure. J Business
Ethics. 1999;18:51-64.
116.Brierley JA, Cowton CJ. Putting meta-analysis to work:
Accountants’ organizational-professional conflict. J
Business Ethics. 2000;24:343-53.
90
3. Medical Practice and Clinical Research
Waiving Informed Consent: Long-Term Consequences for the U.S. Military
Mary L. Cummings, Engineering Fundamentals, Virginia Tech, USA
Keywords: Anthrax, Gulf War, Informed consent, Investigational drugs, Military, Waiver
In December 1990, the Department of Defense (DoD), anticipating the invasion of Kuwait for
Operation Desert Storm, petitioned the Federal Drug Administration (FDA) to waive the federally
mandated informed-consent requirements in the case of two investigational drugs: pyridostigmine
bromide (PB) and botulinum toxoid (BT). PB, administered orally, was thought to be an effective
pre-treatment against the nerve agent soman. The BT vaccine was potentially effective against the
bacterium causing botulism (1). Fearful of the possibility that Saddam Hussein would conduct
chemical and biological warfare against American troops, the Joint Chiefs of Staff felt that these two
investigational drugs could protect U.S. soldiers. The concerns of military leadership were wellfounded. Saddam Hussein had used chemical nerve agents and mustard gas against his own people in
the Iran-Iraq War (2). However, while military intelligence confirmed that Iraq had the capability to
make biological and chemical (nerve agent) weapons, no evidence indicated Iraq had ever made a
weapon with soman (3).
FDA did not approve PB and BT. They were considered experimental and fell under the category
of investigational new drug (IND). Federal regulations stipulate that if any Federal agency, including
the military, desires to use an unapproved drug, that agency must first fully brief the individuals
receiving the IND. This briefing must include mention of associated drug use hazards, and the
potential recipients’ written consent must be obtained. Prior to the Gulf War, informed consent for
INDs could only be waived in extreme emergencies, even for the military. However, the U.S. military
determined that it was not feasible to seek the informed consent of 700,000 personnel deployed to the
Middle East. In 1990, in the months preceding the Gulf War, the military petitioned the FDA to
waive the informed consent regulations. The FDA, not wishing to intervene in national security
policy and with the approval of an Institutional Review Board (IRB), issued the waiver in an interim
ruling in December 1990 (4). However, as part of the approval for the waiver, the military was
required to provide information sheets about PB and BT to the recipients detailing the possible side
effects. In addition, the military was expected to carefully document the use of the INDs as well as
any adverse reactions.
Approximately 300,000 military personnel received the PB pills and 8000 individuals received
the BT vaccine during the Gulf War (5). Despite the specific requirement by the FDA that the
military track data on both drugs, no procedure was ever established to document which personnel
received the drugs and if any adverse side effects were noted (1). Many military personnel
experienced systemic medical problems both during and after the Gulf War that were not combat
related. Such problems have been termed as the Gulf War Syndrome (GWS). Most notably, over
100,000 Gulf War veterans complained of maladies ranging from chronic fatigue to paralysis in the
Corresponding author: Mary L. Cummings, 332 Randolph Hall,Virginia Polytechnic Institute and State University,
Blacksburg, VA 24061, 540-231-6555 (voice), 540-231-6903 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
years immediately following the war (3), and of
these, 20,000 reported debilitating symptoms (6).
In preliminary studies, PB has now been
implicated as the primary catalyst of the GWS,
however the research is still in its early stages
(3).
Waiving Informed Consent
The Federal regulations that govern informed
consent for human subjects fall under the
purview of the Department of Health and Human
Services (DHHS). The regulations state that
informed consent may be waived when using
INDs, but a number of conditions must be met.
No more than minimal risk can exist for the
patient, and after the treatment is concluded, the
participants must be notified of both the
procedure and the possible risks (7). FDA, bound
by the DHHS regulations, established their own
framework of rules regarding INDs. Prior to the
Gulf War waiver, FDA maintained that the
informed consent process could be waived only
in a life-threatening emergency with the patient
unable to communicate and without time to
obtain consent from patient’s legal representative
(7).
The Joint Chiefs of Staff decided it was not
feasible to obtain the informed consent of
700,000 military personnel deployed to the Gulf
War region and that the pending conflict was
essentially an emergency situation by FDA
standards. However, prior to granting the
military informed consent waivers for the use of
PB and BT, FDA required the military to convene
an IRB (1). To meet this Federal requirement for
the BT vaccine, the military actually convened
two IRBs. The first IRB, the U.S. Army Medical
Research Institute of Infectious Diseases
(USAMRIID) Human Use Committee, was the
panel typically used by Army research personnel
to consider protocols involving human subjects.
The USAMRIID concluded that it was unethical
to waive the informed consent of military
personnel who would receive BT (8). They
further recommended that oral, not written,
consent be obtained because oral consent was
feasible, and it also respected the rights of the
soldiers. Six days later, for reasons not stated in
any DoD documents or in any IRB minutes, the
DoD then convened a second, entirely different
IRB, the Surgeon General’s Human Subjects
Research Review Board (HSRRB). The HSRRB
approved the BT protocol as submitted and
recommended that informed consent be waived
(9).
Even though FDA waived the requirement
for obtaining informed consent for the use of PB
and BT in the Gulf War, the approval was
contingent upon the military providing those
service members who received the INDs with
information sheets describing the PB and BT
treatments in detail. The sheets were to explain
the reasons for using the INDs, the symptoms of
botulism and a nerve agent attack, and most
importantly any potential side effects or
reactions. In addition, the soldiers were also
asked to report any of these side effects or
reactions. Apparently, the information sheets
never made it to the Gulf War theater, so the
personnel who received the treatments did not
receive any written information about the INDs.
However, even a cursory glance at the
information sheets that were approved by the
Army for dissemination shows that they were at
best superficial.
Ethical Issues
In 1978, the National Commission for the
Protection of Human Subjects of Biomedical and
Behavioral Research issued the Belmont Report
that identified three principles that are
fundamental in determining whether a research
protocol is ethical. They are: respect for persons,
beneficence, and justice. These are the primary
ethical considerations of an IRB when evaluating
a research protocol (10). The crux of the respectfor-persons principle is the preservation of a
person’s autonomy when making decisions about
his/her own medical care. It is this aspect of the
Belmont Report that is at issue in waiving
informed consent. By swearing an oath to the
military and the nation, service members
willingly sacrifice some autonomy concerning
decisions about their own lives. Enlisting in the
military is a supreme sacrifice and highly
commendable, but should soldiers lose all rights
to autonomy, especially when it comes to their
health? The DoD defends its actions in waiving
informed consent for INDs by stating, “Allowing
a soldier to refuse treatment would endanger him/
her as well as those who would try to save their
lives and ruin mission success”(5). This
paternalistic approach by the DoD overlooks one
critical aspect: What exactly constitutes
“treatment?”
There has been much debate as to whether
the military’s use of PB and BT constitutes
research or treatment. In the clinical trials held
94
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Cummings, Waiving Informed Consent
with no impediments. The military was clearly
circumventing the system and in doing so
trivialized the IRB process and violated Federal
regulations. It appears the military was only
seeking IRB approval as a formality in an
administrative procedure and lost sight of the
purpose of the review. FDA, very concerned
about the military’s use of multiple IRBs when
seeking informed consent waivers, censured the
military in October of 1999 for this violation and
changed the federal regulations regarding
military IRBs (1). As a result, IRBs convened by
the military to evaluate IND protocols are now
required to include at least three members who
are not employees or officers of the federal
government and are not affiliated with the
protocol in any way.
months before the Gulf War, only a select group
of male human subjects were tested with PB and
BT. There was no testing for interactions with
other chemicals or drugs likely to be used with
the INDs, and no long-term studies were
conducted (5). Additionally, persons with health
problems typical of military populations were
never studied in conjunction with the drug
testing, and women never participated in any
trials (2). Is it ethical and reasonable to maintain
that military members receiving drugs tested on a
very small, isolated population were receiving
“treatment?” Despite the fine line between
treatment and research with investigational
drugs, FDA’s own regulations clearly state that
informed consent is required even when the
unapproved drug is to be used in a therapeutic
manner because the drug has not yet passed full
FDA efficacy and safety trials (11).
The respect-for-persons principle was again
violated when the information sheets for the
INDs were “lost” (5, 12). These sheets should
have been paramount in the minds of military
medical professionals overseeing the PB & BT
programs. The IRB approval and FDA
authorization for PB and BT were contingent on
the investigators adhering to the approved
protocols, which included the distribution of the
information sheets. The INDs found their way
successfully to the Gulf War theater, and if DoD
leadership had considered the sheets a similar
priority, they would have been delivered also.
Did the military view the information sheets as
“not feasible” just as they did for informed
consent? When FDA later evaluated the
military’s use of INDs during the Gulf War, it
identified “significant deviations from Federal
regulations published in Title 21, Code of Federal
Regulations (CFR), parts 50 and 312.” (1). FDA
cited several areas in which the military was not
in compliance. Most notably FDA admonished
the military for not disseminating the information
sheets prior to the use of INDs in the Gulf War.
FDA also issued DoD a stern reprimand for not
keeping detailed records on who received the
drugs and, most importantly, any adverse
reactions suffered by military personnel.
Lastly, the most glaring ethical issue was
DoD’s use of two different IRBs. When the
Army’s first IRB found that it was unethical to
administer BT to military personnel without their
informed consent, the DoD convened a second
IRB that produced the desired result of
recommending the waiver of informed consent
Long-Term Consequences
In December 1997, DoD announced plans to
vaccinate all 2.4 million U.S. troops against the
biological threat of anthrax. If not treated in its
initial stages, anthrax is deadly (13). The current
anthrax vaccine is approved by the FDA and was
originally designed for agricultural workers and
veterinarians. It is a six-shot protocol that is
administered over a period of 18 months.
Because of this extended treatment period, DoD
decided that it must vaccinate all 2.4 million
personnel in the unlikely event that all U.S.
forces faced a biological threat.
Almost immediately after DoD made its
announcement, military members began to
protest, based in part on the revelation that
service members were given experimental drugs
without their knowledge in the Gulf War.
Military, medical, and legal critics of the anthraxvaccine decision were not satisfied that the
vaccine was approved by the FDA (13 -15). The
sole manufacturer of the anthrax vaccine,
Michigan Biologic Products Institute (now BioPort) has failed numerous FDA inspections.
Most recently, Bio-Port was cited for 23
violations, some of which included sterility and
potency deviations, and some microbial
contamination (14, 15). In fact, to date the
Michigan plant still has not passed an FDA
inspection (15, 16).
There have never been any published studies
of human efficacy or long-term effects for the
anthrax vaccine (15). Moreover, according to an
April 1999 General Accounting Office (GAO)
report, long-term effects of the anthrax vaccine
have never been studied. To further add to the
95
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
debate over the efficacy of the anthrax vaccine,
the Institute on Medicine has stated that the
licensed anthrax vaccine is only effective against
cutaneous anthrax and furthermore has never
been tested for pulmonary anthrax, which would
be the method of delivery in a combat arena (13).
A chief Army biological researcher wrote in a
1994 textbook on vaccines that “the current
vaccine against anthrax is unsatisfactory” (14).
Despite the military’s assertions that it is only
interested in protecting the welfare of its soldiers,
GAO charges that DoD is extremely negligent in
tracking adverse reactions to the anthrax vaccine,
which was a significant problem with the INDs
used in the Gulf War. In fact, many military
personnel have reported adverse reactions to the
anthrax vaccine. However, in the absence of any
established tracking and monitoring system, there
is no way to accurately identify any percentages.
With the data supporting the questionable
status of the anthrax vaccine and considering
DoD’s past history, it is not unreasonable to
expect military personnel to have doubts about
both the efficacy of the anthrax vaccine and the
military’s plans for implementation. To combat
potential insubordination, DoD court-martialed
those personnel who refused the vaccine, stating
that allowing soldiers to refuse the vaccine would
undermine discipline and be prejudicial to good
order. Many military members, outraged at
DoD’s response and facing involuntary
inoculation, chose to resign from the service
rather than risk their health. The military is
already facing serious retention and recruiting
problems, and DoD’s refusal to make the anthrax
vaccine voluntary is only adding to an already
critical personnel shortage.
Prior to the mandated anthrax vaccination of
all U.S. troops, the military’s policies against the
threat of chemical and biological warfare were
deterrence, containment of the enemy, and use of
other defensive measures such as protective suits
and warning devices (13). It was not until the
Gulf War that troops were inoculated against the
threat of possible biological warfare, and it was
not until 1997 that troops were forcibly
inoculated in peacetime. There has been much
criticism directed toward DoD for implementing
the anthrax vaccine in peacetime. DoD
responded that even though there is no threat of
war, the 18-month treatment period for the
anthrax vaccine requires that it must prepare its
forces for any future contingencies. However,
GAO asserts that based on military intelligence
data, the biological warfare threat for U.S. troops
has not changed since 1990 (14).
A Final Note on Accountability
Accountability is an imperative moral trait
required of all military personnel and is
considered the cornerstone for military command
and leadership. By court-martialing military
personnel who refuse the anthrax vaccine, DoD
is holding these people accountable for their
actions. For those court-martialed, this
accountability will not cost them just their jobs
within the military. In addition, they are
dishonorably discharged and lose all their
veterans’ benefits as well as their retirement
benefits. The nation recognizes the right to make
autonomous health-related decisions for all
citizens, but it appears, not for military personnel
who pay a high price for both autonomy and
accountability.
This exacting level of military discipline and
accountability is unfortunately glaringly absent
from DoD’s use of INDs in the Gulf War.
Especially troubling are the following:
• DoD convened a second IRB for an IND
protocol when the first did not produce the
desired recommendation to waive informed
consent.
• No one was held accountable for the lost
information sheets in the Gulf War. If
military officers lost strategic documents
protecting troops’ safety, they would most
definitely face severe punishment.
• No one was held accountable for the incredible lack of record keeping including tracking adverse reactions during and after the
Gulf War. Not only did military personnel
suffer from a lack of treatment information,
but also the entire medical field suffered
from the loss of critical data.
This clear double standard in accountability will
only continue to haunt the military. Public
reports on the military’s use of experimental
drugs on troops without their knowledge and the
anthrax debacle will only continue to exacerbate
personnel issues. FDA has recently issued more
stringent rulings to prevent some of these ethical
transgressions from occurring in the future and to
compel the military to abide by the laws they are
supposedly defending. However, not until DoD
embraces the Federal policies designed to respect
basic human rights and autonomy will the
96
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Cummings, Waiving Informed Consent
military regain some of its medical credibility
and confidence in leadership.
Acknowledgments
I would like to extend a special thanks to Dr.
Doris T. Zallen for her guidance and technical
support.
Bibliography
1. 64 Federal Register at 54,180; October 5, 1999.
2. Milner CA. Gulf War guinea pigs: Is informed consent
optional during war? The Catholic University of
America Journal of Contemporary Health Law & Policy
1996 Fall; 13: 199-232.
3. Myers SL. Drug May Be Cause of Veterans’ Illnesses.
New York Times 1999 Oct 18; Sect. A:18, (col. 4).
4. 55 Federal Register at 52,187 (1990) codified at 21 CFR
50.23(d).
5. Ryan RP. Should combat troops be given the option of
refusing investigational drug treatment? The Food and
Drug Law Institute Food and Drug Law Journal 1997;
52(4): 377-400.
6. Tuthill KA. Human Experimentation: Protecting Patient
Autonomy Through Informed Consent. The Journal of
Legal Medicine 1997 Jun; 18: 221-250.
7. Code of Federal Regulations: 45 CFR 46.116(c), (1999).
8. U.S. Army Medical Research Institute of Infectious
Diseases memorandum SGRD-UIZ-H. Fort Detrick,
Maryland: Department of the Army; October 5, 1990.
9. Office of the Surgeon General memorandum SGRD-HR
(15-1a). Falls Church, VA: Department of the Army;
October 11, 1990.
10. The Belmont Report: Ethical Principles and Guidelines
for the Protection of Human Subjects of Research.
Washington DC: DHEW Publication (OS) 78-0011;
1978.
11. Code of Federal Regulations: 21 CFR 312.35 (2000).
12. The Pill. “60 Minutes.” CBS television; September 29,
1996.
13. Havrilak GT. The Pentagon’s Anthrax Vaccination
Immunization Program. Military Readiness Project;
1999 Dec.
14. Rempfer TL. Why Am I Resisting the Vaccine? The
Military Trained Me To. The Washington Post 2000 Jan
3; Sect. B:01.
15. Meryl Nass MD. Written testimony before the
Subcommittee on National Security, Veterans’ Affairs,
and International Relations, U.S. House of
Representatives. April 19, 1999.
16. The Court-Martial of Major Sonnie Bates: “60
Minutes.” CBS television; February 6, 2000.
97
Promoting Scientific Integrity: The Long Road Ahead–Some Considerations
from Espírito Santo, Brazil
Jaime Roy Doxsey, Department of Social Sciences, Federal University of Espírito Santo,
Brazil
Keywords: Brazil, Research integrity, Social and human sciences, Teaching research ethics
“We live in an historical moment of transformation of the scientific paradigm which
questions the criteria that scientific rigor in and of itself is ethical.” BB Sawaia, 1999.
While the promotion of research integrity has tended to receive widespread governmental and
institutional support in the United States and Canada, the responsible conduct of research, including
preventing and handling of misconduct, are not always prominent issues in many developing
countries such as Brazil. This paper examines the need to stimulate institutional awareness and
debate on major issues such as production and communication of scientific knowledge as well as the
ethical challenges for developing responsible research practices in the human and social sciences.
A lack of Federal or state legislation, institutional policies or public concern regarding the quality
and the ethics of scientific research do not exempt researchers or universities from establishing
programs to insure research integrity. The institutional context of a medium-sized Federal
government university, the Federal University of Espírito Santo, is examined in an attempt to describe
work conditions, the institutional culture and other obstacles for establishing a program to promote
research integrity.
In Brazil, recent Federal resolutions in the areas of health, medicine and medical research have
established guidelines for human protocol, research integrity, and the protection of human subjects
and have determined a local project review procedure along the lines of North American legislation.
These guidelines extend themselves to all scientific or academic research activities that involve
human subjects. The Brazilian university system and the National Council for Research (CNPQ),
however, have neither acknowledged the relevance of these resolutions for research practices nor
incorporated them into grant procedures.
At the local level, universities, research institutes, academic centers, departments and graduate
programs establish their own policies for research projects and scientific production. Institutional
procedures seldom exist for handling allegations of scientific misconduct or establishing protocols for
human subjects.
The recent expansion of the number of graduate programs also has increased the need for
programs to promote the teaching of research integrity, the ethics of mentoring, and academic career
pressures. Further, data management, recording, retention, etc., require pro-active policies to
anticipate conflicts and incidents of misconduct.
Corresponding author: Jaime R. Doxsey, Department of Social Sciences, Federal University of Espírito Santo, Rua
Pernambuco, 81/1102, Praia da Costa, Vila Velha, Espírito Santo 29.101-335, Brasil, 55-27-3349-8992 (voice), 55-27-33357614 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
What are the implications of these conditions
for research with human subjects in Brazil? Is
the Brazilian population unduly exposed to
doubtful research practices and scientific
misconduct, particularly the lower population
strata (over 50% of the total population) and
more specifically, vulnerable sectors of this
population?
At first glance, the answer would be an
uncategorical “no”. Even considering the lack of
a more systematic analysis of actual research
practices, there is no direct or indirect evidence
that medical, health, human, or social sciences
research in Brazil is unethical. What could be
considered unethical is the lack of priority for
such research at all levels of government in light
of the rising indices of preventable social
diseases, human violence, drug abuse, and the
subsequent decline of living conditions/quality of
public services for the lower strata of the
population.
With financial support and investment in
social policies at an astonishingly low level,
social research tends to be descriptive,
exploratory, or action-oriented. Academic
research seldom receives external or internal
financing, and most funding is limited to
scholarships for undergraduate trainees or the
support of field work.
The lack of a regulatory system of project
approval and norms for the protection of human
subjects should not be misinterpreted as a lack of
research ethics. In a country like Brazil, the few
individuals actively engaged in research with
human subjects do so with great dedication and
considerable respect for their human subjects.
Ethical values are not necessarily culturally
ascribed or limited by adverse institutional and
social conditions.
Nevertheless, what are the actual
circumstances in which the social and human
sciences are being practiced in Brazil? In what
institutional context might it be necessary to
initiate the promotion of research integrity and at
least provide guidelines for misconduct
regulation? How may this promotion of research
integrity be best approached?
Design
This paper is a descriptive essay based on
personal observations and a review of scientific
journals, research methodology textbooks
published in Portuguese, Internet homepages,
records of research projects available in the Pró-
Rectory for Graduate Study and Research,
Federal University of Espírito Santo and the
annual reports of the Office of Research Integrity,
Department of Health and Human Services, U.S.
Office of Public Health and Science. The journal
editions of the Cadernos de Ética em Pesquisa
[Notebooks of Research Ethics], published by the
Brazilian National Commission of Research
Ethics were specially useful in providing
background information for this text.
Results–The Brazilian Context
In Brazil, Federal resolutions first established the
National Commission of Research Ethics
(CONEP) in 1996 and determined guidelines for
human protocol, research integrity, and the
protection of human subjects in 1997. The 1997
resolution determined a project review procedure
in the areas of health, medicine, and medical
research by local Committees of Ethics and
Research. At the present time, there are
approximately 266 Committees of Ethics and
Research (CEPs), the majority of which are
located in institutions related to medical
instruction or university-associated hospitals.
Although the guidelines extended themselves
to all scientific or academic research activities
that involve human subjects, the Federal
Brazilian university system and the CNPQ have
neither acknowledged the relevance of these
resolutions for research practices nor
incorporated them into institutional procedures.
Data from CONEP reveal the registration of
559 projects in 1999. In a classification by
Specialty Topics, most of these projects were
grouped under the topic of “international
cooperation” (78.3%), and a majority within this
category (80%) involved new medications.
Distribution in other topical areas included
human genetics (7.8%), reproduction (5%),
indigenous populations (1.6%), new medical
procedures, and equipment (5.3%) (1).
In observance of the data cited above, it is
not surprising to conclude that medical and
health research formally lead the way in
establishing human protocols for research with
human subjects. Also, it is not accidental that the
majority of the projects reviewed involve
international funding and/or cooperative
agreements. A recent review of the literature
available within Brazil points exclusively toward
bioethics and medical and health ethics as
dominant topics in the field of ethical
considerations (2).
100
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Doxsey, Promoting Scientific Integrity
In the humans sciences, there is little to
report. However, in 1997, the Federal Council of
Psychology determined that new methods or
procedures in the field could be utilized if
presented as research following research norms
for human subjects. The Committee of Ethics in
Research at the Catholic University of São Paulo
(Catholic University – SP) was implemented
through the work of a sociologist who lead
discussions to delimitate general principles
regarding research ethics, which “took into
consideration the specificity, plurality and
scientific creativity of the production of
knowledge in the human sciences” (3).
Unlike the CEPs created in the medical area,
at the Catholic University-SP, the Committee has
developed educational functions to represent the
ethical principles of the institution, serving as a
review board for special recourses. Research
projects that are considered to have special
ethical questions are sent to the Committee by
academic orientators, or by dissertation, thesis, or
research commissions for educational
evaluations. This university understood that
ethical evaluations were already occurring at
other institutional levels and that the
centralization of the approval process in one
committee would be not only impossible but
would fail to capture the different optics of
research ethics.
Another indicator of the extent of concern for
research integrity was presented in a study
entitled: “Analysis of ethical aspects of research
in human beings contained in the authors’
instructions of 139 Brazilian scientific journals”.
(4) Although the study was limited to a review
of scientific journals in the areas of medicine,
nursing, odontology, and the general sciences, the
authors discovered that 79% of the journals made
no reference to ethical considerations in their
notes to potential contributors. Only 12% of the
journals made reference to the necessity of
approval or analysis of the research project by a
Committee or Commission of Ethics in Research.
This author has no knowledge of instructions
to authors in the area of the social and human
sciences. With the growing number of scientific
publications in Brazilian universities, there is
some concern for establishing selection processes
for articles and the evaluation process of the
journals. During May, the Faculty of Education
at the University of São Paulo organized a
conference to discuss the publication policies of
scientific journals in education. Discussion was
focused on the role of the journals in improving
research quality, technical aspects of the journals,
and proceedings for evaluation/selection of
articles. The last session included an item on
scientific and ethical aspects of journal editing.
Increased public concern with electoral
opinion polling has attracted attention in the last
national elections for president and congress, and
most recently in municipal elections. The
concern voiced by media and politicians is
directed, however, to the possible undue
influence of the poll results on the voter and the
political system. No ethical concern for poll
subjects has been registered. Issues regarding
informed consent, the use of the poll results, or
the subjects’ knowledge of the funding sources
have not been publicly evaluated.
Although the lack of governmental support
for scientific and technological research and
development is a constant criticism throughout
the Brazilian society, there is no strong public
support for financing academic research.
Resources from private and international
foundations are centered on corporate interests
with little direct university participation. In
short, there is little grant money, private or
public, which might warrant an institutional
policy being created in order to qualify for grant
applications.
While international funding or “cooperation”
might be instrumental in aligning research
interests in the biomedical sciences to installing
parallel regulatory proceedings for research
ethics, there are no similar external stimuli for
the human and social sciences in Brazil. With no
public pressure or support for human research,
little or no funding, and a lack of issues that
might stimulate institutional response tend to
neutralize the need for more relevant,
modernized research policies in the Brazilian
University system.
A Short Case Study–the UFES
Current research policies at the Federal
University of Espírito Santo deal principally with
the administrative approval of faculty
involvement in research as well as release time
from academic classroom schedules.
Authorization to conduct research is granted by
the department council, after a written evaluation
often by a research commission of peers. A
simplified regulatory system presently requires
project approval by the council of department
heads at the level of the academic center and
101
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
eventual registration of the project in the PróRectory for Graduate Studies and Research.
Details of the project must be outlined on a
basic form that specifies the usual information
regarding the nature of the study, authors,
methods, objectives, and bibliography. No
human protocol is required. References to study
samples, human subjects, and data collection
procedures, when indicated, usually are located
in a section on “methodology.”
Research projects involving human subjects
must have the approval of the Committee on
Ethics in Research only for professors from the
Biomedical Center. This Committee was
registered in March of 1997. No communication
from this committee to other academic centers
has been documented by the institution. The
potential institutional role of this committee
could be to distribute and discuss the present
regulations, which affect other areas of
knowledge.
The lack of information on the necessity for
compliance with existing regulatory standards for
human protocol or the absence of academic/
administrative requirements for recognizing the
ethical consideration of data collection with
human subjects are seen as substantial obstacles
for promoting research integrity. However, the
implications for dealing with possible
misconduct are the most serious.
The first dilemma is the extreme negligence
with which most universities treat their internal
problems of human communication and
academic relationships among faculty and
students, with no viable procedures or
mechanisms to identify, solve, or prevent such
problems. In the case of the public Federal
universities, professors and university
functionaries are classified, by law, as federal
public servants, subject to Federal legislation.
The legislation is basically a disciplinary regime
where duties and obligations are specified.
Denouncements of irregularity/misconduct are
treated administratively in an process that can
consume a year or more.
These laws as well as the university statues
and internal regulations date from the years of
the military dictatorship in Brazil, seldom having
been reformed to establish a less authoritarian
academic administrative structure. These
instruments refer to problems with faculty or
student behavior in terms of order and discipline,
keywords common to public policy of the
military government. Academic problems
involving misconduct in research, plagiarism,
misrepresentation of academic production or
other problems of research integrity can only be
handled administratively under the existing
legislation and institutional procedures (5).
In synthesis, academic or research integrity
as a terminology or concept plays little part in the
actual institutional culture, or at least is not
configured as a formal organizational principle in
the university culture. This is not to say that
academic integrity is not present in many of the
pedagogical and academic actions of students
and faculty, nor in the daily practices of this
institutional culture. Nevertheless, the fact that
academic/scientific ethics or research integrity
are not explicitly registered in formal university
institutional norms considerably complicates the
institutional capacity to develop scientific
integrity and deal with ethical problems of any
nature.
Conclusions
These results confirm the necessity for urgent
institutional action to establish normative
standards that promote a responsible research
environment and a critical consciousness of the
need for training/research in scientific integrity in
all areas of knowledge. However, the
advancement of academic/scientific ethics
depends upon a critical analysis of present
research practices and the recognition of the
protection of human subjects as one component
of research integrity inherently connected to the
ethical production of knowledge.
Institutional research is needed to identify
academic areas with accessibility for a new
approach to teaching research integrity as well a
current researchers’ concerns with research
ethics. Institutional support for such curriculum
reform is vital, but must occur with a greater
strategy to set university goals for excellence in
research with human subjects and to reform
regulations that are obsolete and ineffective in
dealing with problems of academic/scientific
integrity.
Caution is necessary to avoid
“overdeveloped” procedures that do more to
serve the rule makers than to protect the victims
of unethical research practices. Perhaps, instead
of taking the long road and merely reproducing
regulations and administrative procedures for
projects review, or awaiting federal legislation,
local universities such as the UFES should
consider the middle road, one which is not a
102
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Doxsey, Promoting Scientific Integrity
short cut or dodges vital issues, but one which
stimulates a process that provides access to
information, provides debate about research
integrity, and acknowledges institutional needs
for guidelines to avoid scientific misconduct and
to safeguard human subjects, particularly those
subjects in situations of cultural or social risk.
Bibliography
1. Bontempo de Freitas C, Oliviera Lôbo M. CONEP faz
balanço positivo [National Commission of Ethics in
Research – CONEP makes a positive evaluation].
Cadernos de Ética em Pesquisa. 2000;4;4-7.
2. A ética e a bioética em livros [Ethics and bioethics in
books] [Special article]. Cadernos de Ética em Pesquisa.
2000;5;17-19.
3. A ética nas ciências humanas [Ethics in the human
sciences] [interview]. Cadernos de Ética em Pesquisa.
2000;4;14-7.
4. Sardenberg T, Müller SS, Pereira HR, Oliveira RA,
Hossne WS. Estudo mapeia ética em revistas [Study
maps ethics in scientific journals] Cadernos de Ética em
Pesquisa. 2000;5;11-14.
5. Doxsey JR. Ética academica e cientifica numa
perspectiva psicossociológica: a escuta ativa do ouvidor
[Academic and scientific ethics in a psycho-sociological
perspective: the active listening of an ombudsman]. In:
Pinto Lyra R, organizer. A ouvidoria na esfera pública
brasileira [The ombudsman in the Brazilian public
domain]. João Pessoa (PB): Ed. Universitária, 2000. p.
143-57.
103
Ethical Research Practice with Human Participants: Problems, Procedures,
and Beliefs of Funded Researchers
Elana Newman, Department of Psychology, University of Tulsa, USA
Victoria McCoy, Department of Psychology, University of Tulsa, USA
Anthony Rhodes, Department of Psychology, University of Tulsa, USA
Keywords: Cognition, Confidentiality, Ethics, Health, Informed consent, Mental health.
Although Federal and local guidelines provide general advice as to inform researchers regarding
ethical practice (1 - 3), little information is available regarding how researchers carry out such ethical
procedures. Despite the use of Institutional Review Boards (IRBs) to monitor ethical practice, there
is great variability in how these boards operate and what types of policies are deemed acceptable (4).
Similarly, it appears that psychopathology researchers greatly differ in their practices on how to
assess and handle participant distress or injury (5 - 7). In some specialty areas, such as depression,
there is preliminary evidence that most researchers routinely give referrals (8). Nevertheless, the
range of practice is not known.
The need to document how different biomedical researchers implement ethical research policies
is important in order to generate and develop viable and informed research policy. For example, it is
helpful to understand how researchers recruit participants, train staff, obtain informed consent, and
debrief participants (9). Furthermore, specific policies about response and compensation with regard
to responding to participants’ distress, worsening of conditions, confidentiality issues, informed
consent, and other ethical dilemmas across different groups of human research participants is also
needed. Sharing such information among researchers from different disciplines, who use different
methodologies and research samples, can help to identify the range of options and the need for
training initiatives. Finally as technology makes research more global, local community standards of
practice may no longer be adequate to understand good research practice (10). To compound this
issue, distinctions between research and clinical work and research and organizational consulting are
blurring with the trends in program evaluation. Finally, advances in science have made human
experimentation itself more complex. Hence there is a need to share information and understand the
range of ethical practice in the field so we are better able to respond to these challenges and equipped
to create policy in the future.
Currently it is unknown how often research-related injuries and problems occur in the course of
routine research protocols. Although flagrant violations are reported or receive media attention, there
has been no attempt to quantify the prevalence of such problems in routine practice (11). In order to
understand participants’ responses it is also important to ascertain the actual prevalence rates of
research-related costs and injury across a wide range of samples to determine what groups need
Corresponding author: Elana Newman, Ph.D., Department of Psychology, University of Tulsa, 600 South College Avenue ,
Tulsa, OK 74104, 918-631-2248 (voice), 918-631-2833 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
additional safeguards. These risks must be
quantified to include both minor costs (abrasions,
emotional distress) and major costs (death,
disability, and needed hospitalization).
Identification of the subgroups at greatest risk for
research related harm could help inform policy
(12).
Finally the expertise of researchers and
opinions need to be shared. As documented,
opinions and assumptions about possible risks
and benefits of research participation shape
ethical appraisals of research (13 - 17).
Documenting experienced scientists’ opinions
and attitudes toward IRBs and research risk, can
help establish a clearer understanding of the
values that may shape research and research
policy.
The goal of the current study is to delineate
the rates and types of potential research-related
injuries as well as the range of ethical practices
and beliefs. This is important since several
studies document the range of ethical research
practice, but none of them actually assess the
prevalence and types of risks (8).
First, it was hypothesized that there is
considerable variability of research policies and
procedures both within and across types of
research and sample characteristics with those
researchers working with psychiatric illness
being more protective than researchers in other
areas. Policies and procedures were defined as
(a) level informed consent policy, (b) emergency
policies, (c) determination of research- related
risk, (d) debriefing procedures, (e) use of
referrals, and (f) follow-up procedures.
Second, it was hypothesized that the research
risks experienced by psychiatric health groups
will be significantly greater than those
experienced by the medical physical health
group. In addition, it was hypothesized that
researchers who studied psychiatric and medical
samples were expected to report significantly
greater rate of research risks than the nonpsychiatric or medical samples. Research risk
was defined as (a) Incidence of confidentiality
violations for suicide, homicide, and abuse
status; (b) Incidence of participants’ condition
worsening; and (c) Incidence of complaints and
or suits filed against researcher or institution.
Method
We generated a list of 3,684 investigators who
received federal funding for research projects
pertaining to four at-risk groups. Specifically,
researchers who studied humans with
schizophrenia (n = 264), cardiovascular disease
(n = 1472), major affective disorder (n = 899),
and traumatic stress (n = 564) were identified
from relevant NIH institutes using the
Community of Science National Institute of
Health database of funded grants (http://
cos.gdb.org/best/fedfund/nih-select/inst.list.html)
and the Veterans Administration Medical Center
grant database (http://www.va.gov/research/
research.html). These groups were chosen to
represent medically and psychiatric samples that
are hypothesized to be at greater risk for
research-related injuries. In addition, we
identified a pool of 485 federally funded
investigators who study cognition in non-patient
samples to represent a group hypothesized to be a
relatively lower risk for research-related
research.
Relevant grant proposals were identified by
conducting a search of all proposals that had
titles which contained a relevant key word. For
example for studies on depression, depression
needed to be in the title. For traumatic stress
studies, PTSD, trauma or stress needed to be in
the title. A detailed listing of key words and the
systematic manner in which certain protocols
were eliminated is available from the first author.
Studies that crossed topic domains, used minors,
used animals, or were post-mortum human
studies were eliminated from the pool of studies.
All treatment studies were eliminated, since they
have unique risks and benefits that were not
assessed in this study. All projects that were
funded as multi-site collaborative studies were
also eliminated since it was assumed the ethical
considerations might vary across site.
Ultimately, 69 funded researchers who study
cognition, 79 who study schizophrenia, 61 who
study lung-cardiovascular disease, 56 who study
affective disorders, and 49 who study violence/
PTSD were contacted.
A cover letter, 7 page survey form1 , and
return envelope were sent to 314 researchers. A
reminder card was sent one month later to all
responders and non-responders. The survey
began with general information about the
respondent’s demographics, and research and
clinical experience. The researcher was asked to
complete the questionnaire in regard to the most
recent funded grant. Questions pertained to the
setting, sample, type of research, number of
sessions, participant characteristics, staff/training
and supervision. Then questions about informed
106
–––––––––––––––––––––––––––––––––––––––––––––––– Newman, et al., Ethical Research Practice with Human Participants
consent, confidentiality issues encountered,
participants’ reactions, emergency policies, and
injuries were attached.
Results
A total of 101 surveys were returned yielding a
32% response rate. Eleven surveys were dropped
from the analysis because they were post-mortem
studies (n = 4), used minors exclusively (n = 1),
focused on substance abuse, HIV, or personality
disorders (n = 4), animal studies (n = 1) or
couldn’t be classified into the groups based on
the responses (n = 1). Of the 9 researchers who
participated, 52.2% studied mental health (PTSD
n = 12, schizophrenia n = 16, major affective
disorders = 19), 24.4% studied cardiac or health
problems and 23.3% studied “normal” cognition.
Participants
The 90 principal investigators were comprised of
primarily Ph.D. trained researchers (73%) and
M.D.s (19%). There were more males (63%)
than females (37%) represented, and the majority
of respondents were Caucasian (94%). The
respondents’ experience with research ranged
from 2 to 49 years and had received a mean of
2.8 (SD = 1.8) federally funded grants in the 5
years prior to the study. The group of researchers
reported a mean of 70 peer-reviewed
publications, a median of 44 and a mode of 150.
Only 20% reported completing a course in
research ethics during advanced training.
Despite this lack of formal training, 73% felt that
they kept current with ethical issues and 50% felt
they kept current with legal issues in research.
Only 6% and 22% felt they were not current
regarding ethical and legal research issues,
respectively.
Research Procedures
Informed Consent Policy. With respect to
informed consent, the majority of the sample
(97%) provided written informed consent and
48% endorsed using methods to assess
participants’ comprehension of the consent form.
Of the 39 respondents who provided open ended
descriptions of these methods, 25 asked
participants if they had questions, 3 had the
interviewer certify person heard and understood,
3 used independent monitors, 2 relied on other
indicators (fluency, literacy, neurological status),
1 used family consent, 1 used structured consent,
2 asked the respondent to repeat questions, and 2
relied on signature to indicate comprehension.
Although 85% reported no need to investigate if
the identified participant could legally provide
consent, the remaining 15% reported a need
ranging from once (7%) to eighty-five times
(1%).
With respect to informed consent, 53% of
these researchers indicated that there were
instances in which the confidentiality of the
research participant might be broken. As
predicted, this policy differed by type of sample
group [x2 (2, n = 85) = 10.75 p =<.05], with 66%
of those who worked with mental health groups,
55% of those who worked with physical health
groups, and 21% of those who studied cognition
stating instances in which the research team
would consider breaking the confidentiality of
the research record. Among the group who
informed participants about confidentiality
issues, 55% reported communicating this in
specific rather than general terms.
Emergency Policy. Seventy-eight percent
(n = 61) of the researchers endorsed having a
protocol in place a priori to respond to
emergencies. The groups significantly differed
in this policy [x2(2, n =78) =32.15, p <.05] such
that 95% of mental health researchers, 90% of
physical health researchers, and 28% of cognitive
researchers reported such emergency policies in
place. Among the 47 who provided open ended
descriptions of these policies, 15 described use of
emergency on-call personnel, 8 cited they had
“written policies,” 6 used standard local
protocols, 6 cited immediately contacting the
project director or principal investigator, 5
trained staff in Cardio Pulmonary Resuscita tion
(CPR), and 3 discussed continuous monitoring
during research. The remaining four described
emergency medication, medical response plan in
lab and for evacuation, methods for handling
high blood pressure, and one general training
how to respond to a variety of situations.
Determination of Research-Related Risk.
Seventy-eight percent (n = 62) of the researchers
sampled reported keeping records regarding the
“frequency to which individuals experienced
negative and noticeable reactions.” Mental
health researchers reported significant greater
documentation than health or cognitive
researchers [x2 (2, n = 81) = 19.79, p < .05] such
that 88% of mental health researchers, 79% of
physical health researchers, and 52% of cognitive
researchers kept such records.
Debriefing Procedures. Sixty-four percent
(n = 57) of the researchers conducted debriefings
107
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Factors
Manipulation check
Educate participants
Check on participant
Express gratitude
Least important
24 (63%)
1 (2%)
7 (14%)
6 (11%)
Ranki ng
Important
5 (13%)
18 (33%)
12 (24%)
9 (16%)
Fairly Important
8 (21%)
7 (13%)
10 (20%)
26 (46%)
Most Important
1 (3%)
28 (52%)
21 (42%)
15 (27%)
Table 1. Number (and percentage) of participants ranking relative importance of 4 factors in planning debriefing.
after the research protocol. In fact, 70% of
mental health professionals, 42% of health
researchers, and 71% of cognitive researchers
used such debriefings [x2 (2, n = 80) = 5.06,
p =.08]. The majority (80%) of these
debriefings were conducted orally, although 6%
were conducted in writing, with 14% conducted
in both formats; there was no statistically
significant difference among the groups
regarding format [x2 (4, n = 51) = 4.48, p = .34].
The majority of these debriefings were done in
individual sessions (88%) rather than group
(4%), varied (6%) or family formats (2%); this
did not vary significantly among groups format
[x2 (6, n = 51) = 9.05, p = .17]. As can be seen on
Table 1, investigators felt debriefings were most
important for educating participants and checking
on participants. It is interesting to note that
manipulation checks were deemed least
important.
Use of Referrals. Forty-one researchers
(46% of the sample) responded to the item about
referral policy. Among those who responded,
20% reported providing referrals to all
participants, 12% to those participants who
indicated interest, 17% to only those in distress,
42% to those either interested or distressed, and
10% in “other” circumstances. Three researchers
described such other circumstances as “offered to
all deemed appropriate, but given to those
interested;” “two found to have physical
disorders,” and “all those screened with high
Suicidality
Homicide
Child abuse
Elder abuse
Abuse of the disabled
HIV status
Substance abuse
Criminality
Violence toward partner
Other
Never
58 (64%)
76 (91%)
72 (85%)
78 (94%)
78 (94%)
64 (77%)
49 (59%)
68 (83%)
67 (80%)
50 (94%)
blood pressure.”
Given this practice, the number of referrals
for non-emergencies ranged from 0 to 90 (mean
= 4.76, s.d. =13.02; mode =0). The mean
number of referrals for the mental health, health
and cognitive research teams were 8.56 (S.D. =
17.83), 2.29 (S. D. = 4.10) and .40 (S.D. =1.05)
respectively, but these differences did not meet
criteria for statistical significance [F (2, 65) =
2.9, p = .062].
With respect to actual practice regarding
referral for immediate hospitalization, 6
researchers recommended immediate referral for
a condition or concern, (with two researchers
recommending it once, and the rest experiencing
it twice, three times, four times and 10 times). It
is unknown if these referrals were based on
research-related injuries, or other conditions
uncovered during the protocol.
Follow-up procedures. Fifty-four percent
(n = 41) of the researchers reported follow-up
efforts to determine if participants experienced a
worsening of condition. These efforts
significantly differed across groups [x2 (2, n = 76)
= 14.35, p <.01] such that 67% of mental health
researchers, 55% of health researchers, and 8%
of cognitive researchers used such methods. In
terms of actual numbers, 24 researchers reported
conducting a follow-up at least once to check on
a participant.
Infrequently
20 (24%)
5 (6%)
9 (11%)
4 (5%)
4 (5%)
9 (11%)
10 (12%)
9 (11%)
11 (13%)
3 (6%)
Sometimes
4 (4%)
2 (2%)
2 (2%)
8 (10%)
14 (17%)
1(1%)
2 (2%)
Regularly
2 (2%)
2 (2%)
9 (11%)
3 (4%)
3 (4%)
Table 2. Number and (Percentage) of researchers who faced confidentiality issues.
108
Always
1 (1%)
1 (1%)
2 (2%)
1 (1%)
1 (1%)
1 (1%)
1 (1%)
1 (1%)
–––––––––––––––––––––––––––––––––––––––––––––––– Newman, et al., Ethical Research Practice with Human Participants
of research conducted [x2 (2, n = 73) = .42, p =
.81]
Incidence of complaints filed against a
researcher or institution. In this sample, 18%
reported infrequent complaints about research
staff’s conduct. Two percent (n =2) reported
complaints filed against the institution however
none resulted in legal proceedings. On the other
hand, 77% of researchers reported that
participants thanked them, with 33% reporting
this occurring sometimes, and 12% reporting this
as a regular occurrence.
Research Risks
Incidence of confidentiality violations. The
research staff occasionally faced confidentiality
dilemmas as shown in Table 2, with substance
abuse being the most frequently encountered
issue. However, only 8 researchers actually
broke confidentiality. Of these 8, 6 studied
mental health (n = 3 mood disorders, n = 2
schizophrenia, n =1 PTSD), 1 studied normal
cognition, and 1 studied health conditions.
Among those researchers who described the
specific circumstances, two reported needing to
hospitalize at least one participant against his/her
will, three reported having to file at least one
report to the authorities, and two reported
needing to warn at least one person in danger.
Incidence of participants condition
worsening. During the protocol a range of
emotional and physical experiences were
encountered (See Table 3); clearly crying
appeared most often. Although it was rare that a
participant became medically compromised, it
did occur. Twelve researchers (13%) reported at
least one research-related injury. Two researchers
reported that at least one participant had a
research-related infection. Five researchers
reported at least one case of temporary disability,
and none reported research-related death. It
should be noted that only 53% of researchers
reported knowing how many participants
experienced an immediate worsening of
condition (research related injuries) after
completing the research protocol; Knowledge of
research-related injuries was not related to type
Cried
Became hostile or
angry
Experienced Panic
Attacks
Expressed extreme
fear
Reported feeling
spacey
Became medically
compromised
Threatened the
research staff
Other
Discussion
In this preliminary study, 90 federally funded
researchers who work with human participants
responded to a survey about ethical research
practice. There seems to be a great variation in
ethical practice among distinguished researchers,
although all these research participants were
sensitive to research-related ethical dilemmas.
Policies
There is a great deal of variation in research
policy implementation. Although nearly all use
written informed consent, researchers varied in
the detail that they provide participants about the
limits of confidentiality. Although the majority
of researchers developed emergency policies and
debriefing procedures, the nature of these
procedures also varied. Although often required,
32% did not keep records of participants’
negative and noticeable reactions.
Approximately half the researchers reported
Never
35 (42%)
Infrequenty
24 (29%)
Sometimes
16 (19%)
Regularly
7 (8%)
Always
1 (1%)
33 (43%)
35 (42%)
13 (16%)
3 (2%)
0
59 (71%)
17 (21%)
6 (7%)
1 (1%)
0
55 (66%)
16 (20%)
8 (9%)
4 (5%)
0
51 (62%)
18 (22%)
12 (15%)
1 (1%)
0
66 (81%)
14 (17%)
2 (2%)
0
0
71 (87%)
10 (12%)
1 (1%)
0
0
33 (86%)
2 (5%)
1 (3%)
1 (3%)
1 (3%)
Table 3. Number and percentage of researchers who encountered participants’ emotional or physical response to research.
109
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
using follow-up methods to check on
participants’ condition. However, less than half
the sample responded to the item regarding the
use of referrals and those that did respond
indicated a range of practices with respect to
referring to other agencies. As anticipated,
researchers working with psychiatric illness
being more protective and explicit about policies
for emergencies, risk documentation, and followup procedures but not for debriefing.
Risks
With respect to research risk, a minority of
researchers reported having to deal with
confidentiality issues, worsening of conditions,
and complaints from participants. However,
emotional and physical symptoms were
encountered. In particular, 58% (n = 48)
experienced crying, and 12 researchers (13%)
reported temporary research-related injuries.
Given that several of these studies were about
health conditions, it is difficult to evaluate if
these reactions were elicited by research
participation, or were symptoms that individuals
experienced irrespective of research
participation. These reactions need to be
examined in future studies in the context of
baseline functioning of individuals to further
understand if they meet the requirements of
minimal risk. Nonetheless, the data are
consistent with claims that the physical hazards
of being a research participant are minimal even
among medical procedures (18). Although, these
risks appear minimal, they might be an
underestimate given that about half the
researchers did not document or know the
number of participants whose condition
worsened.
Finally, very few researchers received formal
training in research ethics although the majority
were confident that they were up to date in
ethics, and half felt prepared for legal challenges.
Given that researchers thought highly of their
respective IRBs, continuing education may be
best implemented through local IRBs.
There are several limitations to this study.
First sample bias and demand characteristics may
have affected the generalizability of these results.
Although the extensive comments written on
those returned surveys suggest that researchers
were interested in sharing their experiences,
sample bias may have affected the results.
Second, while this study reveals a diversity of
ethical practices, the quality of ethical
implementation is not examined. Hence it is not
known if this diversity suggests unsuccessful or
successful flexibility of methods in responding to
the needs of human participants.
Although the participation rate precludes
generalizing to all researchers, these preliminary
results provide information that can be useful in
designing training and compliance policy. In
particular, the diversity of responses suggests the
need for cross-training across subspecialties to
share perspectives. Individuals with risk factors
may not only present for studies of health and
mental health problems, so it can be helpful to
share approaches across specialties. For example,
although the majority of research-injuries were
identified among those mental heath studies, they
were not exclusively there. Furthermore it is
unclear, given the lack of documentation and
investigation, if this reflects better preparedness
of mental heath researchers or greater risk in
these studies. Future studies may be able to
better examine this by ongoing quality control
(19).
Acknowledgements
The research for this paper was supported by a
University of Tulsa Summer Faculty
Development Grant. A portion of this paper was
presented at Research Conference on Research
Integrity, Office of Research Integrity, Bethesda,
MD, November 2000.
Notes
1. A copy of the survey is available from the first author.
Bibliography
1. American Psychological Association. Ethical principles
of psychologists. Am Psychol 1981; 36: 633-8.
2. Protection of Human Subjects. 45 C. F. R. Subtitle A,
pt. 46. (2001).
3. National Commission for the Protection of Human
Subjects of Biomedical and Behavioral Research. The
Belmont Report: Ethical principles and guidelines for
the protection of human subjects of research (DHEW
Publication No OS 78-0012). Washington DC:
Department of Heath, Education, and Welfare; 1978.
4. Goldman J, Katz MD. Inconsistency and institutional
review boards. JAMA 1982; 248(2): 197-202.
5. Barton JM, Macmillan MS, Sawyer L. The
compensation of patients injured in clinical trials. J
Med Ethics 1995; 21(3): 166-9.
6. Burbach DJ, Farha JG, Thorpe, JS. Assessing depression
in community samples of children using self-report
inventories: Ethical considerations. J Abnorm Ch
Psychol 1986; 14(4): 579-89.
110
–––––––––––––––––––––––––––––––––––––––––––––––– Newman, et al., Ethical Research Practice with Human Participants
7. McCrady BS, Bux DA. Ethical issues in informed
consent with substance abusers. J Consult Clin Psychol
1999; 67: 186-93.
8. Stanton AL, New MJ. Ethical responsibilities to
depressed research participants. Prof Psychol 1988; 19:
279-85.
9. Silverman HJ. Ethical considerations of ensuring an
informed and autonomous consent in research involving
critically ill patients. Am J Respir Crit Care Med 1996;
154(3): 582-6.
10. Edgar H, Rothman DJ. The institutional review board
and beyond: Future challenges to the ethics of human
experimentation. Milbank Q 1995; 73: 489-96
11. Hilts PJ. Safety Concerns Halt Oklahoma Research.
The New York Times 2000 Jul 11; Sect. A:12.
12. Newman E, Kaloupek DG, Keane TM, Folstein S.
Ethical decision making among trauma researchers. In
Kantor GK, Jasinski J, editors. Out of the darkness:
contemporary perspectives on family violence.
Newbury Park (CA): Sage Press; 1997; 271-81.
13. Ceci, SJ, Peters D, Plotkin J. Human subjects review:
personal values, and the regulation of social science
research. Am Psychol 1985; 40(9): 994-1002.
14. Oldham JM, Haimowitz S, Delano SJ. Protection of
persons with mental disorders from research risk. Arch
Gen Psychiatry 1999; 56: 688-93.
15. Scarr S. Race and gender as psychological variables:
Social and ethical issues. Am Psychol 1988; 43: 56-9.
16. Sieber JE. Issues presented by mandatory reporting
requirements of researchers of child abuse and neglect.
Ethics Behav 1994; 4: 1-22.
17. Stanley B, Sieber JE, Melton GB. Empirical studies of
ethical issues in research. A research agenda. Am
Psychol 1987; 42(7): 735-41.
18. Levine M. Ethics and regulation of clinical research.
Baltimore: Urban & Schwarsenberg; 1981.
19. Newman E, Willard T, Sinclair R, Kaloupek D
(forthcoming). The costs and benefits of research from
the participants’view: the path to empirically informed
research practice. Accountability Resear 2001; 8: 2747.
111
Balancing Risks and Benefits of Deception in Assessing Genetic Screening*
Dennis L. Thombs, Department of Adult Counseling, Health and Vocational Education, Kent
State University, USA
Colleen A. Mahoney, Mahoney Consulting Group, Kent, Ohio USA
R. Scott Olds, Health Education and Promotion, Kent State University, USA
Keywords: Alcoholism, Deception, Genetic screening
The Human Genome Project is a massive international research program designed to map the human
genome sequence(1). The fundamental purpose of the program is to spur a transition to DNA
sequence-based biology and biomedical science(2). In addition to revolutionizing medical
diagnostics and therapy, the Human Genome Project will create new challenges in a variety of fields
including law, medical ethics, public health, and health services administration(3). The anticipation of
these changes does not represent a distant concern. A “working draft” of the entire human sequence is
expected by the end of 2001(2).
Against the backdrop of the Human Genome Project, this article critically examines the use of
intentional deception to assess (and anticipate) the utilization of genetic screening for alcoholism
susceptibility. For some time, the manipulation of study participants by deception has been controversial in experimental social psychology(4). This controversy has emerged in health behavior research
as a consequence of the remarkable progress made by the Human Genome Project. Little is known
about the public’s interest and utilization of clinical genetic testing(5). In the specific area of predictive genetic screening, a deception paradigm (described below) has been found useful for assessing
utilization. This paradigm helps estimate utilization when such tools are on the horizon, but not yet
available to the consumer. Intentional deception appears to be necessary because “hypothetical
testing,”(6, 7) honestly described to research subjects as available “sometime in the future,” generates
inflated interest compared to testing described as “currently available”(8, 9).
In an editorial that appeared in the Journal of American College Health,“Hard Questions About
Research Procedures: The Search for Authenticity”(10), Dr. Richard Keeling objected to the use of
deception in a quasi-experimental study conducted by the authors. The report of this investigation
appears in the same issue of that publication “Application of a Bogus Testing Procedure to Determine
College Students’ Utilization of Genetic Screening for Alcoholism”(11). Interested readers may turn
to that article for a full description of the study methods, including the fabricated story concocted to
test student interest in genetic screening for alcoholism susceptibility.
Dr. Keeling’s editorial is an example of a conservative, but perhaps increasingly common position
* Reprinted with permission from the Am J Health Behav 2001; 25(2): 100-105. R. Scott Olds was invited to present this
paper at the RRI conference under the title: Responsible Research Conduct that Balances Risks and Benefits of Deception in
Assessing Genetic Screening Utilization.
Corresponding author: R. Scott Olds, H.S.D., Health Education and Promotion, 316 White Hall, Kent State University,
Kent, OH 44242-0001, 330-672-0679 (voice), 330-672-3063 (fax) , [email protected]
Proceedings: Investigating Research Integrity (2001) –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
on human subjects protection that exaggerates
risk to study participants and discourages potentially valuable inquiry. The conservative position
is based on the following beliefs: 1) deception is
inherently harmful; and 2) deception research is
not carried out under realistic conditions and
therefore is not of value. The authors believe
their views are based on an ethic of measured and
reflective discourse, instead of a “knee-jerk”
response fashioned to serve a particular ideology.
According to Aronson and colleagues (4),
when considering the use of deception in research, investigators must weigh the psychological discomfort participants may experience
against the value of the study. There is no single
set of rules that can be applied to resolve this
dilemma, and reasonable professionals will arrive
at different judgments in this difficult analysis.
To determine college student interest in genetic
screening for alcoholism susceptibility, it was
reasonable to expose them to what was believed
to be modest psychological and social risks. The
Institutional Review Board at Kent State University concurred, and with certain stipulations gave
approval to conduct the study.
The subjects in this study were deceived
about the availability of a genetic screening test.
For up to seven days, 181 students thought they
could schedule a predictive screening test for
alcoholism that does not yet exist. The authors
did not believe that this lie harmed the students
in any substantial way. In broad-brush comments,
Dr. Keeling (10; see page 101 of his editorial)
claims that today’s college students are often
exploited by society and that any challenge to
their “search for authenticity” poses an unacceptable risk to their mental health and/or future
social functioning. It seems that this view is not
unusual in academia today. Such a position
represents “politically correct” discourse that
exaggerates the risks of deception in this study
and casts a broad net of condemnation over all
uses of deception in research. Clearly, humans
have been mistreated in research that employed
deception (e.g., the Tuskegee Syphilis Study), but
distinctions can and should be made in its
application.
In this era of heightened concern about
compliance with Federal regulations on research
involving human subjects, “minimal risks” in
behavioral science research have sometimes been
subtlety redefined as “unacceptable risks.” The
authors have no data to support or dispute such
speculation, but wonder whether the balancing of
risks and benefits has tilted toward the former in
recent years. If so, does this shift represent
increased concern for human subjects? An
iconoclastic interpretation is that the conservative
analysis of risk has been motivated by fears of
lawsuits and a desire to protect the university
from legal action. In addition, doubts about the
quality and usefulness of behavioral science
research in general, may be in operation in some
quarters which only further discourages full
consideration of the potential benefits of such
work.
No data were collected in this study to
support the claim that the students were not
harmed by the deception. However, it should be
noted that the empirical literature does not
support the view that research using deception is
any more harmful than non-deception research
(4). One review of the literature concluded that it
was rare for participants to feel that they had
been harmed by intentional deception (12).
Though empirical studies on the effects of
deception are few, those that have been conducted generally have found that participants
report greater enjoyment from having participated in a deception experiment than in a
nondeception experiment (13). This is probably
due to deception studies being less boring (4). To
address these concerns, in the future, investigators should follow up with participants to determine their reactions to research deceptions.
It is noted that the source of discomfort in
deception research is not only learning later that
one has been deceived, but equally, if not more
important is that the person often learns something painful about themselves or others (14).
Again, data were not collected to support this
hypothesis, but it is strongly suspected that
among those students who were uncomfortable in
this study, the primary source of their discomfort
was their current drinking behavior. As noted, the
sample was over-represented by heavy drinking
students. Participation in the study required them
to reflect on their own alcohol use as well as that
of their family members. Indeed, it was sensed
by the authors that some students were uncomfortable while responding to the questionnaire
and watching the presentation. In other words,
the discomfort that some experienced appeared to
occur before the debriefing, rather than after it
(when they learned they had been deceived).
Some students actually appeared amused during
the debriefings.
The level of discomfort experienced by
114
––––––––––––––––––––––––––––––––––––––––––––––– Thombs, et. al, Deception in Assessing Genetic Screening Utilization
students was probably comparable to being asked
to participate in an anonymous self-report survey
of alcohol use, and probably no greater than
sitting in routine lectures and discussions in
health education courses that deal with any
number of sensitive issues. The discomfort that
some may have experienced was not considered
to be detrimental or bad. Good health education
“shakes up” students by confronting biased
perceptions of risk and challenging existing
social norms. It also is consistent with the
traditional view of higher education, which is to
challenge conventional thinking and behavior
and to engage students in debate about controversial issues.
Dr. Keeling (10) also was critical of the
contention that the study conditions were “realistic.” The authors agree with his observation that
if (or when) genetic testing for alcoholism
susceptibility becomes available, protocols very
likely will require extensive patient counseling
before and after the procedure. So by this benchmark, the study’s procedure was not realistic. The
authors should have been more precise by stating
that “our method was more realistic than using a
procedure that described screening as a future
possibility.” However, at the same time, introducing extensive patient counseling into the study
procedure would have required us to employ a
far greater level of deception. Such a research
design would be considered unethical by virtually all professionals and would justify
Dr. Keeling’s response. This study protocol,
however, does not.
As the study was carried out, participants
were deceived for no more than seven days. They
were debriefed and offered the opportunity to
withdraw their data without penalty. In his
editorial, Dr. Keeling (10) stated,
. . . Having watched a computer-generated
presentation (for 7 minutes) and heard a brief
explanation of the study itself, they were then
required to state their intentions about being
tested immediately. There was little time for
them to ponder the issues and submit a formal
request to be tested. . .(p. 100).
This description of the study’s methods is not
accurate. Careful reading of the methods clearly
stated that students were told they did not have to
make a decision immediately after the
presentation. A questionnaire item allowed them
to respond I am uncertain about whether or not
to be tested (see p.106 of our article)(11).
Further, their participation was always voluntary
and invitational. They were able to cease
participation at any time without penalty.
Dr. Keeling was accurate in describing that over
the next seven days, students were not given
counsel or additional information about the test.
In this respect, the procedure was not as realistic
as future testing probably will be, but neither was
it as unrealistic as described by Dr. Keeling in his
editorial. It is acknowledged that in the future,
people may contemplate the testing decision for
extended periods of time, perhaps even many
years. Obviously, this study does not address
readiness to seek testing over extended time
intervals, but it does provide marketing
information about what to expect if promotion of
genetic screening for alcoholism susceptibility
among high-risk drinkers becomes a public
health goal.
The preliminary findings from this study
suggest that among college students, there may
be little enthusiasm for seeking alcoholism
screening if (or when) it becomes available.
Certainly this issue deserves further investigation. The authors believe the health promotion
profession has an obligation and responsibility to
conduct research that anticipates and informs the
development of sound public health policy. If
future public health policy supports genetic
testing for alcoholism susceptibility, ethical
questions need to be raised by the professions
concerned with public health. This study is part
of the foundation needed to address these questions.
These debates are important and healthy, but
they are not easy. The issues surrounding genetic
testing are complex. Billions of dollars are being
spent on genome research for the purpose of
developing effective technologies to treat and
prevent disease. Yet, relatively little attention is
being given to the behavioral, social, and health
service implications of this technology. There is a
need to better understand the utilization of
predictive screening for a variety of disorders,
including alcoholism. This study should stimulate discussion among health promotion professionals about these aspects of genetic testing.
Bibliography
1. National Human Genome Research Institute. The
Human Genome Project. www.nhgri.nih.gov, 2000;
April, 14.
2. Collins FS, Patrinos A, Jordan E, Chakravarti A,
Gesteland R, et al. New goals for the U.S. Human
Genome Project: 1998-2003. Science
1998;282:682-689.
115
Proceedings: Investigating Research Integrity (2001) –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
3. Khoury MJ and Genetics Work Group. From genes to
public health: The applications of genetic technology in
disease prevention. Am J Pub Hlth 1996;86:1717-1722.
4. Aronson E, Ellsworth PC, Carlsmith JM, Gonzales MH.
Methods of Social Psychological R esearch (Second
Edition). New York: McGraw-Hill; 1990.
5. Croyle RT, Lerman C. Interest in genetic testing for
colon cancer susceptibility: Cognitive and emotional
correlates. Prev Med 1995;22:284-292.
6. Kessler S, Field T, Worth L, Mosbarger H. Attitudes of
persons at risk for Huntington’s disease toward
predictive testing. Am J Hum Genet 1987;26:259-270.
7. Williamson R, Allison MED, Bentley TJ, Lim SMC,
Watson E, et al. Community attitudes to cystic fibrosis
carrier testing in England: A pilot study. Prenat Diag
1989;9:727-734.
8. Craufurd D, Dodge A, Kerzin-Storrar L, Harris R.
Uptake of presymptomatic predictive testing for
Huntington’s disease. Lancet 1989;2:603-605.
9. Tambor ES, Bernhardt BA, Chase GA, Fadden RR,
Geller G, et al. Offering cystic fibrosis carrier screening
to an HMO population: Factors associated with
utilization. Am J Hum Gen 1994;55:626-637.
10. Keeling RP. Hard questions about research
procedures:The search for authenticity. J Am Coll
Health 1998;47:99-102.
11. Thombs DL, Mahoney CA, Olds, RS. Application of a
bogus testing procedure to determine college students’
utilization of genetic screening for alcoholism. J Am
Coll Health 1998;47:103-112.
12. Baumrind D. Research using intentional deception:
Ethical issues revisited. Am Psychol 1985;40:165-174.
13. Christensen L. Deception in psychological research:
When is it justified? Personality and Social Psychology
Bulletin 1988;14:664-675.
14. Smith SS, Richardson D. Amelioration of deception and
harm in psychological research: The importance of
debriefing. Journal of Personality and Social
Psychology 1983;44:1075-1082.
116
Research Integrity and the Direct Involvement of Persons with Disabilities
Nancy B. Robinson, Communication, Arts and Sciences Department, California State
University, Chico, California, USA
Kathleen C. Sadao, Curriculum and Instruction Department, University of the Pacific,
Stockton, California, USA
Keywords: Responsible conduct of research, Preservice training, Problem-based learning, Persons with
disabilities, Affective learning
Teaching students in the health, human service and education professions to be responsible in their
interactions with persons with disabilities, as service providers and researchers, poses unique
challenges to educators to move beyond imparting knowledge to impacting attitudes, values and
ethics. Recent emphasis on outcomes of professional education programs most frequently focuses on
indices of cognitive achievement and performance of specific skills or competencies. Measures of
affective learning, or student attitudes and values toward the persons they serve, are less frequent and
more difficult to document. Universities need to educate professionals who are capable of the
responsible conduct of research. Pre-service education models are shifting from a traditional didactic
approach to the use of case studies and problem solving, in an effort to influence affective learning
and the application of knowledge and skills in real-life simulations. Studies of effective teaching
methods to prepare professionals in the area of responsible conduct of research with human subjects
are clearly needed. Person-focused learning approaches developed from interactive teaching models,
used increasingly in pre-service education in disability services and programs. The use of case studies
tends to promote application of theoretical knowledge and positive changes in affective learning, or
students’ attitudes and values.
Person-focused learning approaches move beyond case studies and directly include persons with
disabilities and family members as partners. Research and teaching-involving people with disabilities
assume that validity is strengthened through the direct involvement of people who experience
disability daily (1). Kvale and Burns discuss threats to validity and the need to reconceptualize
validity in qualitative research (2, 3). Due to the integral involvement of the researcher to conduct
qualitative research, Kvale argued that qualitative research requires attentiveness to the concept of
validity and its social construction with constant reference to the values, attitudes and experiences of
the researcher and participants (2). Further, qualitative research methodology applies to interactive
teaching, in which themes are explored and developed based on real-life scenarios (4). Participatory
action research, a qualitative research process, directly involves key stakeholders in all phases of
investigation (5, 1). In the present study, partnerships with persons with disabilities and family
members began and continued throughout the design, implementation, and evaluation of co-teaching
activities.
Corresponding author: Nancy B. Robinson, CSU Chico, SPPA Program, AJH 107B, Chico, CA 95929-0350, 530 898-6848
(voice), 530-898-6612 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
The goal of the present study is to
demonstrate and evaluate an interactive teaching
method that directly involves people with
disabilities and their family members and the
impact of this model on students’ attitudes and
values, or on affective learning. Although the
use of case study approaches in college level
teaching, particularly with persons with
disabilities, produces positive student learning
outcomes, the differences in approaches to the
uses of case studies are not explored.
Specifically, the researchers sought to examine
the effectiveness of person-focused learning to
promote the responsible conduct of research
among graduate, post-graduate and doctoral
students.
Three major developments in policy,
program development and teaching practices led
to the development of person-focused learning.
First, shifts in legislation and policy began in the
1950’s and 1960’s in the US, which continues
today with increasing emphasis and advocacy for
the rights of people with disabilities to have
equal access to all arenas of community life.
Second, increasing focus on rights and advocacy
for people with disabilities contributed to the
self-determination movement that places
decision-making and life choices with the people
affected, people with disabilities. Third, teaching
practices in higher education shifted from
traditional didactic models to interactive,
problem-solving models that strive to establish
critical thinking skills among students in
preprofessional training programs. The combined
influences of these broadly defined trends in
policy, program, and professional practice are
particularly relevant in higher education, where
the forming of future professionals’ values,
attitudes, knowledge, and skills are critical for
future practice and partnership with people with
disabilities.
Teaching methodology in professional
training programs is changing from a didactic
approach to an interactive model that requires
students to take responsibility for their own
learning (6). Medical education first developed
problem-based learning (PBL) to create a student
driven learning model. PBL was since adapted to
curricular content in several health, human
service, and education disciplines. Beginning
with PBL, four approaches to interactive and
problem-solving approaches to teaching are
briefly described in this paper. The strengths and
contributions of each model are addressed and
the person-focused learning model is highlighted
as the focus of this study and context for
participatory action research.
Problem-Based Learning. As stated above,
PBL began within medical education to increase
the application of medical theory and information
with specific patient case studies and has since
extended to nursing, occupational therapy, and
other fields (7-11). Cockrell, Hughes, Caplow,
and Donaldson described problem-based learning
as a “collaborative learning approach” (12).
Collaborative learning is premised on Vygotskian
concepts that define learning as the social
construction of knowledge. The cooperation and
shared resources that take place in PBL learning
reflect tasks in “real world” settings. These
authors outlined six basic phases in PBL:
(a) encounter with the problem; (b) free inquiry;
(c) identification of learning issues; (d) peer
teaching; (e) knowledge integration and
(f) problem resolution. Based on their
investigation of student’s perspectives of PBL,
Cockrell et al. found three key areas of student
perspectives of PBL: ownership, group
dynamics, and tutor feedback (12). Students
reported a deeper level of understanding and
retention in the PBL process compared to more
traditional teaching approaches and increased
awareness of team building skills. Students stated
a preference for tutors who were non-directive
and non-obtrusive. Students reported that the
benefits of collaborative learning included: a)
learning to become part of a learning community,
and b) learning to speak the language of the
community of professionals within the discipline.
Inquiry-based learning. Inquiry-based
learning (IBL) uses a case-study process to
encourage student responsibility for learning
outcomes. Inquiry-based learning is similar to
PBL in teaching methodology and includes
presentation of case studies and the application of
a problem-solving process that students use to
identify relevant issues that require further
research. However, rather than resolving the case
through a diagnosis, IBL focuses on the inquiry
process using issues that are relevant to the case
(13, 14). As in PBL, students take ownership
from the beginning, as in PBL and work in small,
tutorial groups guided by a faculty member. The
case is discussed and analyzed based on what
information is known, further information
needed, and the identification of learning issues
that require further research. The cases provide a
structure and format that guide students to
118
––––––––––––––––––––––––––––––––––––––––––––––– Robinson & Sadao, Research Integrity and Persons with Disabilities
explore potential solutions to posed problems.
Casebooks are now an accepted technique in
preservice teacher training programs (15). As is
indicated in PBL, the use of a case encourages
group work that inevitably models collaborative
communication skills found in the field. The
paper case leads learners to apply skills learned
to field projects (16). Students then conduct
independent research and at a later session,
present the results of their research that
originated from the initial case study. Faculty
members with the focus on critical analysis of
relevant policy, program, advocacy, financial,
cultural, facilitate summary and wrap-up
discussion and community issues related to the
case.
Family-focused learning. Family-focused
learning (FFL) formed in the context of
interdisciplinary education for health
professionals to provide a model of direct
involvement of family members in the teaching
process (17). Family-focused learning follows
the inquiry based approach through a series of
sessions that begin with identification of issues
around a particular family with an individual
member with a disability, and close with student
presentation of research issues related to the
particular family that is participating in the
teaching and learning process. The key
difference in the FFL, compared to the previous
models described, is that actual families and
people with disabilities participate in the teaching
process with faculty, interact with faculty and
students throughout the development of case
information to be presented and provide
supportive critique to students in their work.
Similar to PBL and IBL, the FFL model requires
an initial session to present concerns and
information that guide student inquiry. In
contrast to the other two models, FFL involves
actual family members who present the “family
story” to students through video and written
media. The development of the video is a joint
venture for the family and participating faculty
members that can require two or more sessions.
When the family is satisfied with the video
presentation, the tape is shared with students of
several health, human services and education
disciplines that identify key issues in a problemsolving process similar to the two models already
described. Following completion of independent
research, students prepare issue papers and
present them to the family and/or individual for
critique in a closing session. Family members
and individuals with disabilities attend the
closing session for the purpose of providing
feedback to students on the scope of their work,
relevance to their particular case, and quality in
addressing the particular issue selected. As in the
IBL closing session, faculty assist students in
summarizing their analyses their individual
research and relate students’ findings to broad
issues affecting families and persons with
disabilities.
Person-focused learning. Person-focused
learning (PFL) incorporates teaching and
learning methods included in the previous
models, but builds on elements found in each
preceding approach. The elements of problemsolving and critical thinking that are hallmarks of
PBL and IBL approaches are also essential to
person-focused approaches. As in the FFL
model, person-focused learning is designed and
implemented with the participation of families
and persons with disabilities. A new element is
the service-learning aspect of PFL. In the PFL
approach, students are required to complete a
project that responds to needs and concerns
identified by the family or individual (18). The
involvement of persons with disabilities,
families, faculty, and students in the development
and implementation of the teaching experience
produces a qualitative shift in teaching
methodology and creates an action research
model (4, 19-21). In the case-study approach,
students respond to the issues presented for the
primary purpose of advancing their own learning.
In the person-focused model, students are placed
in an interactive relationship with family
members and individuals from the outset of the
experience. The student learning goals, from the
faculty perspective, involve: a) application of
theoretical knowledge with real families and
individuals with disabilities; and b) development
of resources that respond to the needs expressed
by families and individuals.
In the current study, the authors were
concerned with the qualitative impacts of the
PFL model on the people involved: students,
families, and persons with disabilities. The
unique features of the PFL model which
incorporate problem solving in a real-life context
and service to families and individuals require
systematic evaluation. The assumption that direct
involvement of actual family members and
people with disabilities increases validity and
thus applicability of the teaching process
required empirical investigation and
119
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
IHE
CSU Chico, CA
UOP, Stockton, CA
Speech Pathology:
AAC
Upper Division &
Graduate
Special Education:
Methods
Upper Division/
Graduate
Dept.
Speech Pathology
Special Education
Students
18 students
40 students
Course
Level
Disability Studies
University of Hawaii
Disability Studies: Team
Work
Upper Division &
Graduate
Interdisciplinary
Disability Studies
13 students
Table 1. Student participants in Person-Focused Learning at three universities.
consideration of the ethics involved. In this
study, the authors sought to systematically
evaluate the reciprocal impact of interactive
teaching on student learning outcomes and
people with disabilities, specifically with people
with disabilities in direct interaction with
students for the duration of semester-long
courses.
The foci of investigation centered on three
questions:
1.
What are student perceptions of the PFL
process, both in the process of interacting
with families and individuals and in learning
outcomes?
2.
What are family member and individual
perspectives of the PFL process, regarding
their partnership role in teaching students
and project outcomes?
3.
What are ethical and logistical considerations for the replication of PFL in human
service training programs, particularly
related to disabilities?
Methods
The study was completed in the context of three
interdisciplinary courses at three different
university sites, with 71 students and 7 families
including persons with disabilities. While course
content differed across the three sites, teaching
methods were similar. Teaching partnerships
used principles of “Family Centered Care,” in
which family concerns drive professional
interventions (22, 14, 23). Key steps in the
teaching partnership included: (a) determination
of family priorities; (b) adaptations to meet
family and individual needs; (c) family input in
project development; and (d) evaluation of
completed projects by family members and
persons with disabilities. Student learning
outcomes were evaluated with qualitative surveys
completed independently. Family and individual
outcomes were identified through semistructured interviews completed with the
investigator.
The courses that provided the context for the
study included a core special education course,
an elective course in augmentative and
alternative communication (AAC), and
interdisciplinary teamwork course. Family
members and individuals with disabilities
participated as teaching partners with faculty
members. Courses were located at California
State University, Chico; the University of the
Pacific in Stockton, California; and the
University of Hawaii. Students who participated
in the courses included three groups, shown in
Table 1.
Characteristics of the seven individuals and
families who participated in the study are listed
below:
•Three adults, three children
•Communication disorders and physical
disabilities in all subjects
•Two individuals with Asian/Pacific Islander
ethnicity
•Five individuals were Caucasian
Course content and learning objectives differed
across the three sites. However, key variables
were held constant in teaching methodology. All
courses included persons with disabilities and/or
family members who participated in the design
and implementation of the curriculum. The major
requirement in each course included direct
interaction with persons with disabilities and
family members in the design and development
of adaptive equipment or technology to meet
needs identified by the individual and family.
Students engaged in a common process that
included identification of needs by persons with
disabilities and/or family members adapted from
participatory action research (5, 1). Eight steps
were completed in the person-focused learning
teaching process. First, faculty developed
120
––––––––––––––––––––––––––––––––––––––––––––––– Robinson & Sadao, Research Integrity and Persons with Disabilities
curriculum information about individuals in
partnership with identified families and persons
with disabilities. Second, students reviewed
available information about the family and/or
individual determine an initial developmental or
environmental concerns identified by the family
and/or individual. Third, student groups
conducted brainstorming regarding potential
family and individual concerns. Fourth, students
prepared interviews based on guidelines provided
by faculty. Fifth, students conducted interviews
with individuals and/or family members. Sixth,
the working group met to identify adaptation or
support project based on results of prior
information and interviews with individual and
family members. Seventh, student groups
presented completed projects to individuals and
family members. Finally, student evaluations of
the process and projects were completed.
The qualitative effectiveness of the personfocused learning process was evaluated by:
(a) student perceptions of learning outcomes; and
(b) perceptions of family members and persons
with disabilities. Methods of evaluation included
student’s self reports and family/individual
interviews.
Self-Report. Students were requested to
complete qualitative comments in response to
questions designed by the investigators.
Questions addressed students’ perceptions of the
learning process and outcomes related to direct
interaction with family members and persons
with disabilities.
Family/Individual Feedback. Individuals
with disabilities and family members were asked
to evaluate their participation in the courses in a
teaching/consultant role. Perceptions of these
participants were also requested regarding the
quality of student projects and interaction with
family members and persons with disabilities.
As the focus of teaching included adaptations and
assistive technology, participants were requested
to evaluate benefits and changes related to
adaptations or resources developed by students.
investigative and problem-based learning in
direct interaction with people with disabilities
and family members. Analysis of student
surveys identified seven themes: (a) attitudinal
change; (b) authentic engagement; (c) critical
thinking; (d) sensitivity to families and
individuals; (e) collaborative teamwork;
(f) preparation for inclusion; and (g) self–
efficacy/skills to adapt materials. Examples of
student comments are included below related to
each theme:
Results and Discussion
Results of the study are discussed in relationship
to perceptions of student learning outcomes and
impacts on family members and persons with
disabilities.
Student Problem-Solving. Student responses
to qualitative questions were analyzed to
determine recurring themes related to
Attitudinal Change.
“There are many things that disabled students
are able to do…most important to focus on
those strengths.” 18c
“I realized how many aspects of a person’s life
can be affected by a disability.” 18c
“It made me realize how difficult it must be to
have a child with a disability, or to be a child
with a disability; everyday actions are so
difficult!” 19c
“I find myself constantly looking at isles in
stores, toys, elevators, etc. to see how they
could possibly be adapted to better suit the
needs of children with disabilities—more
awareness.” 7c
“I think it helped me look at adapting
equipment as a fun responsibility instead of a
required duty.” 8c
“It has helped me to realize that children with
disabilities have a vast amount of needs, and
that each child’s needs are unique. Adapted
equipment may still need further adaptations
to meet a specific child’s needs.” 10c
Authentic Engagement.
“The hands-on work helped me to develop a
better understanding of a family’s needs and
wishes for their children. Though most of
all…learning the true-to-life reality of the
processes involved in working with a family.”
12c
“Actually making the adaptations brings more
involvement and thus more interest, which lead
to more learning.” 12c
“I think with the case study, it is each to
maintain the same frame of reference and not
to expand on ideas or think about new things.
With the adapted equipment, new ideas or
problems are presented and we brainstormed.”
10c
Critical Thinking.
“This assignment makes you think about
aspects of disabilities that normally one
wouldn’t consider.” 2c
“We had discussed the written assignment a lot,
121
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
even before we knew what the questions were.
We were always thinking, how it would help
B.” 6c
individual.” 15c
Self-Efficacy and Adaptive skills.
“The most important part of this assignment
was that it opened a door for me and pretty
much told me that I had the potential to help
any child with a disability.” 3c
“I learned that I take my skills and abilities for
granted. From meeting B., I realized that many
aspects of daily living would be difficult for
her, and in order for them to function at her
level, more things would need to be adapted.”
10c
“Yes, because it provides hands on time that I
will remember more than any case study. It is
also more fun than any case study.” 9c
“I liked the developmental framework and the
way this was all set up. It was very realistic to
what we deal with in our real jobs and it was
very hands on.” 20c
“It makes me become more aware of the types
of things; a lot of things that I would have never
thought of.” 13c
Sensitivity to Families and Individuals.
“Meeting in an informal setting allows both
sides of the team to get to know each other with
out the pressure of a meeting…with the family
relaxed we can start building relationships.”
16c
“Getting to know the family was an important
milestone for us.” 16c
“It has made me realize that the parents are very
important in identifying the a child’s needs.”
16c
“I thought it was very useful to ask T. [the
parent] our questions because we had to know
exactly what her situation was so the outcome
would be helpful.” 5c
Collaborative Teamwork.
“Yes, because we need each other’s specialized
skills along with knowledge and creativity.”
14c
“It was a great idea to work in a group because
everyone has different ideas which we can
bring together. Then everyone has different
talents which were utilized in the production
process.” 12c
Preparation for Inclusion.
“This is something I will have to do in my
classroom so I appreciate the preparation.” 2c
“To find different ways to teach someone the
ABCs and how slow the song needs to be so
that the child can learn.” 9c
“It has made me realize that each child with a
disability is an individual; helping each child
can be done only if that child is looked at as an
Family and individual interviews revealed four
themes: (a) interaction with students; (b) selfvalidation; (c) support networks; and
(d) alternatives to meet individual needs.
Families and individuals commented that they
would participate again. Table 2, below
demonstrates representative feedback provided
by family members and person with disabilities.
Ethical issues identified included the need to
(a) respect individual choice in participation;
(b) confidentiality; (c) honor individual priorities
and (d) respect family differences. Comments
provided by families and individuals at the
completion of each class indicated the possibility
Theme Identified
Family Comments
“Having students come to our home was a highlight of the week for J., he
Interaction with students
looked forward to it all week.”
“Students gave S. attention and made us appreciate his importance.”
“I am getting braver to ask fo r what my son needs.”
“I always knew that J. knows more and the students helped to d ocument
that.”
“It is wonderful for our whole family to participate with the
students…going to the beach together was a first for us.”
“All of the time and support has given S. a chance to get out more.”
“The help provided by the class gave S. a way to communicate that he did
not have before.”
Self-validation
Support networks
Alternatives to meet needs
“We want S. to learn with the other kids and he shows the book to every
one who comes over.”
Table 2. Qualitative themes and family comments regarding Person-Focused Learning Outcomes.
122
––––––––––––––––––––––––––––––––––––––––––––––– Robinson & Sadao, Research Integrity and Persons with Disabilities
of initial reluctance to participate. One parent
commented that she initially was nervous when
meeting the students for the first time,
particularly due to cultural differences between
them. However, this parent later reported that
her feelings changed later after realizing how
much attention and support the students
demonstrated toward her son. This mother’s
comment highlights the need to honor individual
family priorities that may be based on cultural
styles, educational background, language
differences, and other variables. Related to this is
the need to respect and understand family
differences and follow the lead of the individual
or family to determine the most appropriate time
and place to conduct interviews and project
activities.
The results revealed positive qualitative
student learning outcomes. People with
disabilities and family members reported that
their participation provided important benefits
that included perceptions of increased selfefficacy and competence when interacting with
students. Risks were not specifically identified
by families or persons with disabilities, but
inferred from their feedback. The responsibility
to consider risk, which may include risks to
privacy of participants, remains with the
researcher who embarks on teaching partnerships
with families and persons with disabilities.
Comments provided by students in all thematic
areas reported revealed increased awareness and
respect for the life experiences of persons with
disabilities and family members, thus
establishing a foundation for ethical behavior in
future professional roles with persons with
disabilities, including teaching, service, and
research.
Summary
The results of the present study support the
effectiveness of interactive teaching, specifically
Person-Focused Learning, to promote student
learning outcomes that demonstrate respectful
and responsible professional attitudes and
behavior with persons with disabilities and
family members. The specific student learning
outcomes were found in both cognitive and
affective domains, as seen in students’
evaluations of the learning experience. These
findings have implications for preservice training
of health, human service, and education
professionals to establish a foundation for ethical
behavior with human subjects in the career
contexts of service and research.
The qualitative evaluation results of student
learning outcomes indicate that involvement of
persons with disabilities in the teaching process
provides authentic learning that cannot be
replicated with more traditional didactic
methods. Further, involving family members in
the teaching and evaluation process at all levels
follows a participatory action research process
and allows “checkpoints” for subjects to be fully
cognizant of the research agenda and purposes.
Thirdly, including people with disabilities in the
research/teaching process strengthens validity as
recommended by Kvale and Burns (2, 3).
Further, reciprocity in the learning setting is
achieved where students learn the needs of
families and the value their knowledge when
designing materials and technologies to assist
them in the learning environment. The research
participants are valued by the researchers and the
students involved in the assignment and the
student-made products are valued by the families.
The demonstration of a pre-service training
approach that teaches reciprocal relationships
with subjects is perhaps the key finding with
implications for training future professionals in
the area of responsible conduct of research. Not
only did students demonstrate qualitative
evidence of critical thinking in the learning
process, the direct interaction with subjects in the
action research model employed in PersonFocused Learning showed an effect on the
students’ sensitivity toward persons with
disabilities and family members. The
demonstrated effect on students’ sensitivity with
subjects could effect future professional ethics
and conduct. While, further study is needed to
determine attitudes and values that are directly
related to the responsible conduct of research
with human subjects, student attitudes toward
subjects are considered a critical variable of
ethical behavior. The question of what particular
teaching model effectively trains professionals
who are prepared to implement responsible
conduct of research was only partially addressed
by the present study. The attitudes and skills
required for responsible conduct of research are
clearly a constellation of knowledge and ethics
that require further explication.
This qualitative study explored personfocused learning principles in several preservice
courses and revealed positive findings for
students and the families who shared their
123
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
stories. The “realness” of the learning setting
allowed researchers to identify multiple learning
outcomes and ethical issues when involving
people with disabilities in a teaching setting and
research endeavor. Bowen identified the need to
strengthen internal validity through the
integration of qualitative and quantitative
research methodology (24). Further research in
PFL is needed to a) specify affective and
cognitive student learning outcomes; b) quantify
changes in student attitudes; b) compare PFL
teaching to other problem-solving approaches;
c) identify long range impacts on student
learning; d) develop guidelines for replication;
and e) explore the use of PFL to teach
responsible conduct of research. The
philosophical attitude and the research model in
the present study provide a framework for
preservice education and further research to
determine specific professional attributes that
lead to affective, cognitive, and ethical
foundations for the responsible conduct of
research, particularly with persons with
disabilities.
Bibliography
1. Guerrero, JL. Research paradigm shift: Participatory
action research. Working Paper No. June 1995.
Rehabilitation Research and Training Center of the
Pacific. San Diego State University, Interwork Institute.
2. Kvale, S. The social construction of validity. Qualitative
Inquiry 1995; 1 (1): 19-40.
3. Burns, PA. Securing internal validity: How to avoid the
threats [serial online] 1997 [cited 2000 Nov] Available
from: URL:http://trochim.human.cornell.edu/tutorial/
burns/int.html
4. Newman, JM. Action research: A brief overview [serial
online] 2000 Jan [cited 2000 Nov] Forum Qualitative
Sozialforschung / Forum: Qualitative Social Research
[On-line Journal]. Available from: URL: http://
qualitative-research.net/fqs.
5. Whyte, WF. Participatory action research. Newbury
Park (CA): Sage Publications; 1991.
6. Casey, MB, & Hwoson, P. Educating preservice
students based on problem-centered approach to
teaching. Journal of Teacher Education 1993; 44: 361369.
7. Chabon, S. & Wilkerson, DL. Discussions on diversity:
A problem-based instructional model to facilitate
cultural competence. Poster presented at American
Speech, Language and Hearing Association Annual
Conference, Washington, D.C., November 16, 2000.
8. Branda, LA. Implementing problem-based learning.
Journal of Dental Education 1990; 54: 548-549.
9. Bridges, EM. & Hallinger, P. Problem-based learning in
leadership development. 1995; Portland: University of
Oregon. ERIC Clearinghouse on Educational
Management.
10. Gist, GL. Problem-based learning: A new tool for
environmental health education. Journal of
Environmental Health 1992; 54: 8-13.
11. Royeen, C.B. A problem-based learning curriculum for
occupational therapy education. Journal of
Occupational Therapy 1995; 49: 338-346.
12. Cockrell, K.S.; Hughes Caplow, J.A.; Donaldson, J.F. A
context for learning: Collaborative groups in the
problem-based learning environment. The Review of
Higher Education 2000; 23: 347-363.
13. Geln, S. Towards a new model of nursing education.
Nurse Education Today 1995; 15: 90-95.
14. Matteoli, RF.; Verhey, MP. & Warda, MR. Collaboration
between families and university faculty: A partnership in
education. Journal of Psychosocial Nursing 1994; 32:
17-20.
15. Shulman, J., Lotan, RA., & Whitcomb, JA. Groupwork
in diverse classrooms: A casebook for educators. New
York: Teachers College Press; 1998.
16. Boyle, JR. & Danforth, S. Cases in Special Education.
Boston: McGraw-Hill. 2001.
17. Ratliffe, K., Stodden, N., Guinan, M., & Robinson, N.R.
Family partnerships in teaching: family-focused
learning. Journal of intellectual disability research 2000;
44 (3 & 4): 416.
18. Stevens, L. Academic Service Learning within an AAC
Curriculum. 2000 Poster presented at American Speech,
Language and Hearing Association Annual Conference,
Washington, D.C., November 16, 2000.
19. Henry, SK., Scott, JA., Skobel, B., Jones, A., Cross, S.,
Butler, C., & Blackstone, T. Linking university and
teacher communities: A Think Tank Model of
professional development. Teacher Education and
Special Education 1999; 22: 251-268.
20. Hutchinson, NL. & Martin, AK. Fostering inclusive
beliefs and practices during preservice teacher education
through communities of practice. Teacher Education
and Special Education 1999; 22: 234-250.
21. Perry, NE., Walton, C. & Calder, K. Teachers
developing assessments of early literacy: A community
of practice project. Teacher Education and Special
Education 1999; 22: 218-233.
22. Harry, B., Rueda, R., Kalyanpur, M. Cultural
reciprocity in sociocultural perspective: Adapting the
normalization principle for family collaboration.
Exceptional Children 1999; 66:123-136.
23. Gartner, A., Lipsky, DK., Turnbull, A. Supporting
families with a child with a disability. Baltimore: Paul
H. Brookes Publishing; 1991.
24. Bowen, K. The Sin of Omission -Punishable by Death to
Internal Validity: An Argument for Integration of
Qualitative and Quantitative Research Methods to
Strengthen Internal Validity[serial online] 1996 [cited
2000 Nov]. Available from: URL: http://
trochim.human.cornell.edu/gallery/bowen/hss691.htm.
124
4. Conflict of Interest
What is Driving Policies on Faculty Conflict of Interest? Considerations for
Policy Development
Mildred K. Cho, Center for Biomedical Ethics, Stanford University, USA
Ryo Shohara, Institute for Health Policy Studies, University of California, San Francisco,
USA
Drummond Rennie, Institute for Health Policy Studies, University of California, San
Francisco, USA
Keywords: Academic policies, Academic-industry ties, Conflict of interest, Research integrity
There are several factors driving policies on conflict of interest of faculty at academic research
institutions in the United States today. The first is that researchers and institutions have a greater
number, and a wider variety of financial conflicts of interest, especially in the area of biomedical
research. Sometimes, these financial interests appear to lead to very bad outcomes, and when that
happens, public scrutiny of the financial interests increases. Sometimes, this leads to new policy.
What is the current state of academic-industry ties in biomedical research? In 2000, the NIH’s
budget is $17.8 billion (1), while the pharmaceutical industry’s R&D budget is $22.4 million (2).
Krimsky found that 34% of research articles published in the top 14 biomedical research journals in
1992 had undisclosed financial ties of a lead author. These ties included holding a patent on an
invention related to the published research, or being on an advisory board or a major shareholder in a
company whose activites were related to the published research (3). In a review of FDA records,
USA Today reported that 54% of the time, experts hired by the FDA to advise on safety and
effectiveness of drugs have a direct financial interest in the drug or topic they are asked to evaluate
(4). Therefore, academic-industry ties are now the norm, rather than the exception.
Academic-industry ties have been the apparent cause of bad outcomes, including censorship of
data (5, 6), publication bias (7-10), lower quality of research (11), and harm to research subjects,
including death (12). Although it is impossible to determine a causal link between financial interest
and adverse outcome in individual situations, systematically gathered evidence suggests that, in the
aggregate, academic-industry ties can have adverse effects on the scientific process and outcome in
the aggregate (13).
One bad outcome in particular has led recently to public scrutiny and re-examination of policies
on conflicts of interest — the death of Jesse Gelsinger, who was a research subject in a Phase I
clinical trial of gene transfer at the University of Pennsylvania (12). Much attention focused on the
financial ties of investigators and the investigators’ institution with a company that was, in part,
sponsoring the trial. Although, again, it is impossible to prove that there was a causal link between
the financial ties and the death of Mr. Gelsinger, it was a link that was inevitably made, time and
Corresponding author: Mildred K. Cho, PhD, Stanford Center for Biomedical Ethics, 701A Welch Road, Suite 1105, Palo
Alto, CA 94304, 650-725-7993 (voice), 650-725-6131 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
again. A quote from a recent newspaper article
sums up the public perception:
Paul Gelsinger, Jesse’s father, said yesterday
he had undergone a painful change of heart in
the year after his son’s death, at first fully
trusting the researchers and holding them
blameless and then gradually, as disclosures of
apparent wrongdoing emerged, concluding that
he had been duped by scientists who cared more
about profits than safety. (14)
After Mr. Gelsinger’s death, the National
Institutes of Health (NIH) held a public meeting
this year to re-examine some aspects of conflict
of interest policy, and several professional
organizations, including the National Academy
of Sciences, the American Association of
Medical Colleges (AAMC), and Association of
Academic Health Centers (AHC), the American
Association of Universities (AAU), and the
American Association of University Professors
have all assembled internal groups to do the
same.
What are the current policies on faculty
conflict of interest?
Current policies on faculty conflict of interest
exist at several levels, including federal, state,
institutional regulations, editorial policies at
research journals, and statements by professional
societies. All are limited, however, in different
ways. The most widespread federal rules include
the “Objectivity in Research” regulations (15).
These are applicable only to researchers who
apply for research funding from the National
Science Foundation and the Public Health
Service (PHS), which includes the NIH. These
regulations are limited to disclosure of financial
ties that could be construed to affect the publiclyfunded research, and to financial ties that exceed
$10,000 annually or 5% equity interest. Thus,
financial ties in the context of industry-funded
research, where more serious conflicts of interest
might be found, are not covered under these
regulations.
In addition to federal regulations, there are
state laws that might apply to faculty at public
institutions. For example, some states prohibit or
require full disclosure of gifts to public
employees, which include faculty of state
universities. These state laws often do not apply
to private universities, and are not uniform from
state to state.
Institutional policies are mandated by the
federal regulations, which require that
institutions whose faculty apply for PHS or NSF
funding develop and implement their own written
rules for faculty conflicts of interest. These
institutional policies must conform to, but need
not be limited to, federal regulations. Indeed, the
majority of institutional policies go beyond
federal regulations in scope and management of
conflicts of interest, but most do not state
specific limits on financial interests, even when
in conjunction with company-sponsored research
(16). Most of these policies imply or state that
conflicts of interest are dealt with on a case-bycase basis, and seem to rely heavily on disclosure
as a primary mechanism for dealing with conflict
of interest.
Some research journals have developed
policies that require disclosure of authors’
financial interests to editors and reviewers .
However, such disclosures often do not surface
on the pages of the published articles, so their
effects are limited (Krimsky, this volume).
The AAMC, AHC, and the AAU created
guidelines for faculty conflict of interest long ago
(17-19), and although they thoughtfully outline
policy considerations, they are not specific and
are not enforced. Finally, in the wake of Jesse
Gelsinger’s death, two professional societies (the
American Society of Gene Therapy and the
American Society of Human Genetics) have put
forward statements that faculty having financial
interests in companies sponsoring their gene
transfer research is inappropriate and should be
avoided (20, 21). These statements only apply to
gene transfer research, however, and also have no
enforcement power.
What should we do about conflicts of
interest?
The answer to the question, “what do we do
about conflicts of interest?” depends upon the
answers to the questions, “what is conflict of
interest?”, “what is the primary interest of
academic institutions and the government?”, and
“what are the secondary interests we are
concerned about?”
What is conflict of interest? Opinions are
diverse. Many make the distinction between
“actual” and “potential” conflicts of interest.
Others call it scientific misconduct (22).
Depending on how one defines conflict of
interest, one may be led to base policy on
evidence of bad outcomes or on ethical or
professional values. We define conflict of
128
–––––––––––––––––––––––––––––––––––––––––––––––– Cho, et al., What is Driving Policies on Faculty Conflict of Interest?
interest as the co-existence of a primary interest
or duty (such as research integrity, patient
welfare, or education) and a secondary interest
(such as financial gain or recognition) (23). The
policy concern is that the secondary interest
exerts undue influence on the judgements made
in the course of executing the primary interest,
leading to adverse outcomes (such as research
bias or adverse effects on research subjects).
It is important to remember that conflict of
interest rules are activated in the absence of a
“crime” (24). Stark likens them to speed limit
laws. In contrast to laws against murder, which
are aimed at activities that, in themselves, are
deemed immoral and are not in the public
interest, speed limit laws are aimed against
conditions that predispose to the activities that
are not in the public interest. So, while driving at
70 miles per hour may not in itself be wrong in
the way that murder is wrong, high-speed driving
may enhance the chances of causing harm to
others. Some drivers might be quite capable of
avoiding crashes at even 200 miles per hour, but
because it would be difficult and impractical to
determine who they are and whether they are so
capable under all circumstances, the laws are
aimed at preventing the situation rather than
particular outcomes. However, there may be
certain speeds that would be considered
“reckless” in almost any circumstances, and thus
immoral – and there may be analogous financial
interests.
However, there is an important difference
between speed limit laws and conflict of interest
regulations, in that speed limit laws apply to
everyone, whereas conflict of interest laws apply
to groups that have a fiduciary relationship to the
public, such as public officials or professionals.
This distinction is important, because it means
that there are reasons to set the rules by criteria
other than probability of harm to the public,
namely in order to earn or preserve the right to
occupy the special position in society (25).
This definition of conflict of interest implies
that there can be no distinction made between
“actual” and “potential”. The conflicting
interests simply either exist or they do not. They
are, in themselves, not scientific misconduct,
although they may lead to misconduct. The
current definition of scientific misconduct carries
with it the notion of wrongdoing with intent (26),
which is based on the proven existence of a bad
outcome, and is therefore incompatible with a
definition of conflict of interest that is based on
the characteristics of a situation rather than the
outcome.
What is the primary interest? Lack of clarity
about the primary interests of researchers and
their institutions will lead to bad policy, because
one of the points of having the policies is to
protect the primary interests. So, the question is,
what are the roles of academic institutions and
the government in the conduct of science? The
passage of the Bayh-Dole Act gave the
government a new role in academic research,
namely, “to promote the marketing of inventions
developed under federally supported research and
development projects by nonprofit organizations
and small business firms.” (27)
Government specifically encouraged
academic institutions to be involved in the
marketing of inventions. Universities have taken
this encouragement to heart, “… shifting from
ivory tower to revving economic engine.” (28)
The new role of universities as economic engines
leads to expectations that they create jobs and
even whole industries. In fact, the government
has implicitly adopted the values of the business
world, where money is an incentive for
employees to work in the interests of
shareholders. In this model, the secondary
(financial) interest is considered to be in
alignment with the primary interest, rather than
acting as a competing interest. By contrast, the
model of professionalism says that the BayhDole Act and related legislation specifically put
not only faculty but institutions in a position of
conflict of interest. If academic institutions and
their faculty are expected to add economic goals
to their primary missions, can those institutions
be expected to be effective at developing and
enforcing conflict of interest rules for their
faculty? This seems to be a dangerous thing to
ask.
We must be clear about whether academic
institutions should take on economic health as a
primary interest. We must also be clear about
whether we are concerned only with or more
concerned about certain kinds of primary
interests. For example, is only federally-funded
research of concern, or all research? That is,
should policies be directed only at interests that
conflict with government-funded research, or
should they also be directed at interests that
conflict with industry-funded activities, too?
Finally, we should also ask whether clinical
research is of more concern than other research.
There are good ethical reasons to distinguish
129
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
research that involves human subjects from other
research, primarily that human subjects are
subjected directly to risks from the research
itself.
What is the secondary interest? Lack of
clarity about the secondary interests that are of
concern will also lead to bad policy. Current
regulations focus on financial interests, rather
than other, less-tangible interests such as
academic recognition and fame, or personal ties.
This is appropriate for the time being, not
because the intangibles are less damaging, but
because the financial interests are avoidable and
because avoiding them is consistent with the role
of a professional, and enhances public trust.
Financial interests have also increased to a high
level and are deserving of attention merely
because of their frequency. Furthermore, those
who point to the unfairness of concern about
financial interests seem to imply that financial
interests merely replace the non-financial
interests, so that there is no need for special
consideration of the financial interests. However,
the literature suggests that the effect of financial
interests on biomedical research can be detected
as an independent factor, above the background
“noise” of the want for academic recognition and
fame (assuming that it exists uniformly among
researchers).
There is less clarity about what specific kinds
of financial ties are of concern. Current
regulations focus on personal financial ties such
as consulting fees, honoraria, royalties and equity
holdings. They generally do not consider
company-sponsored research per se to be a
conflict of interest, but a growing body of
literature suggests that industry sponsorship in
itself biases research and publication (7-9, 13,
29).
How do we manage conflicts of interest?
Standard methods of managing, or mitigating,
conflicts of interest include (1) disclosure
(e.g., publication of a secondary interest),
(2) mediation (e.g., a blind trust, which puts a
secondary interest under the control of a third
party, or oversight, which puts a primary interest
under the review or control of a third party),
(3) abstention (e.g., recusal from a primary
interest), (4) divesti-ture (e.g., removal of a
secondary interest), and (5) prohibition (e.g.,
permanent withdrawal from a whole category of
secondary interests) (23). At first glance, these
five methods seem to be organized smoothly
along a continuum of stringency. However,
closer examination reveals that there is actually a
qualitative difference between these strategies,
because they are based on different assumptions.
In theory, all of these methods act by
modifying the conflict of interest situation
through either the primary or secondary interest.
However, disclosure is distinct from all the other
methods. It is supposed to act not by virtue of
supplying information to the disclosee, but
because the release of this information is
supposed to make the discloser more aware of
the potential effects and thus affect the
discloser’s behavior (24). Clearly this is a weak
method because of its indirectness. In practice,
the information rarely gets out to a wide
audience, and the discloser knows it, limiting
effectiveness. More importantly, this method
allows the discloser to feel that the act of
disclosing has let him or her off the hook, and
places the burden of management on the
disclosee. Stark points out that disclosure is
based on a model where the role of the discloser
is as an “agent”, or delegate, rather than a trustee.
By this model, the disclosee is assumed to have a
large degree of control over the activities of the
discloser.
In contrast, the other management methods
are based on a trustee or fiduciary model. By this
model, the disclosee is assumed to have little
control over the activities of the discloser and
therefore depends on the discloser to act in the
best interests of the disclosee. Mediation and
abstention carry with them the notion that the
fiduciary position is a role that can be filled by
interchangeable individuals. That is, the
protagonist can be replaced by a third party such
as an oversight committee or another researcher.
Divestiture and prohibition imply that the
protagonist is not replaceable, and therefore the
mitigation of the conflict of interests requires
removal of the secondary interest.
How we deal with conflicts of interest
depends on how we view the players. Are
researchers delegates or trustees? People who
hold elected public office may better fit the
delegate or agency model, since the public has
the power to remove them from office if their
performance is unsatisfactory. Researchers,
however, are more like trustees (especially
clinical researchers) because it is understood that
the public supports their training and activities to
perform tasks that others are not qualified to
130
–––––––––––––––––––––––––––––––––––––––––––––––– Cho, et al., What is Driving Policies on Faculty Conflict of Interest?
perform, and the public is not in a strong position
of control over these activities. The professional
role of scientists and clinicians is fiduciary in
nature, and requires that public interests be
placed ahead of self-interest.
How we deal with conflicts of interest also
depends on how broadly we define the interests
and the conflicts. The goal of academic-industry
ties is to maintain the ability to conduct good
science and to enhance technology transfer for
public good, while preserving research integrity
(including the direction of research) and, in the
case of clinical research, protecting human
subjects from harm. In order to achieve any of
these goals, it is essential to maintain the public
trust and a sense of professionalism, in the
original sense of the word (25, 30, 31), which
includes strong self-regulation (32).
Recommendations for policy development
What are the implications of these definitions of
interests and conflicts of interest for policy
development? First, conflicts of interest should
be defined by characteristics of situations, rather
than by outcomes. This allows taking into
account professional values as well as evidence
that certain situations tend to lead to bad
outcomes. Second, we should not rely on
disclosure as a primary mechanism for mitigating
conflicts of interest. Instead, we should
acknowledge that researchers have professional
responsibilities that are fiduciary in nature. As
trustees, they should be trustworthy. Third,
institutions should remember that institutional
interests play a role in individual conflicts of
interest, as well as the administration of policies
about individual conflicts of interest. Therefore,
institutions should not use policies only as
administrative tools, but also as mechanisms for
communicating institutional values to the public
(24, 31), because the nature of professionalism is
to profess a vow to place the interests of the
public above self-interest (33). The goal is to
provide reassurance to the public that the
institutions have also accepted their fiduciary
role.
Bibliography
1. National Institutes of Health, About NIH: Overview, .
2000.
2. Pharmaceutical Research and Manufacturers of
America, Pharmaceutical Industry Profile 2000, . 2000.
3. Krimsky, S., et al., Financial interests of authors in
scientific journals: a pilot study of 14 publications.
Science and Engineering Ethics, 1996. 2: p. 395-410.
4. Cauchon, D., FDA advisers tied to industry., in USA
Today. 2000. p. 1A.
5. Kahn, J., et al., Evaluation of HIV-1 Immunogen, an
immunologic modifier, administered to patients infected
with HIV having 300-549 x 106/L CD4 cell counts: a
randomized controlled trial. JAMA, 2000. 284: p. 21932202.
6. Rennie, D., Thyroid storm. JAMA, 1997. 277: p. 12381243.
7. Cho, M. and L. Bero, The quality of drug studies
published in symposium proceedings. Annals of Internal
Medicine, 1996. 124: p. 485-489.
8. Davidson, R., Source of funding and outcome of clinical
trials. J Gen Intern Med, 1986. 1: p. 155-158.
9. Rochon, P., et al., A study of manufacturer-supported
trials of nonsteroidal anti-inflammatory drugs in the
treatment of arthritis. Arch Intern Med, 1994. 154: p.
157-163.
10. Stelfox, H., et al., Conflict of interest in the debate over
calcium-channel antagonists. New Engl J Med, 1998.
338: p. 101-6.
11. Rochon, P., et al., Evaluating the quality of articles
published in journal supplements compared with the
quality of those published in the parent journal. JAMA,
1994. 272: p. 108-113.
12. Weiss, R. and D. Nelson, Teen dies undergoing
experimental gene therapy., in Washington Post. 1999:
Washington, DC. p. A1.
13. Blumenthal, D., et al., Withholding research results in
academic life science. JAMA, 1997. 277: p. 1224-1228.
14. Weiss, R. and D. Nelson, Penn settles gene therapy suit.,
in Washington Post. 2000: Washington, DC. p. A4.
15. U.S. Department of Health and Human Services,
Objectivity in Research, 45 CFR 94, 1995, U.S.
Government: Federal Register. p. 35810.
16. Cho, M., et al., Policies on faculty conflicts of interest at
US Universities. JAMA, 2000. 284: p. 2203-2208.
17. AAMC Ad Hoc Committee on Misconduct and Conflict
of Interest in Research, Guidelines for faculty
involvement in commercially supported continuing
medical education. Academic Medicine, 1992. 67: p.
617-621.
18. AHC Task Force on Science Policy, Conflicts of interest
in academic health centers policy paper #1, 1990,
Association of Academic Health Centers: Washington,
DC.
19. AAU Clearinghouse on University-Industry Relations,
University policies on conflict of interest and delay of
publication, 1984, American Association of
Univerisities: Washington, DC.
20. The American Society of Gene Therapy, Policy on
Financial Conflict of Interest in Clinical Research. 2000.
21. The American Society of Human Genetics, Statement on
Gene Therapy. 2000.
22. Bodenheimer, T., Conflict of interest in clinical drug
trials: A risk factor for scientific misconduct. 2000:
Bethesda, MD.
23. Thompson, D., Understanding financial conflicts of
131
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
interest. New Engl J Med, 1993. 329: p. 573-6.
24. Stark, A., Conflict of interest in American public life.
2000, Cambridge, MA: Harvard University Press. 271.
25. Buchanan, A., Is there a medical profession in the
house?, in Conflicts of interest in clinical practice and
research, R. Spece, D. Shimm, and A. Buchanan,
Editors. 1996, Oxford University Press: New York. p.
105-136.
26. Office of Science and Technology Policy, Fact Sheet:
Research misconduct - A new definition and new
procedures for federal research agencies. October 14,
1999.
27. US Senate, University and Small Business Patent
Procedures Act, in Bill S. 414. 1979.
28. Goldberg, C., Across the US, universities are fueling
high-tech booms., in New York Times. 1999: New York.
p. A1.
29. Blumenthal, D., et al., Participation of life-science
faculty in research relationships with industry. New Engl
J Med, 1996. 335: p. 1734-39.
30. DeAngelis, C., Conflict of interest and the public trust.
JAMA, 2000. 284: p. 2237-2238.
31. Korn, D., Conflicts of interest in biomedical research.
JAMA, 2000. 284: p. 2234-2237.
32. Rothman, D., Medical Professionalism — Focusing on
the Real Issues. New Engl J Med, 2000. 342: p. 1281-6.
33. Wynia, M., Medical professionalism in society. New
Engl J Med, 1999. 342: p. 1612-1616.
132
The Commercialization of Academic Science: Conflict of Interest Policies
and the Faculty Consultant
Lisa M. Jones, Postsecondary Education Policy Studies Center, University of Minnesota,
USA
Karen Seashore Louis, Educational Policy and Administration, University of Minnesota,
USA
Keywords: Academic scientists, Conflict of interest, Faculty consultants
Social scientists have studied the effects of faculty consulting on academic productivity - teaching,
research, and service (1- 6) – and used productivity as a proxy for conflict of interest. Most recently,
writers in both the disciplinary and popular literature have addressed conflict of interest and faculty
consultants. However, little empirical research that investigates the connection between
entrepreneurial behavior, consulting, and conflict of interest, exists. This study identifies four specific
behaviors that could compromise scientific objectivity and thus, be classified as conflicts of interest:
research agenda bias, prior review, withholding, and secrecy.
These conflict of interest behaviors are grounded in the norms and counternorms of science
proposed by Merton and Mitroff (7-8). Four norms dominate the roles of scientific researchers:
universalism, dissemination, disinterestedness, and organized skepticism.
Universalism suggests that science is open to all individuals regardless of their personal traits.
The scientific method is used to pursue truth. Dissemination allows for research to become open to
all challenges, subject to verification, and widely disseminated, the antithesis of prior review.
Research advances knowledge and resides in the public domain. Results become communicated so
that others may build upon previous work to move knowledge forward. The purpose of
communication also allows for research to become open to all challenges, subject to verification, and
widely disseminated (9).
The disinterested search for truth enables scientists to explore all information regardless of where
it might lead. Science’s reliance on verification and reliability reflect institutionalized controls to
ensure that knowledge benefits humanity and allows the researchers to proceed objectively. Although
knowledge advancement is the institutionalized role of scientists, some desire credit for their
discoveries vis-à-vis election to the National Academy of Sciences or a trip to Stockholm (e.g., Nobel
Prize). Conflicts then arise over the priority of discovery that further fuels secrecy. Furthermore,
academic science is a competitive industry that encourages researchers to withhold results for
personal aggrandizement either through enhanced reputation or financial gain. Entrepreneurial
behavior is a perceived threat to the researchers’ disinterestedness in the pursuit of knowledge for its
own sake. Burton Clark views entrepreneurialism as “a characteristic of social systems...taking risks
Corresponding author:Lisa M. Jones, Ph.D., Postdoctoral Associate, Postsecondary Education Policy Studies Center,
University of Minnesota, 330 Wulling Hall, 86 Pillsbury Drive S.E., Minneapolis, MN 55455-0208, 612-625-0773 (voice),
612-624-3377 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
when initiating new business practices where the
outcome is in doubt...(10)” The scientist
maintains a vested interest in the research
outcomes. When individual scientists establish
research agendas based on profitability, science is
not served. The payoff between basic research
discoveries and economic profitability often
requires time that neither society nor the
marketplace are willing to grant academics. This
creates the appearance that basic research
projects compete with commercially viable
proposals for funds.
Finally, Merton described organized
skepticism as the “temporary suspension of
judgment and the detached scrutiny of beliefs”
that affords scientists with the opportunity to
examine results using empirical or logical criteria
(11).
The search for truth rests upon the
foundations of basic research. When academic
scientists pursue lines of inquiry regardless of
their commercial viability, the public interest is
served. However, shifting political forces place
equal or even greater importance on
commercially viable academic science that could
stimulate economic growth expeditiously (12).
This study examines life sciences faculty
who report earning additional income by
consulting for non-profit organizations, industry,
and government and their engagement in actual
conflict of interest behaviors. This study limits
the definition to consulting activities for financial
remuneration, and examines individuals who
select consulting as a major source of
supplemental income from nonprofit
organizations or government agencies, private
enterprise, or both public and private.
Furthermore, the study examines behaviors of
those who consult exclusively with one company.
Methods
The data source used for this study is part of the
Academic-Industry Research Relationships
Study in Genetics and Other Life Sciences. The
analyses here are based on data from the broader
study’s 1994-1995 national survey of 3,169 U.S.
faculty in the life sciences. Fifty researchintensive institutions were selected based on the
levels of National Institutes of Health funding for
1993. All medical-school departments and other
academic life-science departments and graduate
programs were identified using the 1994
Peterson’s Guide to Graduate Programs in
Biological and Agricultural Sciences. One
medical clinical department, one non-medical
clinical department, and two non-clinical
departments were randomly selected from each
institution. Both the Peterson’s Guide and
University Bulletins identified 4,000 faculty that
included non-clinical, clinical, and researchers
funded by the Human Genome Project (HGP).
A stratified random sample of faculty, half of
whom were clinical and half of whom were nonclinical, were selected from a list of faculty
across the 200 departments. Special provisions
were made to include the HGP researchers
because of the broader study’s interest in
behaviors of genetics researchers. Ineligible
faculty (those who were deceased, retired, or not
located) were omitted from the sample, leaving a
final sample size of 3,169 faculty.
Data Collection
The data collection process occurred from
October 1994 through April 1995 by the Center
for Survey Research at the University of
Massachusetts. Each participant was mailed a
survey packet, which included a cover letter,
coded postcard, and questionnaire. The
questionnaire and postcard were to be returned
separately to protect respondent anonymity.
Reminder/thank you postcards were mailed
shortly after the initial mailing. Follow-up calls
conducted from late November to mid-February
to non-respondents generated an additional 190
cases for analysis. We received useable
responses from 2,052 faculty, for a total response
rate of 65 percent.
For this substudy, the sample consists of
the 1,032 non-clinical faculty respondents.
Selection of the individuals was assured by
including only faculty who do not conduct
clinical trials on “ drugs, devices, or diagnostic or
therapeutic technologies.” The non-clinical
faculty was chosen because previous research
conducted using the complete sample shows that
these individuals are on the “front end”
(entrepreneurial) of the commercialization
process. Furthermore, the industry relationships
between clinical faculty and corporations are
structured around clinical trials rather than new
discoveries (12).
Variables
Faculty gender, academic rank, average annual
research budget, average level of entrepreneurial
behavior, and average income earned above
salary were used as independent variables in the
134
–––––––––––––––––––––––––––––––––––––––––––––––––––––– Jones & Louis, The Commercialization of Academic Science
statistical analyses. The entrepreneurial behavior
scale constructed consists of the following survey
items: “Has the research that you do at your
university resulted in....(Check one for each
item)...patents applied for, a startup company.”
Individuals could check either yes (coded as “1”)
or no (coded “0”). The next question used for
this scale was: “For the firm with which you
currently have the greatest involvement, which of
the roles listed below do you have? (Check all
that apply)...equity holder, company owns or
licenses a patent based on your research.” If the
respondent left the item blank, it was coded as
“0” for no. A check mark was coded as “1” for
yes. The reliability for the entrepreneurial
behavior scale offered a standardized alpha of .69
(n = 1032).
Conflict of Interest measures
Research agenda bias. One conflict of interest
measure concerns external influence on research
topics: “To what extent has your selection of
research topics been affected by...(Check one for
each item) a) the likelihood of commercial
application of the results.” Participants were
offered the following response options: Not at all
(coded as “0”); very little (coded as “1”); to some
extent (coded as “2”); or, to a great extent (coded
as “3”). The results were collapsed into a
dichotomous variable coded “1” for yes and “0”
for no.
Prior review. Another conflict of interest
measure considers the publication relationship
between faculty and the sponsor. The following
item measured prior review: “Have you
personally conducted any research at your
university, the results of which are the property
of the sponsor and cannot be published without
the sponsor’s review or consent?” Yes was coded
as “1” and no as “0”.
Secrecy. This variable identifies the
relationship between commercial science and
publication of results. “Has your university
research resulted in findings that were never
published for proprietary reasons?” was the item
used to measure secrecy. Yes was coded as “1”
and no as “0”.
Withholding. The final conflict of interest
measure for this study considers the sharing
relationships between academic researchers.
This item asks individuals to report their denial
of others’ requests for research tools: “In the last
3 years, have any other university scientists
requested any results or materials that you did
not provide?” Yes was coded as “1” and no as
“0”.
Statistical analysis
Unless otherwise noted, statistical significance
and the direction of reported relationships
between consulting and conflict of interest
behaviors were tested by multivariate linear and
logistic regressions. The equations were adjusted
for academic rank, gender, institutional control
(public or private), academic program ranking,
institutional location (metropolitan versus nonmetropolitan), supplemental income amount, and
levels of entrepreneurial behavior.
Results
Sixty percent (n =616) of this sample (n =1032)
report that they have consulted with either public
(35.2%) or private (24.5%) enterprises at least
once. This contrasts with the 26% of the
respondents who consult with either group as a
major source of supplemental income. Table 1
shows the consultants’ characteristics broken
down by gender, academic rank, average research
budget, average level of entrepreneurial behavior,
and average income earned above salaries.
Males account for 82% of the sample, thus it is
not surprising to see them represented more than
females in the consulting categories (x2 = 24.74 p
< .001). Full professors represent 54% of the
total sample and are also consult more than
assistant and associate professors (x2 = 16.88 p <
.05). However, the assistant professors that
consult work more with private enterprise than
the public sector. One possible explanation for
this finding is that assistant professors may have
established relationships with companies during
their graduate training. The results further
indicate that those who consult exclusively with
one company tend to be male, full professors.
Furthermore, private enterprise consulting faculty
have larger research budgets than nonconsultants, which supports a Louis et al. (13)
earlier study that suggested that research budget
reflects entrepreneurial behavior as it indicates a
commitment to large-scale research. Private
enterprise consultants also report more
entrepreneurial behaviors. The analysis indicates
the specific entrepreneurial activities of these
individuals: 65% have applied for patents (x2 =
63.99 p < .01); 20% have started new companies
(x2 = 15.19 p < .01); 23% hold equity in a
company (x2 = 82.87 p < .001); and 15% are
involved with companies that own patents from
135
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
their university research (x2 = 31.94 p < .001).
When faculty who consult exclusively with
one company were compared with those who do
not (including non-consultants), exclusive
consultants report higher levels of entrepreneurial
behavior, research budget, and amount earned
above their institutional salaries. Table 2 shows
the mean differences between these groups.
Exclusive consulting offers greater financial
rewards for the academic scientist, which should
increase the potential for them to defy research
behavioral norms for self-aggrandizement.
The analysis indicates the specific
entrepreneurial activities of those who consult
exclusively with one company: 72% have
applied for patents (x2 = 30.41 p < .001); 35%
have started new companies (x2 = 33.65 p <
.001); 35% hold equity in a company (x2 = 83.61
p < .001); and 30% are involved with companies
that own patents from their university research
Characteristics
Rank
Gender
Consulting:
No
Consulting
Public
Consulting
Private
Consulting
(x2 = 70.09 p < .001).
Conflict of interest variables. When
consultants were asked to report on the conflict
of interest variables used in this study, we found
that of those who answered “yes”, the majority
were private enterprise consultants. Table 3
shows these results. Private enterprise and
nonprofit/government consultants were most
represented in research agenda bias (x2 = 26.58
p < .001); prior review (x2 = 37.15 p < .001);
withholding (x2 = 11.49 p < .01); and trade
secrets that resulted from university research (x2
= 10.61 p < .05). The results for secrecy were
not statistically significant.
Logistic regression analyses. Entrepreneurial
behavior level (0 to 4) is associated with private
enterprise consulting when gender, academic
rank, teaching, publication numbers, service,
research budget, and amount of supplemental
income are held constant. The most meaningful
Research
Budget
Male
Female
Assist.
Assoc.
Full
79%
21%
1 3%
24%
63%
239,752
Entrepreneurial
Behavior
.43
Income
o ver
Salary
4,9951
80%
20%
8%
29%
63%
355,494
.472
3,8803
96%
4%
1 7%
22%
61%
3 97,3374
1.145
1,52016
Table 1. Consultant characteristics (N=1032) reported in percentages and means. 1 Difference between non- and public
consultants (p < .001) 2 Difference between public and private consultants (p < .001) 3 Difference between public and
private consultants (p < .05) 4 Difference between public and private consultants (p < .001) 5 Difference between nonand private consultants (p < .001) 6 Difference between public and private consultants (p < .001)
Research Budget
Entrepreneurial Behavior
Income over Salary
365,5681
1.762
22,1703
269,196
.48
5,595
Consulting:
Exclusive
All Others in Sample
Table 2. Mean differences between exclusive consultants and all others in the sample on research budget, entrepreneurial
behavior, and amount earned over income. 1 (p < .05) 2 (p < .001) 3 (p < .001)
Research
Bias***
Consulting:
No Consulting
Public
Consulting
Private
Consulting
Behaviors
Prior
Withholding**
Review***
Secrecy
Trade
Secrets*
23%
24%
1 1%
9%
9%
8%
ns
ns
6%
7%
43%
2 9%
18%
ns
12%
Table 3. Consultant reports (N=1032) of conflict of interest behaviors. ***p < .001 ** p < .01 * p < .05
136
–––––––––––––––––––––––––––––––––––––––––––––––––––––– Jones & Louis, The Commercialization of Academic Science
variable in the equation is the private enterprise
consultant status (t = 9.32, p < .001), followed by
publication numbers (t = 4.48, p < .001). The
strength indicates that private enterprise
consultants appear more likely to engage in
entrepreneurial activities than either public
consultants or non-consultants. The full model,
which explains 15% of the variance, suggests
that faculty who consult with private industry
and who have higher publication numbers are
more likely to engage in entrepreneurial
behaviors than others.
There is a modest correlation between
supplemental income and private enterprise
consulting (r = .32, p < .001), and exclusive
consulting (r = .32, p < .001). Supplemental
income amount was not regressed on consulting,
however, because of these correlations. The
model, which accounts for 15% of the variance,
indicates that publication numbers, service levels,
and total research budget from all sources is
closely aligned with supplemental income
amount. The most salient independent variable is
service (t = 5.86, p < .001), followed by
publications (t = 3.73, p < .001) and overall
research budget (t = 3.61, p < .001).
Correlations show weak relationships
between private industry consulting and research
agenda bias (r = .16, p < .001), withholding (r =
.09, p < .01), and prior review (r = .18, p < .001).
Additionally, those who consult exclusively with
one company are correlated with research agenda
bias (r = .08, p < .001) and prior review (r = .15,
p < .001).
Logistic regressions were conducted to test
whether or not consulting with private enterprise
affects research agenda bias, prior review,
secrecy, and withholding. The models to test
private enterprise consulting effects included the
following control variables: faculty attributes,
institutional characteristics, academic
productivity measures, and entrepreneurial
behavior levels.
The first regression shows that the level of
entrepreneurial behavior (x2 = 74.05, p < .001) of
the faculty member as well as academic program
ranking and metropolitan location affects
whether or not they allow commercial potential
or funding opportunities to determine their
research agenda. This finding suggests that
faculty in highly ranked programs in
metropolitan areas are less likely to allow
external factors such as commercial viability and
funding to affect their research topics. However,
as levels of entrepreneurial behavior increase, the
odds that they define research topics according to
non-research-related dynamics increase by a
factor of 1.65.
The second regression tests the relationship
between consulting and prior review. The results
indicate that private enterprise consulting has a
negative effect on prior review, while
supplemental income amount and level of
entrepreneurial behavior has a positive effect (x2
= 68.16, p < .001). The probability that private
enterprise consultants will publish results only
after sponsor’s review decreases by a factor of
.50. However, the likelihood of prior review
increases by a factor of 1.59 for rising
entrepreneurial behavior levels and 1.24 for
supplemental income amount. Essentially, a
private enterprise consultant is less likely to
conduct research not published without the
sponsor’s consent. But, increased entrepreneurial
behavior and supplemental income do affect
prior review.
Private enterprise consulting does not appear
to affect withholding research tools from other
scientists who request them in either tested
model. Faculty in private institutions are less
likely to withhold (by a factor of .59), while
supplemental income increases the likelihood of
withholding (by a factor of 1.26). When
entrepreneurial behavior level is added, the
negative effect of institutional control remains
constant, while the supplemental income effect is
slightly lessened (x2 = 34.90, p < .001). Levels
of entrepreneurial behavior increase the chance
that one will withhold from others by a factor of
1.37. The results indicate that faculty in private
institutions are less likely to withhold from other
scientists even when controlling for levels of
supplemental income and entrepreneurial
behavior.
Finally, academic program ranking decreases
the likelihood that a scientists’ university
research results in trade secrets by a factor of .56
while level of entrepreneurial activity increases it
by a factor of 2.67 (x2 = 58.30, p < .001). This
model accounts for 21% of the variability for this
variable.
The models generated to explain why some
scientists conduct research that is never
published for proprietary reasons were not
statistically significant. Thus, issues related to
secrecy as defined in this study were not
examined in this analysis.
Analyses on the effects of exclusive
137
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
consulting on the conflict of interest variables
showed results that are similar to the private
enterprise consultant for research agenda bias (no
effect), prior review (negative association), and
withholding (no effect). These important
findings suggest that even the faculty member
who consults exclusively with one company is
unlikely to violate the research norms of the
academic enterprise.
Discussion
The results do not indicate that conflicts of
interest occur with any significant frequency; to
the contrary, the results show that academic
scientists are able to balance their
institutionalized scholarly roles with commercial
science. Faculty remain embedded in their own
social organizations which in the case of the
consultant includes the university, the discipline,
and the government, organization, or company
with whom one consults. Rowan and Miskel
argue that these social organizations generate the
norms that direct individual behavior (15).
Although conventional wisdom suggests that
when the faculty consultant serves multiple
masters, academic roles and norms are sacrificed
for self-interest, the results imply that the
consultant maintains an allegiance to the norms
of teaching, research, and service. Given these
criteria, the faculty in this study can be perceived
as actors within the institution of academic
science, rather than simply as a set of actors who
operate within a single organizational entity.
This argument is founded on the capacity of
faculty members to interact in a variety of
situations that appear to have competing interests
and values while they perfect their craft. If
academic science is the institution, the
institutionalized roles and norms embedded in
the scientific method become the criteria
consultant-scholars use to make decisions in their
relationships with commercial scientists.
University faculty have a societal contract
that affords researchers with academic autonomy
in exchange for a commitment to improve social
welfare through teaching, research, and service
(16). The question that drives university conflict
of interest policies is whether or not faculty
fulfill these institutionalized roles without
serving their own self-interest. If they fail to
fulfill their duties or pursue their own selfinterest in the course of their academic activities,
critics would argue that they are involved in a
conflict of interest. However, the conflicts that
academic scientists face are complex and do not
allow for a simple explanation.
Despite the lack of a positive relationship
between private enterprise consulting and the
conflict of interest variables tested in this study,
the need to protect universities, disciplines, and
the public from academic renegades remains.
Current methods such as disclosure to both
academic journals and universities provide an
important mechanism to alleviate conflict of
interest. However, these policies should be
grounded in conflict of interest behaviors, rather
than potentials, and enforced by individuals in
the academic community. Emanuel and Stein
reported that one out of three authors of journal
articles held financial stakes in reported research
outcomes and failed to disclose such in their
publications (17). If self-regulation of the
academic enterprise should continue without
external interference, enforced disclosure
becomes an important tool to prevent conflicts of
interest from bleeding into research activities.
The results of this study offer some
important implications for how academic policies
should be conceived. First, policy development
and implementation should rest upon data.
Empirical data provides a foundation for the
formulation of effective and enforceable policy.
The policies developed in this arena span the
boundaries between the disciplines, funding
agencies, academic institutions, and private
sector companies. Rather than establish
guidelines in isolation of one another, policies
could become aligned across these boundaries to
establish both consistency and clarity.
Ultimately, compliance becomes evaluated at
both the department and disciplinary levels.
Consistency and clarity across boundaries will
permit faculty to make informed choices.
Second, policymakers should develop clear
guidelines within their institutional and agency
sectors. Policies that guide rather than constrain
faculty behavior could aid faculty understanding
of specific behaviors that constitute conflict of
interest. Furthermore, clearly articulated
guidelines should identify the consequences of
individual action so faculty will understand the
ramifications of their behavior.
Finally, academic institutions could identify
consulting as a component of the faculty reward
structures. Boyer and Lewis suggested that
consulting could become a means for faculty to
involve themselves in both community and
institutional service (1). Consulting activity
138
–––––––––––––––––––––––––––––––––––––––––––––––––––––– Jones & Louis, The Commercialization of Academic Science
could become an element of faculty development
programs that stimulate faculty vitality and,
ultimately, productivity.
Acknowledgements
This research made possible by a grant funded by the
National Center for Human Genome Research, National
Institutes of Health (Grant Number HGO0724-01),
administered by the University of Minnesota and Harvard
University. For inquiries contact: Dr. Lisa M. Jones,
Postsecondary Education Policy Studies Center, University
of Minnesota, 330 Wulling Hall, 86 Pillsbury Drive S.E.,
Minneapolis, MN 55455-0208, 612-625-0773 (voice), 612624-3377 (fax).
Bibliography
1. Boyer CM, Lewis DR. Maintaining Faculty Vitality
Through Outside Professional Consulting. In SM Clark,
DR Lewis (editors). Faculty Vitality and Institutional
Productivity: Critical Perspectives for Higher
Education. 1985.
2. Fairweather JS. Faculty Work and Public Trust:
Restoring the Value of Teaching and Public Service in
American Academic Life. Needham Heights (MA):
Allyn and Bacon Publishers; 1996.
3. Patton CV, Marver JD. Paid Consulting by American
Academics. Educational Record 1979: 175-184.
4. Rebne D. Faculty Consulting and Scientific Knowledge:
A Traditional University - Industry Linkage.
Educational Administration Quarterly 1989; 25(4): 338357.
5. Teague GV. Faculty Consulting: Do Universities Have
“Control”? Research in Higher Education 1982; 17(2):
179-186.
6. Marsh HE, Dillon KE. Academic Productivity and
Faculty Supplemental Income. Journal of Higher
Education 1980; 51(5): 546-555.
7. Merton RK. The Sociology of Science: Theoretical and
Empirical Investigations. Chicago: University of
Chicago Press; 1973.
8. Mitroff II. Norms and Counternorms in a Select Group
of the Apollo Moon Scientists: A Case Study of the
Ambivalence of Scientists. American Sociological
Review 1974; 39(4): 579-595.
9. Louis KS, Anderson MS. The Changing Context of
Science and University-Industry Relations. In H
Etzkowitz, A Webster, P Healey (editors). Capitalizing
Knowledge: New Intersections of Industry and
Academia. Albany (NY): State University of New York
Press; 1998.
10. Clark BR. Creating Entrepreneurial Universities:
Organizational Pathways of Transformation. United
Kingdom: Pergamon; 1998: 4.
11. Merton, RK. op cit. 1973: 277.
12. Etzkowitz H, Webster A. Entrepreneurial Science: The
Second Academic Revolution. In H Etzkowitz, A
Webster, P Healey (editors). Capitalizing Knowledge:
New Intersections of Industry and Academia. Albany
(NY): State University of New York Press; 1998.
13. Louis, KS, Jones, LM, & Anderson, MS. Bought and
Sold: Academic Ideas in the Marketplace. Paper
presented at the meeting of the Association for the Study
of Higher Education, Miami, FL. (1998, November 6).
14. Louis, KS, Blumenthal, D, Gluck, ME, & Stoto, MA.
Entrepreneurs in Academe: An Exploration of
Behaviors Among Life Scientists. Administrative
Science Quarterly, 1989; 34, 110-131.
15. Rowan B, Miskel CG. “Institutional Theory and the
Study of Educational Organizations”. In J Murphy, KS
Louis (editors). Handbook of Research on Educational
Administration, Second Edition. San Francisco (CA):
Jossey-Bass Publishing; 1999.
16. Slaughter S, Leslie LL. Academic Capitalism: Politics,
Policies, and the Entrepreneurial University. Baltimore
(MD): The Johns Hopkins University Press; 1997.
17. Emanuel EJ, Steiner D. Institutional Conflict of
Interest. N Engl J Med 1995; 332 (4): 262-267.
18. Boyer CM, Lewis DR. Maintaining Faculty Vitality
Through Outside Professional Consulting. In SM Clark,
DR Lewis (editors). Faculty Vitality and Institutional
Productivity: Critical Perspectives for Higher
Education. 1985.
139
5. Understanding Misconduct
Preventing Scientific Misconduct: Insights from “Convicted Offenders”
Mark S. Davis, Justice Research & Advocacy, Inc., USA
Michelle L. Riske, Justice Research & Advocacy, Inc., USA
Keywords: Equity theory, Prevention, Rationalization techniques, Research misconduct, Responsible conduct
of research, Scientific misconduct
The mere seriousness of certain social behaviors implies the need to prevent them. In the case of
conventional crime, for example, survivors of homicide victims or the victims of physical or sexual
assault, when asked what they want most, often will say they wish the incident had never happened.
For them, a successful homicide prosecution does not bring back the lost loved one. A long prison
term for the rapist will not restore the victim to the state she enjoyed prior to the crime. As a result,
we strive to identify and implement various ways of reducing opportunities for both offending and
victimization.
Although the perceived harm in research misconduct may not be as great as in violent crime, its
consequences nevertheless can have disastrous and far-reaching effects. After-the-fact measures such
as the investigation of allegations and the sanctioning of the guilty, while necessary for justice and the
vindication of the moral order, seldom can undo the harm caused by each instance of fabrication,
falsification, plagiarism, or other serious departure from the norms of science. The retraction of a
published paper cannot restore the time wasted by other investigators pursuing pointless lines of
research or by editors and referees reviewing meaningless results. An apology and a signed voluntary
consent agreement by one found guilty of research misconduct does not automatically lift the taint
from the supervisor and colleagues in whose lab the misconduct occurred. And for those who suffer
from life-threatening diseases and consequently hold out hope for a cure, the broken trust of falsified
clinical trials has far more devastating effects. To be sure, the shock waves emanating from a single
incident of research misconduct can create untold collateral damage, including the tarnishing of
reputations of scientists, institutions, and of the enterprise of science itself.
In view of our collective inability to undo the damage and effect restoration to all parties in these
cases, the prevention of research misconduct is a desirable end. The question then becomes, what can
the scientific community do to keep research misconduct from occurring in the first place? The
purpose of this preliminary analysis is to explore largely untapped data sources in order not only to
advance theoretical work in this area, but also to glean information of practical import.
In order to tackle the challenge posed by prevention, we must acknowledge that prevention can
occur at more than one level. Douglas Weed, employing public health’s notions of primary and
secondary prevention, suggests that we first need to know something about etiology, and he argues
that there are causal factors both internal and external to the scientist who engages in research
misconduct (1) . Examples of internal causal factors would include psychological problems, financial
motivations, or perhaps the desire to hurt others. Causes external to the scientist, on the other hand,
Corresponding author: Michelle L. Riske, J.D., M.A., 849 Cleveland Ave., Amherst, OH 44001, 440-988-8455 (voice), 440988-8455 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
are factors such as the pressure to publish,
inadequate training or supervision, or the fierce
competition for research grants.
In either case, successful prevention requires
that we somehow interrupt one or more processes
that lead to an instance of research misconduct.
For example, if we knew that individual
psychopathology was responsible for research
misconduct, we perhaps could administer the
Minnesota Multi-phasic Personality Inventory
(MMPI), the Narcissistic Personality Inventory,
the Psychopathy Checklist, or other psychometric
tools to help us screen out applicants who were
predisposed to engaging in unethical research
practice. In an effort to address an external cause
such as inadequate supervision, we might
institute regular meetings between lab
supervisors and their staff members.
Objectives
This pilot study focuses on two individuallevel explanations for research misconduct. First,
Cressey’s research on embezzlement in financial
institutions was examined (2). Cressey’s
subjects, who largely perceived themselves to be
respectable people, had three characteristics in
common:
1. A non-shareable financial problem, for
example, one the individual could not
discuss without suffering humiliation;
2. An awareness the problem could be solved
by violating the position of financial trust;
and
3. Suitable rationalizations for the embezzlement of funds to resolve their self-conception as a trusted person.
Applying Cressey’s work to scientific
researchers, is it possible that some have nonshareable problems, not necessarily financiallybased, which motivate them to engage in
research misconduct? The possibilities could
include the inability to produce replicable work
under pressure, a perceived lack of talent for
research, or personal problems such as marital or
emotional difficulties. For example, William
Summerlin, the protagonist in one of the bestknown cases of research misconduct, intimated
that he had been under a lot of pressure from the
head of the lab to produce results. Could the
inability to withstand this sort of pressure
constitute a non-shareable problem?
In addition to possibly having such nonshareable problems, how do researchers who
engage in misconduct formulate rationalizations
for their behavior? And what form might these
rationalizations take? Sykes and Matza, in their
research on juvenile delinquency, discuss several
of what they refer to as “techniques of
neutralization” including (3) :
• Denial of a victim (Who am I really hurting
by fudging these data?)
• Denial of an injury (What is the harm?)
• Condemnation of the condemners (They’re
out to get me.)
• Denial of negative intent (I never meant to
hurt anyone.)
• Metaphor of the ledger (For most of my time
here in the lab I’ve been a hard-working,
loyal employee. I’m entitled to a slip or two.
All in all, I’ve done more good than bad.)
Is it possible that individuals who commit
research misconduct may employ one or more of
these techniques in order to justify their conduct?
The second perspective employed for this
study was social psychology’s equity theory,
which speaks to perceived fairness in dyadic
relationships (4). Equity theory is exemplified in
the common phrases “You scratch my back and
I’ll scratch yours” and “One good turn deserves
another.” Social beings have come to expect
reciprocity when dealing with others. If people
perceive they are getting less from a relationship
than they are given, they may suffer distress. It
is common, then, for the ostensibly exploited
person to take measures to relieve this distress
and restore a sense of equity. In the case of
research misconduct, scientists may be more
likely to engage in misconduct if they believe
they were deprived of what was rightfully theirs,
such as the co-authorship on a publication or a
coveted promotion. Accordingly, individuals
may engage in scientific misconduct as a form of
retaliation against a coworker or supervisor if
they believe that they have been slighted or
exploited.
Design
Two sources of data were gathered for this study.
The first was information from the case files of
individuals against whom a finding of scientific
misconduct was made by the Office of Research
Integrity (ORI). A standard data collection form
was used to record data including the institution,
type of alleged misconduct, information from the
respondent, response of the institution, and
finding by the ORI. A member of the research
144
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Davis & Riske, Preventing Scientific Misconduct
subjects with the case file reviews. Upon
completion of the interviews, the subject list was
given to ORI. Both data collection instruments
were approved by an Institutional Review Board
and by the U.S. Department of Health and
Human Services, Office for Protection from
Research Risks.
team read each case file and wrote narrative
responses to the items on the data collection form
summarizing information primarily pulled from
the investigative reports by the universities and
from the investigative reports of ORI and its
predecessors. These narrative responses were
analyzed for this part of the study. A total of 21
case files were reviewed for the initial pilot
study. These case files included 16 cases
reviewed as part of a pretest, as well 5 additional
cases that included cases closed prior to the
formation of the ORI, i.e., these cases were
handled by the Office of Scientific Integrity
(OSI), ORI’s predecessor.
The second source of data consists of
interviews with scientists against whom a finding
of scientific misconduct was made by the ORI.
Subjects who were included in the first nine case
files used as part of the pretest comprised the
sample for this portion of the data collection
process. Because some scientists approached
could not be located or were unwilling to
participate in the interviews, only three out of the
nine contacted were interviewed. It is possible
that the experience of having been accused and
found guilty of research misconduct was so
unpleasant that some subjects have little interest
in dredging up the past. One scientist who
declined to participate in the study summed up
his feelings in an e-mail to the senior author:
“I am very sorry to disappoint you but after
more then ten years I have no inclination to
discuss this issue with anybody. With my very
poor English I found it useless to talk about
the inquisition. I have no idea what is a (sic)
subject and goal of your research, but I wish
you a (sic) success in your work in the name of
justice, science and humanity.”
One of the interviewees summed up his feelings
more bluntly when thanked for his time:
“The time is not the problem; it’s the pain of
having to relive this crap.”
The researchers signed a confidentiality
agreement with ORI to protect sensitive case file
information. The researchers also took additional
steps to ensure confidentiality during the data
collection process, by excluding the subjects’
name and case file number from the data
collection instruments. Subjects were identified
by the assignment of a subject number. To match
files with subjects being interviewed, a list
including the subject name, institution, ORI case
number, and subject number was created. The
information was only used to link interview
Methods of Analysis
Because theoretical work on scientific
misconduct is relatively meager, we chose to use
a qualitative approach borrowed from
phenomenological psychology. Rather than first
searching for evidence of specific theories or
propositions, the investigator examines the data
more for “explication” than explanation (5). This
results in the listing and preliminary grouping of
terms or phrases revelatory of, in this case,
etiology. As a check against possible bias created
by prior knowledge or other factors, the analyst
extracts exact phrases rather than interpreted
concepts. Another analyst approaches the data in
the same way, identifying exact wording to
convey possible sources of causation. The
second step involves the two analysts coming
together to compare and reconcile their lists. In
the third step, the analysts group the phrases into
common themes or constructs. Finally, the
constructs are examined to see if they relate back
to the selected theoretical approaches in order to
help us interpret and discuss the relevance of
these constructs or central themes in explaining
the etiology of research misconduct. For
example, in looking at Cressey’s notion of the
non-shareable problem (6), the analyst would
group together those extracted phrases suggesting
such themes as psychological issues, marital
difficulties, financial pressure, lack of
knowledge, difficulty with expectations of a
supervisor, lack of supervision, or other problems
an individual might reasonably be uncomfortable
sharing with others.
Data obtained from the case file reviews and
from the interviews eventually will be content
analyzed using the QSR-NUDIST software.
Content analysis is a means of systematically
analyzing textual information to find recurring
themes, issues, and motifs, which can then be
isolated, counted, and interpreted (7, 8) . If the
appropriate statistical criteria are met, the data
will also be analyzed to examine relationships
among variables in order to assess, for example,
if a certain type of misconduct or rank is
145
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
associated with the existence of a non-shareable
problem.
The Sample
The data collected was part of a pilot study to test
the efficacy of the data collection instruments
developed, which were then used as part of a
larger study examining all individuals against
whom a finding of scientific misconduct was
made by the ORI as of December 2000. A total
of 21 case files were reviewed for the pilot study.
Many of the respondents held academic positions
as either Senior Faculty or Junior Faculty (each
Number of Subjects
10
8
8
8
6
category included 8 out of the 21 subjects).
Senior Faculty included professors, associate
professors, and directors/heads of departments,
institutions or clinics. Junior Faculty is defined
as assistant professors, postdoctoral students,
research fellows and residents. Other
researchers, including research associates,
predoctoral students, and administrative
assistants, made up the remaining positions (5
out of 21). It should be noted that tenure status
could not be gleaned from the case files.
With respect to the types of research
misconduct committed by these 21 respondents,
38% of the cases were for plagiarism, 19% were
for fabrication, and 19% were for falsification.
Fabrication/falsification made up 14% of the
cases, and the remaining 10% were for a
combination of falsification, fabrication, and
plagiarism.
5
Results
4
Data from the case files reviewed were analyzed
using the qualitative phenomenological
approach.
2
Etiology
The
systematic search for possible etiological
Type of Position
factors related to our two theoretical
Fig. 1. Researcher's Academic Position
perspectives yielded data in support of both
theories.
Phrases or
Combination
elements
10%
Fabrication
extracted from
19%
the case files
showed evidence
Fabrication/
of non-shareable
Falsification
problems such as
14%
publish-or-perish
pressure, lack of
knowledge or
experience,
difficulty with
supervisor’s
Falsification expectations/lack
19%
of supervision,
and personal
problems. These
phrases were
usually extracted
from information
contained in the
Plagiarism
University
38%
investigative
Fig. 2. Classification of Research Misconduct
reports or the
Senior Faculty
Junior Faculty
Other
Researchers
146
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Davis & Riske, Preventing Scientific Misconduct
Job Pressure
Lack of Subject
Matter Knowledge
Personal Problems
Personal insecurity
Problems with
Supervision
Enormous pressure to
produce/ Pressure to
produce by supervisor
Understanding of grant
application process/First
proposal
Time factors - short
deadlines/ Short cut to
save time
Different interpretation of Personal/Family
the normal modes of
difficulties
responsible authorship
Supervisor was
demanding in research
results
Pressure to keep the
system working
Understanding of the
principles of attribution
in review articles
Lacked proper scientific
guidance from supervisor/
Unsupervised
Insecure position
Not able to handle
position/ Saddled with
responsibilities which in
hindsight were out of
proportion to subject’s
training and experience
Under personal pressure
from supervisor to
publish data in order to
secure a research
position.
Isolated laboratory with
few peers to discuss
situation or possible
career alternatives
Never trained in
appropriate record
keeping
Negligent oversight/
Deficiencies in oversight/
Supervisor’s oversight
was inadequate
Medical illness
Could not fully satisfy
the expectations of the
supervisor/ If supervisor
had more realistic
expectations this incident
might never had occurred
Difficult job situation/
Stressful job situation
Table 1. Etiology - Non-shareable problem
investigative reports from ORI; therefore, the
information is hearsay, as the actual statements
made by the respondent or other interested
parties were usually not contained within the case
files.
Information obtained from the interviews
also provided evidence in support of a nonshareable problem by the respondent, which may
have contributed to his misconduct. For
example, one interviewee stated:
“How am I going to get a position where I don’t
have to worry every 2-3 years about if I don’t
get my grant I’m gonna be out on the street.
This feeling of being a second, kind of a second
class citizen. Um, the pressures to produce
papers. And, you know, it was, I knew I was
doing something wrong, but I reached a point
where I didn’t care.”
The data also contained summarized statements
from respondents indicating rationalization
techniques of denial of an injury, condemnation
of condemners, and denial of negative intent.
Although information extracted from the case
files did not definitively point to instances where
the subject engaged in conduct in order to restore
a perceived loss of equity in a dyadic
relationship, some of the phrases taken from the
case files suggest possible motivation by the
subjects that could indicate retaliatory conduct in
response to perceived exploitation. For example,
some of the subjects said that they falsified data
in order to annoy colleagues or that they were not
recognized for their scientific expertise. Other
subjects discussed competition in relation to
positions within the university or institution and
competition for submitting papers for
publication.
Implications for Prevention
If we look at the preliminary evidence for our
theoretical questions, we can infer some tentative
implications for prevention. Information
pertaining to lack of proper supervision or
training suggests that it might be prudent for
universities to implement better procedures and
guidelines for supervisors with respect to
employee oversight and monitoring
responsibilities. We found some support that
periodic reviews or audits of research notebooks,
as well as the original data collected for all
experiments, by the supervisor may help to
reduce research misconduct. Ensuring that
147
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Denial of an Injury
Condemnation of the
Condemners
Denial of Negative Intent
No harm done because the
experiments were preliminary,
unimportant, and had not been
published
Worked on several of the articles
which were used as references for
the proposal and therefore
permitted to incorporate these
materials into the proposal
If there was some faulty reporting
of findings, that it was minimal
since it was not the central issue
of the study
Subject had opposite and
competing opinions to research
performed by colleagues of the
complainant
Allegations by complainant were
an attempt to “get rid of” the
subject from the University
Fabricated sampling times were
preliminary and never intended to
be published
Going to tell supervisor the truth
after the subject had a chance to
obtain valid counts, but the
subject didn’t have the chance
Table 2. Etiology - Neutralization Techniques
employees are properly trained on all
experimental techniques prior to performing such
experiments could also help reduce the
researcher’s lack of knowledge on the subject
matter, as well as apprehension about
acknowledging that as a problem. Similarly,
discussing the serious ramifications of research
misconduct can also discourage some of the
denial its perpetrators use to rationalize their
actions with such conduct; for example, that
there indeed is harm associated with these actions
that affects a variety of actors and institutions,
including, most importantly, the patient
population.
The three interviews conducted to date have
also provided some insights for prevention. One
subject credited the careful handling of data for
his own demise:
“...when the technician did the, you know, do
the random stuff, yes, there would be a copy
on the computer, but he would also print out
the data, you know, a paper copy and put that
into their books. So, it was, you know, like, it
was also like a guarantee that I would
eventually be found out and that it could all be
traced back.”
So upon his returning to the lab from an extended
trip:
“...basically they sat me down and confronted
me with the fact that these data sets don’t fit.
And, it was a situation of, uh, what do you say
if you’re caught red-handed? You know all the
original data was there. It was very easy for
them to go back to the original sets and see
that there were discrepancies.”
This same interviewee briefly contemplated
trying to cover up the misconduct, but again
realized:
“...it was truly a situation where the record
keeping system that I had set up was such that
there was no way I could possibly go back
through all the computer files and alter those.
There was, you know, everything, the techs had
always printed out paper copies, so there was
shelves of three ring binders with all the data.
It was a situation of, it can’t be done.”
One interviewee felt that training might help
prevent some research misconduct:
“I think that there should be more training,
study in just the basics of the scientific method
and, you know, what is appropriate, you know,
Evidence of possible motivation to retaliate Evidence of possible motivation to exploit
Made up data to annoy a colleague
Future dependent on rapid success in the laboratory
Some friction between subject and others in the lab
Laboratory placed too much emphasis on short-term
productivity
Bitter relationship between subject and supervisor
Failed to make changes because upset with others
Attempt to get rid of subject
Competitive pressure for tenure-track positions
Insecure position
Personal animosity against the subject/Prejudice
against the subject
Table 3. Etiology - Equity Theory
148
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Davis & Riske, Preventing Scientific Misconduct
what is not appropriate in terms of experimental
methodology, in terms of statistics, in terms of,
if you’re going to discard data, you know, what
are the other experimental reasons for
discarding the data? For example, oh yeah, I
had a sudden sneeze and I sneezed and I
botched test tubes or I knocked over this
particular test tube, or I tested this particular
agent and found that, oh my gosh, I actually
added 10 times the amount of a particular
component, you know, those are valid reasons
for discarding data. You know, I don’t think
there’s enough emphasis placed on teaching
people the proper scientific method.”
3. Sykes, G, Matza, D. Techniques of neutralization: A
theory of delinquency. American Sociological Review
1957; 22: 664-670.
4. Walster, E, Walster, GW. Equity: theory and research.
Boston (MA): Allyn & Bacon; 1978.
5. von Eckartsberg, R. Existential-phenomenological
research. In von Eckartsberg, R, editor.
Phenomenological inquiry in psychology: Existential
and transpersonal dimensions. New York: Plenum Press;
1996, p. 3-61.
6. Cressey, D. Other people’s money: The social
psychology of embezzlement. New York: The Free
Press; 1953.
7. Denzin, NK, Lincoln, YS. Part I: Method of collecting
and analyzing empirical materials. In Denzin, NK,
Lincoln, YS, editors. Collecting and interpreting
qualitative methods. Thousand Oaks (CA): Sage; 1998,
p. 35-45.
8. Maxfield, MG, Babbie, E. Research methods for
criminal justice and criminology. 2nd ed. Belmont (CA):
West/Wadsworth Publishing Co; 1998.
Another subject offered what he referred to as an
“easy” solution to the problem of fabrication and
falsification:
“What you do, is you have, uh, open laboratory
meetings where everyone in the laboratory
knows what everyone else is doing. Uh, you
say you did an experiment that took a hundred
rats, but only five rats came into the, into the
lab, it’s pretty clear that you didn’t do a hundred
rats. Uh, if you’re not there doing the work,
uh, that people think you’re doing or know that
you’re supposed to be doing, uh, so I think, uh,
open laboratories, with regular, uh,
presentations of data prevent that.”
Conclusions
We used a qualitative approach to explore
selected aspects of individual-level etiology of
research misconduct. These preliminary data
offer some tentative support for our theoretical
perspectives. More definitive conclusions will
have to await the collection and analysis of the
data from the larger study.
This research-in-progress also offers support
for certain forms of prevention. These
suggestions, rather than the product of wellmeaning, but less-than-well-informed
commentators, come from those most intimately
involved in actual cases. Returning to the
analogy of crime, learning from those who have
engaged in research misconduct is not unlike
debriefing convicted burglars on what would
have kept them from choosing a particular
dwelling as a target. Who should know better
than those who have done it?
Bibliography
1. Weed, DL. Preventing scientific misconduct. American J
Public Health 1998; 88(1): 125-129.
2. Cressey, D. Other people’s money: The social
psychology of embezzlement. New York: The Free
Press; 1953.
149
The Relative Efficiency of Research Misconduct Investigations Involving
Personal Injury vs. Injury to the Scientific Record
Andrew J. Hogan, Department of Medicine, Michigan State University , USA
Ronald J. Patterson, Department of Microbiology and Molecular Genetics, Michigan State
University, USA
Robert L. Sprague, Department of Kinesiology , University of Illinois-Urbana/Champaign,
USA
Keywords: Efficiency, Research misconduct, Type of injury
Research misconduct investigations conducted by universities and other research institutions are
sometimes highly contentious affairs whose findings are disputed both internally and externally. The
central question of the research reported in this paper is whether certain features of the typical internal
research misconduct investigation contribute to the likelihood of closure or to continued contention.
Most research misconduct investigations undertaken in institutions that receive Federal research
contracts and grants follow the investigational model proposed by the National Institutes of Health
(NIH), described here as the tribunal model. In civil law, similar types of disputes (civil fraud,
misappropriation of property, etc.) are dealt with in adversarial proceedings. One measure of the
efficiency of the typical model for conducting a research misconduct investigation is to determine
how often that model produces a definitive finding, or alternatively how often it leads to further
proceedings.
The objective of this study was to test whether the presence of personal injury associated with a
research misconduct allegation influences the likelihood of a post-investigation proceeding (lawsuit,
grievance, legislative hearing, administrative inquiry, etc.), in the context of the use of the tribunal
model of investigation. We hypothesized that the standard tribunal model, which was designed
principally to protect the integrity of the scientific record, might not be very efficient in addressing
misconduct allegations in which a personal injury was the central feature.
Materials and Methods
Data. Cases were identified in the files of Dr. Robert Sprague of the University of Illinois-Urbana/
Champaign, which contain 1,100 references on the 231 research misconduct cases (hereafter referred
to as the “Sprague files”). The Sprague files consist primarily of copies of news stories in scientific
journals, such as Science and Nature, or academic trade journals, such as the Chronicle of Higher
Education and Lingua Franca.
Sixty-three cases were identified as having adequate documentation of alleged misconduct
Corresponding author: Andrew J. Hogan, Ph.D., Health Services Research Division, Department of Medicine, Michigan
State University, B210 Clinical Center, East Lansing, MI 48824, 517-353-4361 (voice), 517-432-1326 (fax),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
involving either a personal injury or an injury to
the scientific record. A personal injury case was
one in which a person directly involved in the
misconduct allegation identified some kind of
personal loss, usually misappropriation of
intellectual property — plagiarism or the
unauthorized use of confidential information
from grants or articles under peer review. A
scientific record case was one involving some
form of contamination of the scientific record.
Scientific record cases usually involved
falsification/fabrication, but sometimes involved
misappropriation of the intellectual property of
non-parties to the allegation.
Post-investigation proceedings included
grievances filed within the institutions, lawsuits,
complaints to regulatory or funding agencies, and
requests to legislative or administrative bodies. A
post-investigation proceeding was classified as a
due process case if one or more of the parties
raised due process issues (hearing notification,
right to call or cross-examine witnesses, impartial
decision-makers, etc.) related to the research
misconduct investigation.
In the tribunal model of a research
misconduct investigation, an individual files an
allegation with an institution, and the institution
forms a panel to investigate the allegation. The
panel is responsible to gather evidence, call and
examine witnesses, and make a finding; in
common parlance, the tribunal is prosecutor,
judge and jury. The standard NIH-tribunal model
often attenuates some due process rights
commonly found in adversarial proceedings, in
particular rights to call or cross-examine
witnesses and to present evidence. Current NIH
policy suggests that the complainant in such an
investigation be treated as a witness, rather than
as a party.
In an adversarial proceeding, one party
(complainant) accuses the other party
(respondent) of misconduct. The parties gather
and present evidence, call and examine and
cross-examine witnesses. The institution provides
an adjudicator to process the allegation, hold
hearings and render a decision. We were able to
identify no unambiguous cases in which the
adversarial model was employed in a research
misconduct investigation.
Data Collection and Reliability. We
reviewed 221 documents related to the 63
identified cases. For each document, a form was
completed (see Appendix A) identifying the case
name and the document number in the Sprague
files. The abstractor (Hogan or Patterson)
identified the type of misconduct alleged
(fabrication/falsification, misappropriation of
intellectual property, other serious deviations,
retaliation, or other). The abstractor then
determined the nature of injury based on whether
there was an injured party known to the
individual alleged to have committed
misconduct; if so, the case was classified as one
involving personal injury, otherwise as injury to
the scientific record. Next the abstractor coded
for the type of institutional investigation (tribunal
or adversarial), based principally on whether the
complainant was a witness or a prosecutor.
The abstractor then determined whether there
were other proceedings consequent to the
institutional research misconduct investigation,
such as:
• Internal grievances, discrimination complaints, etc.
• Lawsuits, complaints/appeals to administrative agencies, complaints/appeals to legislative bodies.
In those cases where there was some sort of postinvestigation proceeding, the abstractor
determined whether due process issues were
raised.
Finally, the abstractor examined each
document regarding the role of the institutional
legal counsel as being supportive, neutral, or
obstructive of the procedural fairness of the
institutional investigation. The abstractor looked
for any references to the role of institutional legal
counsel regarding the selecting or preparing
witnesses, selecting or preparing panelists,
selecting or preparing administrators, handling
administrative problems/complaints, issues of
attorney-client privilege, providing or
withholding information, applying legal
indemnification, deliberating or making findings,
the preparing or editing of reports, the protecting
of parties’ due process rights.
To assure the reliability of the abstraction
process, the first 20 cases were reviewed by both
abstractors to establish interrater reliability using
a data collection tool. Review of the reliability of
the initial cases indicated a 94% agreement on
which documents were relevant to each case, a
70% agreement regarding the type of
misconduct, and a 91% agreement on whether
the injury was personal or to the scientific record.
There was a 60% agreement on which documents
indicated the type of institutional investigation,
152
––––––––––––––––––––––––––––––––––––––––– Hogan et al., The Relative Efficiency of Research Misconduct Investigations
but 100% agreement on the type of institutional
investigation. There was also 100% agreement
regarding the existence of post-investigation
proceedings. The reasons for the discrepancies in
the classification of misconduct allegations were
discussed and resolved before finishing the
abstraction of the remaining cases.
In the subset of cases where due process
issues were raised, any controversies regarding
the role of the institutional attorney in the
research misconduct case tended to increase the
likelihood of a post-investigation proceeding by
more than six-fold (see Table 2). However, this
result was only marginally statistically
significant.
Results
No unambiguous cases where the original
research misconduct investigation was
administered using the adversarial model were
found. All of the results related to research
misconduct investigations which were conducted
under the standard tribunal model.
Of the 63 cases described in the 221
documents reviewed, 41% of cases resulted in a
post-investigation proceeding, and 69% of these
involved a due process issue. Of the 63 cases,
41% of cases involved personal injury, and 70%
of personal injury cases resulted in a
post-investigation proceeding. Of the personal
injury cases resulting in a post-investigation
proceeding, 61% of these proceedings involved a
due process issue.
Ten percent of the 63 cases involved some
controversy regarding the role of the institutional
attorney. Although we looked for instances where
the role of the institutional attorney was
supportive of procedural fairness, only negative
statements appeared in the literature examined.
Twenty-one percent of cases arose in the context
of a funded grant.
Multivariate logistic regression analysis was
performed to determine the likelihood of
post-investigation proceedings. The results are
presented in Table 1. Personal injury cases are at
least 10 times more likely to result in a postinvestigation proceeding than cases involving
injury to the scientific record. When allegations
are made in the context of a funded grant, the
likelihood of a post-investigation proceeding is
reduced, although this effect is only marginally
statistically significant.
Parameter
Personal injury
Attorney
controversy
Grant context
Odds Ratio
10.34**
95% Bounds
Upper
Lo wer
36.46
2.94
3.71
33.39
0.41
0.22*
1.12
0.04
Table 1. Logistic Regression Analysis:
Likelihood of Post-Investigation Proceeding.
n=63, ** = p < 0.05; * = p < 0.10
153
Parameter
Personal Injury
Attorney
controversy
Grant Context
Odds Ratio
3.39**
95% Bounds
Upper
Lower
11.28
1.028
6.50*
46.16
0.92
0.35
1.88
0.07
Table 2. Logistic Regression Analysis: Likelihood of PostInvestigation Proceeding Involving Due Process.
n=63 / ** = p < 0.05 / * = p < 0.10
Conclusions
Because we were able to identify only two
ambiguous cases of research misconduct
investigations possibly employing an adversarial
model, we were not able to determine whether
the adversarial model would result in fewer
post-investigation proceedings than the tribunal
model arising out of misconduct investigations
involving personal injury.
Under the standard tribunal approach to
research misconduct investigations, cases
involving personal injury are much more likely
to produce a post-investigation proceeding. We
speculate that the tribunal approach frustrates the
ability of personally injured complainants to seek
redress. From the lofty perspective of protecting
the integrity of the scientific record, personal
injury cases may often appear trivial or
unimportant and clouded by interpersonal
bickering that borders on the unprofessional.
Very often personal injury cases involved
intellectual misappropriation disputes between
students or junior faculty and senior faculty
members. In such cases, the administrators and
the members of the tribunal conducting the
investigation tend to be more the peers of the
respondent than the complainant. Complainants,
rightly or wrongly, often believe that the
investigation is biased toward the respondent and
that the tribunal procedures prevent them from
making the most effective cases against the
respondent.
ORI’s recent policy statement about treating
whistleblowers as witnesses will probably
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
increase the likelihood of a post-investigation
proceeding by giving complainants even less
standing than they previously held.
In some cases the external funder offers a
post-investigation appeals process including a
full due process hearing, for example, the
Departmental Appeals Board of the Department
of Health and Human Services. The existence of
this appeal mechanism may alter the conduct of
the original investigation, leading to fewer
post-investigation proceedings. The existence of
an external appeal mechanism may discourage
some institutions that might be tempted to bias a
research misconduct investigation toward an
outcome most favorable to the institution’s
reputation or financial interests; the possibility of
disclosure and/or reversal at an appeals hearing
could act as a check on such institutional
behavior.
Institutional attorneys may face conflicts of
interest when fair treatment of the parties to an
investigation is not perceived to be in the
institution’s interest. Legal representation of an
organization presents many potential ethical
pitfalls for attorneys, especially when conflicts
arise within an organization, as is the case when
a university must investigate a research
misconduct allegation against a faculty member
or student.
While most judges are attorneys, most
attorneys are not judges and most attorneys are
trained to act as advocates for their clients. Some
institutional attorneys may see their roles as
advocates for procedural fairness, but they also
understand that a finding of misconduct can carry
heavy financial and reputational consequences
for the university as well as the individual
respondent.
Moreover, any of the parties to a misconduct
investigation could become a potential litigant
against the university because of decisions made
during the case by university administrators.
Therefore there may be a strong tendency to act
as legal advisor to university administrators as
opposed to advocates for a fair and impartial
investigation.
In this research, it is difficult to determine
whether controversial actions by institutional
attorneys was a cause or consequence of postinvestigation proceedings, since the timelines
necessary to distinguish cause from effect are
often missing in the kinds of documents
reviewed. Also the frequency of such reports are
low, but this could arise from the confidentiality
of attorney-client communications as well as
from lack of incidents to report.
Caveats. Most reports of research
misconduct are from news stories in scientific or
trade magazines (Science, Nature, Chronicle of
Higher Education). Reliance on these sources
could introduce a possible reporting bias, since
only the most disputatious cases would be
considered news worthy. This reporting bias
could significantly affect the prevalence data
presented earlier, but probably would not have a
major effect on the results of the multivariate
analysis.
NIH/ORI reports on the outcomes of research
misconduct investigations were also a major
source of cases. NIH/ORI reports also contain
relatively few plagiarism/ownership cases, which
might tend to underestimate the number of
personal injury cases.
Some observers believe that the handling of
research misconduct cases has improved over
time. The results of this study found a slight and
statistically insignificant temporal decline in the
number of cases resulting in post-investigation
proceedings. However, this decline was
confounded by a concurrent decline in the
number of cases reported over time. Because the
cases presented here were identified from the
scientific news literature, this latter decline could
be a function of either fewer cases (better
management) or less reporting (declining
newsworthiness) or both. A separate study based
on a fixed baseline of research misconduct
allegations in the institutions in which they arose
has been proposed to disentangle these
confounded effects.
Case References (Reference Identifier in
Sprague Database, citation)
5310. Marshall E. EPA faults classic lead poisoning study.
Science 1983; 222: 906-7.
6411. Weissman v. Freeman. D. New York 1989 Feb 23;
Nos. 225, 353.
6438 . Investigations, S. o. O. a. Fraud in NIH grant
programs. 1988.
6439 . Crewdson J. Science on the hot seat. Chicago
Tribune 1989 Mar 19; Sect. 5, 9, 13.
6483 . Dakins D. Copyright lawsuit illuminates debate over
academic standards. Diagnostic Imaging 1989 May;
54-60.
154
––––––––––––––––––––––––––––––––––––––––– Hogan et al., The Relative Efficiency of Research Misconduct Investigations
6484. Mervis J. Bitter suit over research work asks “Who
deserves the credit?’. The Scientist 1989; 1, 4.
6486. Blum D. Yeshiva professor wins copyright lawsuit.
Chronicle of Higher Education 1989 Apr 5; A14.
6494. Raymond C. Institutes of Health find that Purdue
professor plagiarized experiments. Chronicle of
Higher Education 1989 Jul 12; A2.
6496. Marshall E. USGS reports a fraud. Science 1989;
244: 1436.
6512. Barinaga M. NIMH assigns blame for tainted studies.
Science 1989; 245: 812.
6519 . Gupta’s defense. Science 1989; 245: 1192.
6521. Erben H. Carelessness, or good faith? Science 1989;
245: 1165-6.
6523. Coughlin EK. Geologist accused of misidentifying
fossils calls charges ‘misleading’. Chronicle of
Higher Education 1989 Sep 13; A9.
6532. Raymond C. Allegations of plagiarism of scientific
manuscript raise concerns about ‘intellectual theft.’.
Chronicle of Higher Education 1989 Jul 19; A4, A7.
6551. Water T, Oliwenstein L. Science duds. Discover
1990: 42-44.
6565. Zurer P. NIH panel strips research of funding after
plagiarism review. Chemical & Engineering News
1989; 24-5.
6603. Greenberg DS. Squalor in academe Plagiarist gets
promoted, victim is out of her job. Science &
Government 1990, May 1; 1, 5-7. Gualtieri, C. T.
“Letters.” .
6622. Henderson J. When scientists fake it. American Way
1990; 56-62, 100-1.
6623. Heidi S. Weissmann v. Leonard M. Freeman. 684 Fed
Sup: 1248.
6629. Kaplan, R.J. (1989). Letter.
6636. Scandal upon scandal. Nature 1990; 343: 396.
6661. Gordon G. Lab aide sues U-M, ex boss. Detroit News
1990 May 6: 1A, 9A.
6689 . Culliton BJ. NIH misconduct probes draw legal
complaints. Science 1990; 249: 240-2.
6728. Hamilton DP. NIH sued over misconduct case.
Science 1990; 2419: 471.
6738. Twedt S. Federal study raps Pitt medical school. The
Pittsburgh Press 1990 Sep 9: Sect. A:1,7.
6744. Hallum J, Hadley S. NIH Office of Scientific
Integrity: Policies and procedures (Letter to the
Editor). Science 1990; 249: 1227-8.
6745. Weismer G. Orofacial impairment in Parkinson’s
disease (Letter to the Editor). Neurology 1990; 40:
191-3.
6747. Hamilton D. Science misconduct legalese thickens.
Science 1990; 249: 1102.
6812. Bradshaw RA, Adler AJ, et al. Bridges committee
procedures (Letter to the Editor). Science 1990; 250:
611.
6831. Pinney GW. ‘Whistle-blowers’ are encouraged to
come forward at ‘U,’ dean says. Minneapolis Star
Tribune 1990 Oct 9: 4B.
6832. FDA investigating doctor at ‘U’ who led drug study.
Minneapolis Star Tribune 1990 Dec 9; 4B.
6842. Hamilton DP. NIH misconduct procedures derailed.
Science 1991; 251: 152-3.
6874. Cordes C. Professor asserts Stanford lied to U.S.
about misconduct probe. Chronicle of Higher
Education 1991 Mar 13; A21, A24-A25.
6920. Hallum JV, Hadley SW. NIH misconduct procedures:
Effect of new ruling (Letter to the Editor). Science
1991; 251: 1296.
6945. Even misconduct trials should be fair. Nature 1991;
350: 259-260.
6990. Rooney P. UI researcher steps down after admitting
plagiarism. News Gazette 1991 Apr 11; A1, A10.
6991. Rooney P. Colleagues see pressure in plagiarism case.
News Gazette 1991 Apr 12; A3. Ames, Rusell
University of Illinois 19913
7057. London R. Judge strikes rules on scientific fraud
inquiries. New York Times 1991 May 3; Section 2,
11.
7079. Weiss T. Too many scientists who ‘blow the whistle’
end up losing their jobs and careers. Chronicle of
Higher Education 1991 Jun 26: A36.
7090. Monson N. Misconduct in the labs. Health Watch
1991: 24-33.
7097. Wheeler D. U.S. has barred grants to 6 scientists in
past 2 years. Chronicle of Higher Education 1991;
A1, A6-A7.
7216. Wiernik J. Suits charge plagiarism in research. Ann
Arbor News 1991 Sep 18; B1- B2.
7279. Woolf P. Fraud in science: How much, how serious?
1989 Hastings Center Report 11, No. 5: 9-14.
7284. Blakely E, Poling A, et al. Fraud, fakery, and
fudging. : 313-330.
7327. Rennie D. The Cantekin affair. JAMA 1991; 266:
3333-7.
7328. Winslow R. Earache studies lead to painful dispute.
Wall Street Journal 1991 Dec 18; B1, B3.
7329. Lead researcher under investigation. Science 1991;
254: 1575.
7486. Ernhart CB, Scarr S. Lead challenge (Letter to the
Editor). Science 1992; 255: 783.
7501. Moore JA. Cases of misconduct related to speech and
hearing science. Paper for HSS 1991; 494/Psych
493.
7551. Roberts L. Misconduct: Caltech’s trial by fire.
Science 1991; 253: 1344-7.
7584. Wheeler DL. Scientists begin to question
confidentially of fraud investigations. Chronicle of
Higher Education 1992; A1, A8, A11.
7632. Woolf PK. Deception in scientific research.
Jurimetrics Journal 1988; 29: 67-95.
7673. Adler T. Psychologists criticize study of low lead
levels. Monitor 1992 Mar; 40-1.
7689. Woolf PK. Accountability and responsibility in
research. Journal of Business Ethics 1991; 10:
595-600.
7757. Rigert J, Lerner M. ‘U’ researcher could face federal
charge. Star Tribune 1992 Aug 22; 1A, 4A-1A, 5A.
8020. Anderson M. Impostors in the temple. New York:
Simon & Schuster 1992.
155
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
8070. Rigert J, Lerner M. In courting industry, ‘U’ has
invited trouble.” Minneapolis Star Tribune 1992 Dec
31; 1A, 14A.
8186. Scarr S, Ernhart C. Of whistleblowers, investigators,
and judges. Ethics & Behavior 1993; 3: 73-93.
8187. Ernhart CB, Scarr S. On being a whistleblower.
Ethics & Behavior. 1993; 3: 199-206.
8197. Minnesota psychiatrist accused of fraud in study.
News Gazette 1993 Feb 17: D-2.
8217. Rigert J. Top child psychiatrist at ‘U’ indicted on
charges of fraud. Minneapolis Star Tribune 1993 Feb
17; 1A, 11A.
8218. Sweeney P. U professor faces charges on drug study
irregularities. Saint Paul Pioneer Press 1993 Feb 17;
1A, 6A.
8259. Cordes C. Research project on ear infection
dramatizes challenge of conflicts. Chronicle of
Higher Education 1993 Mar 3: A23, A28.
8262. McCuen RH. Institutional oversight to enhance
integrity in research. Journal of Professional Issues
in Engineering 1990; 116: 27-37.
8298. Penn S. Misconduct in the modern university. Thistle
1992; 1, 7.
8299. Loh P. Vest sued for coverup of scientific
misconduct. Thistle 1992; 1, 8-1,11.
8323. Zorza M. “Letter.” 1993 May 10.
8419. Cotton P. Harassment linked to fraud, lawsuit alleges.
JAMA 1992; 267: 783-784.
8439. Burd S. Shalala’s criticism of U.S. polices on
scientific misconduct troubles some lawmakers and
watchdog groups. Chronicle of Higher Education
1993; A31, A34.
8478. Lueders B. Hunger strike feeds UW controversy.
Isthmus 1993; 5, 7.
8479. Kerr S. Feds freeze fraud-finders A Wisconsin
connection? Shepherd Express 1993 Jun 17-24;. 3.
8501 . Leo A. Paquette, Ph.D., Ohio State University. NIH
Guide for Grants and Contracts 1993; 2-3.
8502 . James H. Freisheim, Ph.D., Medical College of
Ohio. NIH Guide for Grants and Contracts 1993;
1-2.
8506 . Mark M. Kowalski, M.D., Ph.D., Dana Farber
Cancer Institute and Harvard University. NIH Guide
for Grants and Contracts 1993; 2.
8523 . Burd S. New policy of naming scientists who are
found guilty of fraud renews debate on federal role.
Chronicle of Higher Education 1993 Jun 30; A24,
A25.
8549 . Final findings of scientific misconduct. NIH Guide
for Grants and Contracts 1993; 2-3.
8585 . Holland E. Committee finds validity in allegations
against Paquette. On Campus 1993 Aug 19.
8586 . Lore D. OSU hunts core of plagiarism. Columbus
Dispatch 1993 Aug 23; C1.
8587 . Cleverley M. Chemistry professor has student
support. The Lantern 1993 Aug 26.
8604 . Polsby M. NIH spasmodic torticollis. Bethesda,
MD: National Institutes of Health. 1993 Jul 30.
8606 . Polsby M. “Letter.” 1993 Aug 20.
8607 . Rosen JE. “Letter.” 1993 Jun 26.
8608 .
8609 .
8611 .
8612 .
8627 .
Rosen JE. Statement of Dr. Jane Rosen. Undated.
McCutchen CW. “Letter.” 1993 Aug 24.
Rosen JE. “Letter.” 1993 Aug 24.
Rosen JE. “Letter.” 1993 Aug 12.
Taubes G. Fraud busters. Review of Academic Life
1993; 47-53.
8648 . Grassley CE, Cohen WS. “Letter.” 1993 Sep 23.
8791 . Hilts PJ. U.S. scientist says superiors stole credit for
her vaccine. New York Times 1993 Dec 18; 1010.
8841 . Nimmons D. Sex and the brain. Discover 1994;
64-71.
8902 . Butler K, Ohland G. Making the case against
discrimination. LA Weekly 1991 Jan 18.
8915 . Anderson D, Attrup L, Axelsen N, Riis P. Scientific
dishonesty & good scientific practice. Copenhagen:
Danish Medical Research Council 1992 .
9097 . Lock S, Wells F, et al., Eds. Fraud and misconduct
in medical research. 1993.
9185 . Immunologist fined $5,000 in lab misdeed. New York
Times 1994 Aug 31; A16.
9186 . Final findings of scientific misconduct. NIH Guide
for Grants and Contracts 1994; 23: 2.
9417 . Findings of scientific misconduct. NIH Guide 1995.
9542 . Hilts PJ. A university and 4 scientists must pay for
pilfered work. New York Times 1995 May 19; A20.
9568 . Bauer HH. Ethics in science. [Internet] Discussion of
Fraud in Science; 1995 May 1.
9603 . Taubes G. Plagiarism suit wins: Experts hope it
won’t set a trend. Science 1995; 268: 1125.
9614. Herling J. Embarrassment in Hong Kong. Chronicle
of Higher Education 1995 Mar 24; A43-A44.
9616 . Parrish CJ. The Lumigen Project - or how Wayne
State University went into business and lost its shirt
and a lot more! Newsbriefs Wayne State University
1994 Sep 7;. 1-7.
9617 . Shellum B. Chemical conspiracy. Detroit Free Press
1995 Apr 7; 1E-2E.
9618 . Wilson M. Judge: WSU chemist stole major
discovery. The Detroit News 1994 Aug 17; 1A, 6A.
9636 . Lytton H. Frequency of FFP. [Internet] 1995 Jun 11.
9642 . Kytle R. HHS release. [Internet] 1995 Jun 23.
9649 . Holden C. More on genes and homosexuality.
Science 1995; 268: 1571.
9654 . Crewdson J. ‘Gay gene’ study comes under fire.
Chicago Tribune 1995 Jun 25, 1, 6-1, 7.
9662 . Benditt J. Conduct in science. Science 1995; 268:
1705-1718.
9684 . Marshall E. NIH’s “Gay gene” study questioned.
Science 1995 Jun 30; 268: 1841.
9707 . Findings of scientific misconduct. NIH Guide 1195;
24.
9720 . Integrity, O. o. R. Annual Report. Department of
Health and Human Services 1994 Sep.
9727 . Researcher assessed damages for distribution of a
cell line. ORI Newsletter 1994; 3: 1-2.
9739. Burd S. Nursing professor found guilty of scientific
misconduct. Chronicle of Higher Education 1995
July 7; A27.
9757. Bivens LW. Findings of scientific misconduct. Office
of Research Integrity. Notice 1993 June 21.
156
––––––––––––––––––––––––––––––––––––––––– Hogan et al., The Relative Efficiency of Research Misconduct Investigations
9784 . Wilchek M. [Letter to the Editor]. Science 1995;
269: 1034.
9812 . Office of Research Integrity, Annual Report 1994.
1995.
9826 . Findings of scientific misconduct. NIH Guide 1995;
24, No. 33.
9854 . Findings of scientific misconduct. NIH Guide 1995;
24, No. 34.
9890 . ORI/AAAS Conference plagiarism and theft of
ideas. 1993.
9959 . Case summaries. ORI Newsletter 1996; 4, No. 2: 5.
9963 . Findings of scientific misconduct. NIH Guide 1996;
25, No. 15.
9972 . Findings of scientific misconduct. NIH Guide 1996;
25, No. 13.
9974 . Miller JP. AIDS research scientific fraud. [Internet]
1996 Apr 30.
9975. Findings of scientific misconduct. NIH Guide 1996;
25, No. 13.
9976. Findings of scientific misconduct. NIH Guide 1996;
25, No. 10.
9983 . Jayaraman KS. Court allows fossil fraudster to
return. Nature 1996; 380: 570.
10003 . Cherry M. Black scientist faces inquiry at South
African university. Nature 378: 324.
10079. Findings of scientific misconduct. NIH Guide 1996;
25, No. 22.
10084. Alan L. Landag. ORI Newsletter 1995; 4, No. 1: 5.
10110 . King RT. How a drug firm paid for University
study, then undermined it. Wall Street Journal 1996
Apr 25.
10116 . Jock F. MIT torn by bitter dispute over missile.
Science 1996; 271: 1050-2.
10153 . Findings of scientific misconduct. NIH Guide 1996;
25, No. 22.
10154 . Findings of scientific misconduct. NIH Guide 1993;
25, No. 22.
10163 . Gail L. Daubert. ORI Newsletter 1996; 4: 3-4.
10164 . Jamal Z. Farooqui. ORI Newsletter 1996; 4: 4.
10165 . Andrew Friedman. ORI Newsletter 1996; 4: 4-5.
10166 . Joan Gans. ORI Newsletter 1996; 4: 5.
10212 . Hilts PJ. Dispute on gulf war missile continues to
fester at MIT. New York Times 1996 Feb 24; 11.
10229 . Zinberg D. A cautionary tale. Science 1996; 273:
411.
10260 . Findings of scientific misconduct. NIH Guide 1996;
25, No. 27.
10358 . Garber P. MCG looking at changes admid research
investigation. Augusta Chronicle 1996, October 5.
10463 . Rex D. International recruitment highlights need to
track scientific behavior. Nature 1996, September
12; 383: 107-108.
10605 . Vipin Kumar. ORI Newsletter 1996; 6-7.
10693 . Findings of scientific misconduct. NIH Guide 1997;
26, No. 15. Findings of scientific misconduct. NIH
Guide 1997, May 9; 26, No. 15.
10719 . David F. Eierman. ORI Newsletter 1995; 3.
10724 . Yahya Abdlahi. ORI Newsletter 1996; 5.
10772 . Ann Marie Huelskamp, M.H.S. Johns Hopkins
University School of Medicine. ORI Newsletter
1997; 5.
10789 . Zalewski D. Do corporate dollars strangle scientific
research? Lingua Franca 1997; 51-59q.
10802. Findings of scientific misconduct. NIH Guide 1997;
26, No. 22.
12345 . Mazur A. The experience of universities in
handling allegations of fraud or misconduct in
research. Project on Scientific Fraud and
Misconduct: Report on Workshop Two, AAAS 1989.
157
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Appendix A
DATA COLLECTION SHEET FOR ORI ABSTRACT RESEARCH PROJECTS
CASE NAME________________________
DOCUMENT NO.__________
TYPE OF MISCONDUCT ALLEGED:
(check all that apply)
Fabrication/Falsification
______
Misappropriation of Intellectual Property ______
Other Serious Deviations
______
Retaliation
______
Other:___________________________________
NATURE OF INJURY
(Is there an injured party known to the alleged misconductor?)
Personal Injury _____ Injury to the Scientific Record ______
TYPE OF INSTITUTIONAL INVESTIGATION
(Is the complainant a witness or a prosecutor?)
Tribunal ______
Adversarial______
OTHER PROCEEDINGS CONSEQUENT TO THE INSTITUTIONAL INVESTIGATION
Internal (grievances, discrimination complaints, etc.)
If yes, was due process an issue?
External:Lawsuits
If yes, was due process an issue?
Complaints/Appeals to administrative agencies
If yes, was due process an issue?
Complaints/Appeals to legislative bodies
If yes, was due process an issue?
____
____
____
____
____
____
____
____
ROLE OF INSTITUTIONAL LEGAL COUNSEL
As regards the following, was there any evidence as regards the role of institutional legal counsel as being
(S)upportive, (N)eutral, (O)bstructive or (U)nknown of the procedural fairness of the institutional
investigation? (circle one in each line)
Selection or preparation of witnesses:
Selection or preparation of panelists:
Selection or preparation of administrators:
Handling administrative problems/complaints:
Issues of attorney-client privilege:
Providing or withholding information:
Application of legal indemnification:
Deliberation or making findings:
Preparation or editing of reports:
Protection of parties’ due process rights:
S
S
S
S
S
S
S
S
S
S
158
N
N
N
N
N
N
N
N
N
N
O
O
O
O
O
O
O
O
O
O
U
U
U
U
U
U
U
U
U
U
Ethical Evaluation of Misconduct Cases
Doric Little, School of Medicine, University of Hawaii, Honolulu, USA.
Martin D. Rayner, Pacific Biomedical Research Center, University of Hawaii, Honolulu,
USA
Key Words: Ethical evaluation, Misconduct, Scientific divorce
The policies governing the actions of the Ethics Committee at the University of Hawaii were
developed during the late 80’s when the dominant paradigm for Ethics investigations was the
“whistleblower” model. In this model a person of relatively low power in the academic hierarchy
complains of scientific or ethical misconduct perpetrated by a person of higher rank and/or power,
typically within their own academic unit.
For such cases to be handled in an appropriate manner (and to ensure that whistleblowers feel free
to come forward) the confidentiality of the complainant must be carefully protected. Administrative
procedures should minimise the chances that the accused person can use his/her academic power: a)
to have the complaint disregarded without adequate investigation and/or, b) to instigate reprisals
against the whistleblower. However, innocent faculty also need to be protected from frivolous or
malicious complaints. Thus, an initial Inquiry (Phase 1) was required, during which the existence of
the complaint is withheld from the accused, with the accused being informed and interviewed only
after the complainant has convinced the Review Panel that a thorough investigation is justified. At
that point, a full Investigation (Phase 2) is initiated, the accused is informed of the complaint while
his/her lab notebooks, computer files and other pertinent sources of information are immediately
sequestered. The accused then has the opportunity to present detailed rebuttal. If the evidence in
support of this rebuttal seems inadequate, then the committee so reports to the Administration and a
more formal Phase 3 Hearing is set up. It is only after the innocence of the accused has been
reasonably established (typically following the completion of Phase 2) that more difficult issues may
be considered, such as the possibility that the complaint was motivated by envy or by malice.
Furthermore, to conclude that the complaint is malicious requires the committee to assess the
motivations of the accuser at the time the accusation was made. Thus, even if strong suspicions exist,
it is not likely that sufficient evidence will be uncovered to confirm suspicions of malicious intent.
Despite the even-handed principles involved in this approach, the Inquiry Phase of such
investigations is necessarily limited to evidence provided by the complainant. And, more
significantly, both Phase 1 and Phase 2 primarily address the guilt or innocence of the accused. While
we understand that this sharp focus is appropriate in some situations, our experience suggests that this
is not necessarily a “one size fits all” model. This committee has experienced scientific misconduct
cases in which this approach prevented a fair and balanced Inquiry. We suggest that specific
circumstances exist in which policies based on this model may need to be modified to ensure an
appropriately ethical analysis of the complainant’s case.
Corresponding author: Martin D. Rayner, Interim Director, Pacific Biomedical Research Center, 1993 East-West Rd.,
University of Hawaii, Honolulu, HI 96822, 808-956-5184 (voice), 808-956-9574 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Year
Complaint
Outcome
92
92
93
93
94
95
96
96
97
98
99
99
99
99
00
Intel. Prop. Theft
Plagiarism
Plagiarism
Admin. Miscon.
Plagiarism
Authorship
Intel. Prop. Theft
Intel. Prop. Theft
Intel. Prop. Theft
Misapp. of funds
Theft/fabrication
Intel. Prop. Theft
Intel. Prop. Theft
Sci. Miscond
Hum Subj. issue
Sustained
Sustained
Dismissed
Dismissed
Sustained
Dismissed
Dismissed
Dismissed
Negotiated
Reimbursed
Dismissed
Dismissed
Sustained
Sustained
Sustained
Whistleblower
protections
Required
Required
Required
No
Required
No
No
No
No
Required
No
No
No
No
No
Collaboration
Breakdown
No
No
No
No
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
$$ issues
No
No
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
No
Table 1. Analysis of cases presented to the University Ethics Committee from 1992 to 2000
Results
Despite the many high-profile cases, nationally,
which seemed to fit the whistleblower model
during the 80’s and early 90’s, we have noted
significant changes in the nature of the
complaints coming before our committee over
the last five years (see Table 1). As shown in this
Table, six of the nine cases occurring after 1995
involved issues of intellectual property. Before
this time, however, only one case out of six
involved a clear intellectual property dispute.
Seven out of the nine cases since 1995, but only
one out of the six earlier cases, involved
breakdowns in scientific collaborations.
Similarly, five out of the nine post-1995 cases
involved high financial stakes, whereas none of
the earlier cases seem to have been primarily
motivated by financial considerations. Finally,
whereas four out of the six early cases required
whistleblower protections to protect the identity
of a junior complainant, only one complaint out
of nine cases since 1995 benefited from such
protections. Thus, whistleblower protections are
still needed, although cases that fit that specific
model are no longer a major part of our
workload.
Discussion
Ethics Evaluations in A Changing World
Two nation-wide trends may well have been
responsible for these changing patterns. First,
changes in funding patterns have increased the
payoff for collaborations between potentially
competing laboratories. Second, as scientific
information has become increasingly regarded as
potentially marketable intellectual property, it is
inevitable that disputes will arise as to the
ownership of that property. The stakes are further
raised when University Administrators suggest
that returns to research units from the marketing
of such intellectual property should become a
significant component of the budgets of
academic research units. In apparent response to
these trends, our recent cases have been
motivated primarily by disputes over the
ownership of potentially valuable intellectual
property. These situations are not consistent with
the whistleblower model on which our Ethics
policies and procedures are based - making them
difficult to evaluate. However, these cases cannot
be dismissed as being merely “authorship
disputes” beneath the level of interest of those
whose duty it is to evaluate true scientific
misconduct issues, in view of the very high
stakes which may be involved. Finally, we have
seen such cases start at the level of an authorship
dispute, only to later expand into full-scale
accusations of data fabrication.
Nevertheless, our university’s policies as
well as the general awareness of the scientific
community remain tuned to the whistleblower
160
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Little & Rayner, Ethical Evaluation of Misconduct Cases
model. So, as one might well expect, our cases
continue to be presented in the “approved”
whistleblower format, promising to reveal
significant instances of scientific misconduct.
If one fails to understand their origins, such
cases can be difficult to evaluate. In one such
instance we were unable even to conclude that a
valid case existed under our restrictive rules for
Phase 1 Inquiries. What does one do when Phase
1 of a “denial of authorship” complaint leads to
the complainant eventually submitting letters
from the accused in which the accused pleads
with the complainant to accept authorship on the
paper in question? Should the accused have been
interviewed during Phase 1, in this case, so as to
gain additional understanding of the background
against which the complaint was made? The
initial decision that there was no case to pursue,
precipitated a seemingly endless series of
requests for external intervention and/or reevaluation of our committee’s policies. We need
to do better than that.
Similarly, other recent cases before our
committee have seemed to involve inherent
conflicts between the superficial appearance and
the underlying realities of each case. The stage
now seems set for continuing problems arising,
in part, from our evaluative approaches. Perhaps,
significant changes should be proposed in both
the published procedures and investigative
approaches so as to permit effective evaluation of
cases that do not fit the whistleblower paradigm.
However, these cases raise arguments for
modifications of our procedures that might, if
implemented, remove key protections for more
classic whistleblowers.
This seems a potentially dangerous situation
in which it would be all too easy for university
faculties and administrations to make serious
mistakes while acting from the highest ethical
motivations. To address these concerns recent
cases have been re-evaluated to search for
potentially generalizable patterns within what
had seemed to be “property disputes”. Such a
pattern could provide the theoretical grounding
from which a more systematic approach could be
developed towards this different class of
misconduct complaints.
Excluding situations involving “priority of
discovery” issues, or situations of outright theft
(none of which we have yet seen), when two
groups feel that they both have valid claims to
some piece of the same pie this is probably a pie
they baked together. In other words, the majority
of such disputes seem to arise from the
breakdown of formerly effective collaborations.
And, since most collaborations collapse from
personality conflicts, it is hardly surprising that
such breakdowns lead to disputes over the
custody of intellectual property. The comparison
with that other graveyard of failed collaborations,
the divorce courts, is inescapable. The level of
acrimony over rights to intellectual property
seems fuelled by these underlying personal
issues, just as rights to child custody may
become the focus of a parent’s sense of violation
in a divorce situation. An Ethics Committee that
must stick its head into a “scientific divorce”
needs to be well aware just how high the
emotional stakes may have become for the
individual contestants regardless of the monetary
worth of the objective data.
The committee will need to remember that
not all fights are about money. Some fights are
incomprehensible from any other motive than to
humiliate the opponent. And they will need to
recognise that when it takes at least two people to
bake such a pie, it often takes two to spill it on
the floor. Of course, the participants in this
“divorce” may not have behaved equally badly,
but the party most wronged is not necessarily the
one who complains the most loudly. This is
dangerous territory for an investigative
committee, where the most fundamental
assumptions of the whistleblower model may no
longer be valid.
Formulating a working hypothesis
The essence of the issue is this: whereas the
whistleblower model appropriately evaluates the
validity of the complaint, in a “scientific divorce”
it cannot be assumed that the substance of the
complaint is valid. Furthermore, it was clear that
our case load in Hawaii would not be sufficient
to permit even a minimally rigorous prospective
study of such cases - which is why we are
presenting our ideas to this meeting. If analysis
of our experience resonates with the experience
of other similar committees, perhaps they will
also take up this issue.
“Scientific divorces” may need to be
evaluated by different procedures. In these cases
one should not focus on the guilt or innocence of
the accused, but rather survey the ethical
landscape in which the breakdown of
collaboration occurred. Specifically, it is not
appropriate to assume that the complaint is valid
or that the complainant is not a material
161
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
contributor to the situation under investigation.
To support this approach, the preliminary
instructions given to our Review Panels were
changed. When the initial complaint indicated
that either an intellectual property dispute, or a
breakdown in collaboration, was involved, it was
suggested that both the complainant and the
“accused” needed to be interviewed during Phase
1. In other words, it may be impossible to
determine whether or not misconduct is likely to
have occurred unless both parties are
interviewed. In a situation of this kind, however,
the committee needs to be aware that the
complainant will have had time to purge any files
that might prove embarrassing, although the
accused may well have been taken by surprise.
Additionally, even in Phase 2 of the
investigation, we suggested that the Review
Panel delay considering whether the accused
might be guilty or innocent of misconduct. First,
they should focus their attention on a different
question: “What happened to create the present
conflict?”. However, they should be prepared to
take as much detailed testimony as necessary to
answer that very simple question. Only when the
committee has reached a clear consensus as to
“what happened”, should they attempt to
consider which actions taken by each participant
might rise to the level of scientific misconduct.
The danger here is that such open-ended
investigation can get out of hand – the Chair of
the Review Panel may need to remind its
members that focus should be maintained on
immediately relevant events.
These instructions appear to have
substantially facilitated the appropriate ethical
evaluation of difficult cases. Our Review Panels
have been models of good committee interactions
where all decisions have been unanimous
following considerable discussion but without
significant disputes. This surprising degree of
agreement resulted from a comprehensive
consensus as to “what really happened” –
committee members have all felt comfortable
that “blame”, where blame has been needed, was
fairly assigned. Finally, shared understanding of
the underlying issues allowed them to make
some very tough calls in potentially explosive
cases. Even in these hard cases, committees
appear to have appropriately surveyed each
situation without bias and to have resolved the
issue appropriately.
Next steps
The most effective method needs to be explored
by which to merge this “Ethical Landscape
model” into policies written to protect
whistleblowers. We would like to avoid a triaging
mechanism which would separate cases into, for
example: intellectual property cases, misconduct
cases and “harm/rights” cases with different
guidelines (as in the the separate courts of our
legal system). Instead, we have hoped to find
some way to treat all our cases from an ethical
perspective, while at the same time preserving
our protections for whistleblowers. We now
believe that ALL cases can be addressed from
this ethical approach in which we do not ask “is
the accused guilty?” but instead ask “what really
happened?” Once the Panel can answer that
question, then they can consider the extent to
which each participant has behaved in an ethical
or unethical manner - and we are ready to ask
whether any of these behaviors rise to the level of
scientific misconduct. By contrast, Phase 3 of
the investigation (when this is necessary), should
be the point at which standard legal models are
introduced.
Fortunately, only one small change in our
policies is required to implement this approach.
The Review Panel needs the discretion to
interview the accused during Phase 1, should
they conclude that this can be carried out without
threat to the complainant. Given that freedom,
the Panel can then adopt either the “standard”
approach to Phase 1, or the “ethical landscape”
approach, as seems most fitting to the case under
investigation.
Nevertheless, the open-ended investigational
approach advocated here can lead to unusual
situations. For example, in one recent case the
Committee’s final report to the University
Administration recommended censure not only
for the accused but also for the complainant
(whose actions contributed to the wrongdoing),
as well as for a third party who facilitated the
situation to his own benefit. To have reported
only on the guilt of the accused would have
seemed a violation of our Committee’s ethical
duty in this instance.
Acknowledgements
We acknowledge receipt of travel funds from
Alan Teramura, Senior Vice-President for
Research at the University of Hawaii.
162
Potential Cultural Factors In Scientific Misconduct Allegations
Walter M. Meyer, III, Department of Psychiatry and Behavioral Sciences, The University of
Texas Medical Branch, Galveston, USA
George M. Bernier, Jr., Department of Internal Medicine, The University of Texas Medical
Branch, Galveston, USA
Key words: Authorship disputes, Ethnicity, Gender, Scientific misconduct
Since 1993, The University of Texas Medical Branch has had 16 allegations of scientific misconduct.
They were each examined carefully during an inquiry by a faculty committee and the scientific
integrity officer for evidence of fabrication, falsification or plagiarism. Only one of them was judged
to be scientific misconduct. It involved plagiarism, which was acknowledged by the respondent, and
this case will not be discussed further in this document. The remaining 15 allegations did not reach
the stage of investigation. They involved a variety of other types of complaints: an authorship dispute
in 4 cases, inadequate sharing of data in 3 cases or allegations of questionable research practices in
the remainder. Since many of these disputes involved individuals who were not born in North
America and were raised in different cultural settings, the authors hypothesized that cultural factors
underlie many of these allegations. In order to examine this question, they have done a retrospective
review of the 15 allegations.
Methods
A retrospective review of these 15 allegations was done to detect the possible involvement of gender,
academic status, ethnic factors or cultural concerns. To determine whether any ethnic or cultural
group appeared to be overly represented as complainant or respondent, the cultural/ethnic background
status of the entire faculty, post-doctoral fellows and research technical personnel was compared to
those involved in these allegations.
Results
The 15 complaints involved 29 people; 13 White (10 European descent, 3 Middle Eastern descent),
one African American and 15 Asians (9 Indians and 6 Chinese). See Table I for ethnic distribution of
the complainants and respondents. One of the Indians was involved in two separate instances, once as
a respondent and once as a complainant. All the Asians were born and raised outside of the United
States. Six of the complainants were White (4 European descent, 2 Middle Eastern descent) and 3 of
these were born and raised outside of North America. Seven of the respondents were White (5
European descent, 2 Middle Eastern) and two were born outside of North America. The one African
American individual, born in the United States, was a respondent. Nine Asians (4 Chinese and 5
Indians) were complainants and 7 Asians (2 Chinese and 5 Indians) were respondents.
Corresponding author: Walter J. Meyer III, Department of Psychiatry and Behavioral Sciences, University of Texas Medical
Branch, Galveston, TX, 77555, 409-747-8355 (voice), 409-747-8351 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Complainants
White US born
White Foreign born
Asian, Indian
Asian, Chinese
African American
Total
White,
US
1
0
2
0
0
3
White,
Foreign
2
1
0
0
0
3
Asian,
Indian
1*
0
3*
0
1
5
Asian,
Chinese
1
1
0
2
0
4
African
American
0
0
0
0
0
0
Total
5
2
5
2
1
15
Table I: Number of complainants and respondents by ethnic group
* One person was a complainant and a respondent
Three subjects involved in these allegations
were technicians, seven were post-doctoral
fellows and the remaining 19 individuals were
faculty. One faculty was involved in two
allegations, once as a complainant and once as a
respondent. The complainants and the
respondents were of very similar ages, mean ages
of 45.7 and 44.0 years, respectively. In ten cases,
the complainants were older than the respondents
and in five they were younger. Ten of the
complainants were of lower rank in the university
than their respective respondents. Only five of
the 29 individuals were female (two Whites, two
Indians and one Chinese). These 5 women were
involved in a total of 3 allegations.
Six of the allegations involved individuals
from different ethnic groups. The remainder
involved individuals from the same ethnic or
cultural background. Of the six disputes
involving more than one ethnic group, three
involved White of European origin and Indians;
two, a White and Chinese; one, a African
American and an Indian. Nine disputes involved
individuals from the same ethnic group: two
involved Chinese; three involved Indians; and
four involved White. Among the disputes
involving White as both complainant and
respondent, one involved both parties being from
the Middle East; one involved both parties born
in the USA and of European descent; one
involved a complainant born in an eastern block
country and a respondent born in the USA; and
the last involved a foreign born middle eastern
(Egyptian) complainant and an American born
Israeli respondent. Two of the allegations
involving Asians referred to deep-seated distrust
of individuals from similar backgrounds in their
country of origins. In one instance, the
complainant stated that he knew that the
misconduct had occurred because people from
the village of the respondent were evil. In the
other instance, the complainant referred to the
political leanings of the respondent as they
related to their country of origin, i.e., brands of
communism.
To determine whether any ethnic or cultural
group appeared to be overly represented as
complainant or respondent, the cultural/ethnic
background status of the entire group of
university employees (faculty, bachelor level
technicians or post-doctoral fellow) was
compared to those involved in complaints. All
but one female professor was or had been
employees of the university. Only five of the
individuals were female (two Whites and three
Asians). The faculty is 24% female and 17% of
these allegations involve females.
There is a great difference in the ethnic
distribution of the total faculty compared to those
individuals involved in scientific misconduct
allegations. The medical school has a faculty of
750 individuals (550 White, 39 Hispanic, 24
African American and 136 Asian). Of the 136
Asian, at least 55 are from India and 43 are from
China. Table II illustrates the differences in
ethnic distributions between the faculty, bachelor
level research technicians and post-doctoral
fellows at large and those individuals involved in
scientific misconduct disputes. There is a
significant difference between the individuals
involved in scientific misconduct allegations and
the total group of individuals in the same
category for the faculty (p <.0001 by chi-square),
the technicians (p <.0001 by chi-square) and the
post-doctoral fellows (p <.001 by chi-square).
The country of origin was not discerned for the
faculty. But there does seem to be among the
White individuals an unexpectedly large number
of individuals born in the Middle East.
Discussion
In the early 1990’s many universities started
164
––––––––––––––––––––––––––––––––––––– Meyer & Bernier, Potential Cultural Factors in Scientific Misconduct Allegations
Total
Total
Total
Involved in
Scientific
Misconduct
Disputes
Faculty*
Technicians**
Postdoctoral ***
Faculty *
Technicians**
Postdoctoral ***
White
73.0
Nat. Am. Hispanic
3.2
5.2
56.6
40.0
52.6
33.0
28.6
4.6
3.0
5.3
0.0
0.0
9.8
4.0
0.0
0.0
0.0
Indian
0.5
Asian
18.1
Total
100
0.0
1.0
0.0
0.0
0.0
29.0
52.0
42.1
67.0
71.4
100
100
100
100
100
Table II: Differences expressed as percent of total in ethnic distributions between the faculty and postdoctoral fellows at
large and those individuals inovlved in scientific misconduct disputes
*Significantly different p <.0001 by Chi Square, **Significantly different p <.0001 by Chi Square
***Significantly different p <.001 by Chi Square
establishing a very formal process to consider
scientific misconduct charges. The initial
definitions were focused on fabrication,
falsification and plagiarism but did leave an
opening for examining ‘other practices that
seriously deviate from those that are commonly
accepted within the scientific community for
proposing, conducting or reporting research’ (so
called unusual or questionable practices) (1-3).
The allegations or complaints were usually none
of these; rather they reflected personal disputes
between the complainant and respondent.
Questionable research practices were particularly
difficult to define and often the scientific
integrity officer and/or relevant faculty
committee were called upon to make a judgment
of intent. Therefore these disputes were almost
always impossible to discern with any assurance
for fairness. In order to gain insight into these
types of complaints, a fairly large amount of
work has been done nationally to examine the
nature of the complaint. In fact, certain types of
complaints such as authorship complaints were
rejected as scientific misconduct. Also the Office
of Science and Technology Policy has
established, but not formally implemented, a
more narrowed definition to exclude questionable
research practices and to include with fabrication,
falsification, and plagiarism only the
inappropriate use of documents which might be
seen as part of the review process (4). Even with
this narrower definition the complaints about
authorship, data ownership and access and
questionable or sloppy research practices will
continue to plague the university committees and
scientific integrity officers.
In contrast to open discussion about the
nature of the complaints and allegations, almost
nothing has been written about the nature of
those who made the complaints or those who
were the target of the complaints. The little we
do know refers only to the respondents who have
been determined to have committed scientific
misconduct. We know little about those who
brought the complaint forward because of the
appropriate concern about damaging the
whistleblower. Also almost nothing has been
written about those allegations, which did not
meet the definition of scientific misconduct as
defined by falsification, fabrication, and
plagiarism. One study of authorship disputes
received at the Ombuds office of Harvard
Schools and affiliated hospitals reported that the
number of disputes has greatly increased between
1991-2 to 1996-7 (5). Women were involved in
the majority (53%) of the complaints and non-US
citizens were involved in 21% of them (5). The
current study seems to be the only other venture
into this area. This study identifies a higher than
expected number of individuals who were born,
raised and partially educated outside of the
United States. In addition, the complaints are
often against individuals from the same ethnic
background and gender as the complainant. This
data is provocative. If substantiated in other
universities, it indicates a need to reexamine our
education of faculty and post-doctoral fellows
concerning the proper use of the scientific
misconduct complaint process. Also other
mechanisms need to be identified to help settle
these misunderstandings among scientific
colleagues.
There are significant hazards to doing this
type of retrospective review. This type of
endeavor invites accusations of racism, gender
bias, and other un-American activities, such as
racial profiling. In order to get different
perspectives on this issue, the authors had the
Director of our Affirmative Action Office and a
member of our Institute of Medical Humanities
165
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
review this manuscript. We are attempting only
to describe as a group the complainants and
respondents, not to speculate why one group
rather than another might utilize the scientific
misconduct complaint process to address other
related issues in the research group setting. One
speaker at the recent ORI conference on research
(6) suggested that misconduct complaints are
increasing because of the increased collaborative
nature of research and increased difficulty in
obtaining funding. Only three of our allegations
involved collaborations outside of the
complainant’s research group. Four of our
allegations could be linked to some financial
factors but they did not seem to be the main
issue. Usually the complaint involved very poor
communication between the respective parties.
Some ground rules for working together need to
be taught as part of the research curriculum.
Conclusions
The vast majority of complaints did not involve
scientific misconduct as currently defined. This
retrospective review suggests that cultural
concerns may contribute to the complaints to the
scientific integrity office. Proportionally the
Asian group is over represented in the scientific
misconduct complaint process. This report
documents for one university the magnitude of
the apparent influence of cultural differences in
the scientific misconduct complaint process. On
the surface, this retrospective review suggests
that cultural differences account for many of the
authorship and other scientific misconduct
disputes. Since the vast majority of complaints
in this retrospective review did not involve
scientific misconduct as currently defined, we
believe there is a need for an increased
educational effort on the part of the university to
orient faculty, bachelor level research technicians
and post-doctoral fellows on the appropriate use
of the scientific misconduct process and to
develop other mechanisms to help them resolve
conflicts with fellow scientists. Guidelines for
data ownership and management (7), authorship
of grants, and authorship of papers (8) have been
recently established on our campus to aid in this
process.
Bibliography
1. U.S. Public Health Service. Responsibility of PHS
awardee and applicant institutions for dealing with and
reporting possible misconduct in science. 54 Federal
Register 151; 42 CRF Part 50 (8 August 1989), 3244651.
2. Panel on Scientific Responsibility and the Conduct of
Research. Responsible Science: Ensuring the Integrity
of the Research Process, I. Washington DC: National
Academy Press; 1992.
3. Dresser R. Defining scientific misconduct: the relevance
of mental state. JAMA 1993; 269: 895-97.
4. Office of Science and Technology Policy. Federal policy
on research misconduct. 65 Federal Register 235; (6
December 2000), 76260-64.
5. Wilcox LJ. Authorship: the coin of the realm, the source
of complaints. JAMA 1998; 280: 216-7.
6. Little D, Rayner M. Ethical evaluation of misconduct
cases. Research Conference on Research Integrity.
November 18-20, 2000.
7. Meyer WJ. Data Management: recording, retention,
access and ownership. In: David TP, editor.
Management of Biomedical Research Laboratories: A
National Conference Proceedings. Tucson, AZ:
University of Arizona Press; 1999. p. 173-78.
8. The University of Texas Medical Branch Web Page
accessed at http://www.utmb.edu/avpr/integrity/
memo.htm.
Acknowledgment
The authors wish to express their gratitude to
Deborah Reynolds for her help in preparing this
manuscript.
166
Whistleblowers in Environmental Science, Prevention of Suppression Bias,
and the Need for a Code of Protection*
Elihu Richter, Occupational and Environmental Medicine, Hebrew University School of
Public Health and Community Medicine, Jerusalem, Israel
Colin L. Soskolne, Department of Public Health Sciences, University of Alberta, Edmonton,
Canada
Joseph LaDou, International Center for Occupational Medicine. University of California
School of Medicine, San Francisco, USA
Tamar Berman, Occupational and Environmental Medicine, Hebrew University School of
Public Health and Community Medicine, Jerusalem, Israel
Keywords: Code of protection, Environmental science, Suppression bias, Whistleblowers
Suppression bias is the distortion in the estimate of findings on hazard and risk inimical to special or
national interests, and is well known (1-4). The direct and indirect repercussions of suppression bias
are issues of direct importance not only to environmental scientists and health and safety
professionals, but also to the public itself. These repercussions raise questions as to the adequacy and
degree of protection provided by professional organizations, research institutions, and the legal
system against such suppression bias.
Suppression bias is rooted in the way societies react to troublesome information, as we know
from the tradition of shooting the messenger of bad news. The trial of Socrates served as the classic
case study of the risks to messengers. The jurors of Athens, a city besieged from without and insecure
from within, convicted Socrates and sentenced him to death for corrupting the morals of the youths of
Athens (5-6). Legal scholars have pointed out that Socrates would be convicted by a modern jury for
the same reasons that he was convicted by the jury in Athens: his teachings undermined order,
stability, and state security. For Athenians, there was a Benthamite rationale for putting Socrates to
death: silencing him was necessary to preserve the greatest good for the greatest number in a society
weakened by external wars and internal divisions (7).
Environmental scientists and occupational health and safety professionals measure and report
health risks from exposures to toxic and physical agents so that preventive measures can be put into
* An earlier version of this paper was published in the Int Journ of Occ and Env Health, 7:68-71, 2001, based on a
presentation at the Collegium Ramazzini Annual Meeting in Carpi Italy on October 29 2000. We thank the Councilors and
members of the Committee on Ethics and Philosophy of the International Society for Environmental Epidemiology for
advice and encouragement, the International Student Foundation and the Collegium Ramazzini for financial support, and Dr
Herbert Levine (Bethesda MD), Mr Ron Dror (Tel Aviv) and Dr Richard Laster (Jerusalem) for their helpful insights.
Corresponding author: Elihu D. Richter, Hebrew University-Hadassah School of Public Health and Community Medicine,
POB 12272, Jerusalem, Israel, 972-2-6758147 (voice), 972-2-6784010 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
effect. We define epidemiologic messengers, or
whistleblowers, as persons who are subjected to
harassment, lawsuits, ostracism, job loss, loss of
funding, intimidation, abuse, threats, or even
force after reporting such risks, or are prevented
from investigating or reporting risks altogether.
In most scientific fields, the rewards go to
investigators who report «positive findings». But
in the environmental sciences, the situation is the
opposite. In environmental and occupational
medicine, and in epidemiology and related
disciplines, “positive” findings about hazards
and risks are threatening to powerful interests.
Investigators who study or report these risks are
therefore at increased risk for harassment by the
very nature of their work.
Ultimately, suppression of information about
hazards and their health risks may itself become
hazardous to public health. There has not been
sufficient recognition of the possibility that such
pressures may serve to deter investigation or
assessment of health risks from exposures, and
thereby delay or block the implementation of
preventive measures. So far, there have been few
systematic efforts to examine the impact of such
pressures on the direction, content, and work
output of environmental epidemiologists,
physicians in occupational medicine, and other
scientists. Nor has there been sufficient attention
as to how to respond to these pressures.
Methods
This paper reviews past reports and summarizes
work now being carried out by the ISEE
Committee on Philosophy and Ethics and the
Collegium Ramazzini. This work documents
episodes of harassment of environmental
scientists and episodes of responding to requests
for assistance from environmental messengers
subject to harassment. We also make
recommendations for future action by
governmental organizations, which define
standards for research policy.
Findings
In the 1980’s, the United States Environmental
Protection Agency (EPA) published a document
which described the hazards unique to
environmental scientists and the forms of
harassment to which they may be subject. It
made the point that harassment is most likely
directed at younger or less well-known scientists,
employees of government or industry, or
members of the exposed population itself in
settings where protection of human rights is
weak. However, information is not readily
available on the degree to which this or other
Federal agencies defined institutional
responsibilities to protect investigators from
external or internal harassment.
The context and content of the problem
Martin (8) has listed the five methods of
suppression bias. These are: (a) preventing
creation of data (b) controlling, (c) blocking,
(d) distorting data, and (e) attacking researchers.
This simple list shows that using harassment to
block dissemination of data on hazard and risk
and attacking researchers who report such
findings are only part of a syndrome of
suppression bias, leading to what is known as
lamppost toxicology or epidemiology. Martin and
Deyo have reviewed the driving forces, context
and methods of harassment of epidemiologic
messengers or whistleblowers, and have provided
case studies (1, 2, 8).
The reported distribution of the problem:
sentinel episodes
Does suppression bias deter the prompt
detection, reporting and prevention of hazard and
risk? If so, is this bias systematic, episodic, or
sporadic and what are its distributions and
determinants? The details of whistleblower
harassment are not frequently publicized (9), but
below we present a list of episodes that have
come to light in the past years from reports
gleaned from the professional and lay literature,
and from our own direct contacts.
Cases of suppression by a governmental
institution
• Cate Jenkins, an environmental scientist
with the US EPA, claimed that chemical
industry studies had consciously minimized
the hazard of dioxin (10-11). She received a
written reprimand for writing down what she
knew about the history of the dioxin incinerator regulations (12-13), and was transferred from her position.
• Omar Shafey, an epidemiologist in the
Florida State of Health, was forced to leave
his position after publishing an epidemiologic report on complaints of acute illness in
residents exposed to drift from aerial spraying of malathion, used to control the Medfly
168
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Richter, et al., Whistleblowers in Environmental Science
(14).
• Desi Mendoza Rivero, a Cuban physician,
was imprisoned after he issued statements
regarding an epidemic of dengue fever (15).
• Grigory Pasko and Alexander Nikitin,
government scientists in Eastern Europe,
were accused of treason and subjected to
physical abuse after they reported dangers
from nuclear waste in Murmansk (16-17).
From newspaper reports, it appears that
Pasko’s subsequent acquittal was reversed.
(17)
• Melvin Reuber, a toxicologist at the
Frederick Cancer Research Facility in
Maryland (which is part of US National
Cancer Institute) studied links between
pesticides and cancer. As a result of his
studies, he is one of the world’s leading
critics of pesticides. In 1981, he was subjected to an attack on his work and his
credibility that shattered his career (18-19).
• In the United Kingdom, a Health and Safety
Executive (HSE) memo indicates that
several researchers and health and safety
activists who exposed poor health and safety
practices were targeted for special surveillance (20).
Cases of suppression by an academic
institution
• John Coulter, a medical researcher at the
Institute of Medical and Veterinary Science
in Adelaide, South Australia was dismissed
from his post after releasing a report that
ethylene oxide was mutagenic (21).
• Robert van den Bosch of the University of
California, Charles Lincoln of the University
of Texas, and Robert Fleet of Texas A&M
University all suffered abuse because of their
research on the hazards of pesticides (22).
• David Kern, an occupational physician and
epidemiologist at Brown University Medical
School, received notice that his tenure would
not be renewed and his clinic closed after he
reported numerous cases of interstitial lung
disease in nylon flockers at Microfibres (23).
• In Israel, Dr Jerome Westin was greylisted
for any governmental or academic appointments after publishing findings on massive
contamination of the nationwide milk supply
with organochlorines (24).
Cases of suppression by industry
• In the 1940’s, Randolph Byers, the Harvard
pediatrician, was sued for defamation and
damages by the Lead Industries Association
for publishing findings on brain damage
from acute lead poisoning in children from
nibbling paint chips (25-26).
• Doug Johnson, a safety specialist for
Tatitlek, Chugach, and Chenega Corporation
in Alaska was fired after raising environmental concerns regarding Alyeska’s oil spill
response program in Prince William Sound
(27).
• Myron Mehlman, a Mobil Oil Corporation
toxicologist, was fired after advising a Mobil
subsidiary in Japan to stop selling gasoline
with hazardous levels of benzene, a known
carcinogen (28).
• Alexandra De Blas of Australia was threatened with a suit for defamation by a mining
company when she attempted to publish a
thesis about environmental impact of its
operations. (29).
• Dr Yoram Finkelstein, an Israeli
neurotoxicologist with important publications on organophosphates and lead, is
currently the target of a SLAPP (Strategic
Lawsuit against Public Protestors) lawsuit
for libel after writing a medical opinion on
the health risks from emissions of hexavalent
chromium, Cd, lead, Ni, and other pollutants
from an aluminum foundry (30).
Survey Results
At the Annual Conference of the International
Society for Environmental Epidemiology (ISEE)
held in 1999 in Greece, the Committee on Ethics
and Philosophy distributed a questionnaire to the
delegates. Out of 10 individuals who completed
the questionnaire, five reported harassment
following publication of research findings on
health risks from environmental exposures. The
following is a brief description of these cases:
• Male MD, age 47, a scientist in a major
Cancer Institute in Italy, experienced ostracism after publishing findings on asbestos
exposure in a petroleum refinery and lung
169
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
cancer.
• Female MD, MPH, age 60, was threatened
with loss of her job after publishing findings
on TCDD exposure and cancer.
• Male MPH, PhD., age 53, experienced
ostracism and the threat of job loss after
publishing findings on cancer mortality in
Vietnam veterans exposed to Agent Orange.
• Two Female MD, investigators age 59 and
47, experienced both ostracism and confiscation of data after publishing findings on
ethylene oxide exposure and breast cancer.
Pressures on institutions
Deyo et al have reviewed Congressional
harassment of the CDC Injury Prevention Unit
following its epidemiologic work on impact of
gun control laws on violent deaths (2).
Actions to date:
The International Society for Environmental
Epidemiology (ISEE) Committee on Ethics and
Philosophy and the Collegium Ramazzini
Committee to Protect Whistleblowers are
working in parallel to provide moral and
professional support to whistleblowers (31). The
ISEE has already developed procedures designed
to provide an international service of advice,
referral, and support for environmental
whistleblowers which was first presented in a
special workshop at the ISEE International
Conference in Athens in 1999 (not far from the
site where Socrates was convicted.) The
Collegium Ramazzini is now doing the same, and
is planning to expand media reporting of
whistleblower harassment, with particular
attention to occupational medicine professionals
in developing countries. The aim of both
professional societies is to establish systems for
monitoring and reporting harassment and abuse
of whistleblowers, and to offer support and
assistance should it be requested.
In 1996-97, before ISEE developed these
procedures, it reacted to two situations in which
investigators were subject to political pressures
resulting from the publication of their findings.
In the case of Dr. Herbert Needleman, ISSE sent
a petition signed by many of its members to the
University of Pittsburgh asking that its review of
the validity of his findings on the effects of low
level lead exposure on intelligence scores,
behavior, and mood status be insulated from
outside pressures and be governed by the criteria
used for peer review. In the second case,
Professor Viel from France reported to the Ethics
and Philosophy Committee being the target of
job threats following publication of papers in the
British Medical Journal on risks for cancer
among seashore residents living near nuclear
power plants. This investigator also reported
pressures from the nuclear industry to reveal the
identity of individuals whose health records were
part of an epidemiologic study. The Ethics and
Philosophy Committee convened an ad hoc
subgroup, under the late Professor John
Goldsmith, one of its founding members, which
communicated with Professor Viel, and offered
to provide moral support for the issues raised. In
both the Needleman and Viel cases, the issues of
concern were resolved, but it is not known
whether and to what degree ISEE’s response
played a role. Both Needleman and Viel are
well-known senior investigators who published
their work in prestigious journals. Their
situations are exceptions to the rule that most
whistleblowers do not have the protection of
status and seniority, their findings or warnings
may not be particularly original, and they may be
prevented from either from publishing their
findings or completing investigations in
progress.
Through 2001, ISEE has responded to two
cases, that of Yoram Finkelstein and Omar
Shafey, and is working on a third, that of a
pathologist sent to prison in Belarus.
Discussion
The case studies above provide support for the
hypothesis that powerful governmental, military,
economic, and political interests are often the
driving forces and the sources of legal and illegal
harassment of environmental messengers and, at
times, the institutions they work for. But most of
the case reports are from Western countries with
developed research cultures and codes for the
protection of human rights. The high-risk
settings for exposure to pressures against
environmental scientists are those where research
is most needed, i.e., where exposures and risks
are severe, where there are few environmental
scientists, and occupational safety and health is
not properly regulated and enforced by law. The
risks are increased where legal safeguards for
human rights are weak, and where access to a
free press is blocked.
Yet, data are not readily available to examine
170
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Richter, et al., Whistleblowers in Environmental Science
the working hypothesis that the exposure settings
in which scientists are at greatest risk for threats,
harassment, and legal pressure are those in which
they are most needed. Africa, Latin America,
Asia, the Mid-east and Eastern Europe are the
regions of the world with the worst
environmental and occupational health problems,
the fewest environmental scientists, and the
weakest safeguards to protect the rights of
investigators. The situation is probably the worst
for physicians working in occupational medicine
who serve remote populations, given their
relatively low status on the professional totem
pole. In many of these countries, the situation for
environmental scientists parallels the situation
with regard to human rights, and suppression
bias, like poor working conditions, is accepted as
part of the normal research environment. It
therefore stands to reason that in these regions,
the absence of information on harassment of
researchers can almost be said to be evidence of
the effectiveness of suppression bias as a
deterrent to investigation of environmental
hazards. So far, neither the ISEE nor the
Collegium Ramazzini have received requests for
help from these settings.
In the developed countries, we need to ask
whether a more subtle institutional form of
suppression bias could be taking hold. Academic
institutions are entering into strategic business
alliances, most often with biotechnology and
pharmaceutical firms (32). The close ties
between university and business are a frontal
assault on the last vestiges of “academic
freedom” of the faculty members. Moreover, the
diminishing role of governments in funding
public health research causes academic
institutions to pursue corporate funding. This
trend furthers the alliance of university and
business, and increases the likelihood of
suppression bias.
We suggest that suppression bias and the
occurrence of environmental hazards circularly
reinforce each other. Alibek has pointed out that
in the former Soviet Union, suppression of
information on health hazards to personnel and
the environment from activities related to
weaponizing bacterial and viral organisms for
bioterrorism led to a scenario in which safety was
jeopardized over and over again in the name of
national security. He described a scenario in
which suppression bias resulting from the
harassment of epidemiologic messengers
endangered public health (33).
Institutional safeguards against harassment in
environmental science
Until now, research on ethics in environmental
epidemiology has focused on the obligations of
individual researchers to comply with norms of
truth and not engage in scientific misconduct
(34-35). But there has been insufficient
discussion of the obligations of institutions to
protect their workers and their findings from
external harassment when their findings are
embarrassing, costly, or threatening to powerful
interests. Such harassment serves as a deterrent
to investigating and reporting information about
hazards and risks.
Measures to protect messengers in
environmental and occupational epidemiology
should be required of grant recipients of research
contracts around the world and should become a
worldwide institutional norm.
Messengers can be wrong
The statements made by epidemiologic
messengers on the presence of a hazard or risk
may be right or they may be wrong. We suggest
that pressures, harassment, and abuse are no
substitute for access to the peer-review process.
At the same time, there is the need to be
concerned about pressures on this peer review
process by new trends in the academic world to
forge alliances between industrial or
technological interests and the research
community.
What Next?
Professional societies derive their legitimacy
from their mission in promoting the public good.
Investigation and reporting environmental
hazards and their risks are essential to prevent
damage to public welfare. As we noted at the
outset, the protection of epidemiologic
messengers derives from the primacy of
protecting public health. Ironically, Benthamite
rationales —stretched somewhat—could have
served to acquit Socrates were it to have been
shown that his teachings were necessary for
protection of the greatest good for the greatest
number, or more fundamentally, for the health
and welfare of all individuals, in keeping with
traditions of the sanctity of preserving individual
human life.
Organizations concerned with ethics in
171
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
science in recent years rightfully called attention
to the need to establish rigid standards for
preventing scientific misconduct by individuals.
The first generation of work on ethics in research
focused on setting standards, procedures and
codes of practices which defined responsibilities
of individual scientists at all levels, to work
according to codes of truth, quality assurance and
quality control, precision and accountability
(36-37). This first generation of work addressed
issues raised by whistleblower scientists who
drew attention to scientific misconduct in the
laboratories of their superiors. These episodes of
misconduct led to the distortion of findings,
failures in quality assurance and quality control,
and lapses in precision and accountability. The
issue at hand now is standards for preventing
institutional misconduct. There has been no
parallel effort of equivalent force to enact
standards that prevent misconduct by
institutions—be they the scientist’s employer or
other bodies—which results in harassment of
epidemiologic messengers.
We suggest that failure to ensure proper
access to independent peer review insulated from
internal and external pressures is a form of
institutional misconduct. The same statement
applies to failure to provide protection against
legal harassment, such as occurs with SLAPP
lawsuits. Therefore, the second generation of
work in ethics and scientific integrity has to deal
with a new and different set of problems. These
pertain to the need for standards, procedures, and
codes of practice that define the responsibilities
of institutions and organizations to prevent the
harassment of individual environmental scientists
who either attempt to investigate or report
findings on hazard and risk which go against
powerful interests that could be damaged by such
information.
The issues at hand here are not quite the
same as those having to do with investigations of
scientific misconduct, i.e., the falsification or
fabrication of research results. In investigations
of scientific misconduct, there is a more or less
level playing-field for right and wrong: the peer
reviewed literature and its well elaborated codes
and norms for evaluating scientific evidence. In
the case of whistleblowing in environmental and
occupational epidemiology, the problem is to
promote access to this level playing field, and to
ensure that the playing-field is indeed level.
There is a need to ensure that outside interests,
often commercial, economic or political, do not
obstruct access to or distort the process of peer
review.
There is a need to recognize a dissonance
between the emphasis of the first generation of
ethics on promotion of research integrity and that
of the second on prevention of suppression bias.
Often there is a two-stage scenario in which
investigators—or officials in need of a rapid
estimation of hazard or risk— are first blocked
from access to the exposed populations and
relevant databases, and then their reports are
disqualified because they are incomplete,
imperfect or imprecise. In short, the very criteria
used to define the quality of investigation may
serve as barriers to reporting its substance. This
situation —in which being precisely wrong is
considered preferable to being approximately
right—is the classic scenario of delay.
One form of harassment of environmental
epidemiologists and other investigators is to
subject their databases and records to a battery of
legal subpoenas. If transparency is our norm, it
is hard to fault such challenges. However, such
subpoenas pose potential challenges to the
privacy of research on individuals, and may serve
as a deterrent to their giving permission to use
data on individual exposure and risk. But, in the
case of environmental epidemiology and related
fields, the ultimate form of harassment is to deny
the investigator access to databases, so as to
prevent a complete investigation. In
epidemiology, in particular, barriers to accessing
databases on denominators can be particularly
devastating, because they effectively forestall
precise calculations of risk. Such barriers, by
delaying or impeding investigations, may not
only block research, but they permit the
persistence of situations hazardous to the health
and safety of the populations themselves. We see
use of the term “sound science” to disparage
attempts to make do with limitations of
estimates of risk based on studies not meeting
“gold standard” requirements because data sets
may not be complete. (38).
A second form of harassment is lawsuits for
libel. To address this hazard to environmental
scientists, there is a need to explore the use of
insurance policies modeled after those available
to writers. Grants to environmental scientists
should include budgeting for such insurance.
Conclusions
Until now, there has been no watchdog address
172
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Richter, et al., Whistleblowers in Environmental Science
for environmental and occupational
epidemiologists to which to turn for assistance.
We suggest that major granting agencies
follow the lead of ISEE and the Collegium
Ramazzini in protecting environmental scientists
from harassment. We call for studies on the
impact of harassment of research scientists on the
detection and prevention of health risk. We call
for the development and use of codes for
protecting environmental scientists from
harassment when they are engaged in this
mission. We recommend that measures to protect
messengers in environmental and occupational
epidemiology be required of recipients of
research grants or contracts around the world.
These codes should become a worldwide
institutional norm. Codes that protect
epidemiologic messengers in environmental and
occupational medicine will serve also to protect
the public health.
Acknowledgements
Councillors of ISEE and Collegium Ramazzini,
which provided a grant for EDR’s work, and an
anonymous donor for TB’s work.
Bibliography
1. Martin, B (1981) The scientific straightjacket: the power
structure of science and the suppression of
environmental scholarship. The Ecologist. Vol 11(1):
33-43.
2. Deyo, RA, Psaty BM, Simon G, Wagner EH, Omenn GS
(1997) The messenger under attack- intimidation of
researchers by special-interest groups. New England
Journal of Medicine. 366 (16): 1176- 1180.
3. Relyea, HC (1994) Silencing Science: National Security
Controls and Scientific Communication. Norwood, New
Jersey: Ablex Publishing.
4. Hess, DJ (1999) Suppression, bias, and selection in
science: the case of cancer research. Accountability in
Research. 6: 245-257.
5. Plato (1963) The trial of and death of Socrates. New
York: Heritage Press.
6. Dror R. Notes on the trial of Socrates. Based on
conversations with the late Professor Yeshayahu
Liebowitz (personal correspondence, June 1999).
7. Bentham, J (1907) An introduction to the principles of
morals and legislation. Oxford: Clarendon Press.
8. Martin, B (1999) Suppressing research data: methods,
context, accountability, and responses. Accountability in
Research, Vol. 6: 333-372.
9. Bahr, MH. “Violations of academic freedom: official
statistics and personal reports”, Social problems, 14,
1967, pp. 310-320.
10. Cate Jenkins, “Newly Revealed Fraud by Monsanto in
an Epidemiological Study Used by EPA to Assess
Human Health Effects from Dioxins”, EPA
memorandum to Raymond Loehr, Chairman, Executive
Committee, EPA Science Advisory Board, February 23,
1990. As cited in Rachel’s Hazardous Waste News #171.
Dioxins and Cancer–Fraudulent Stories. March 7, 1990.
11. Jenkins C. (September 3, 1991) Affidavit: Recent
Scientific Evidence Developed after 1984 Supporting a
Causal Relationship between Dioxin and Human Health
Effects. Full text of the affidavit is available online from
Agent Orange Victims, Families, and Friends in Touch.
http://softvision.netfirms.com/ao_vets/.
12. Memo from Gerald H. Yamada, Office of [EPA’s]
General Counsel, to Catherine L. Jenkins, dated
February 26, 1993. As cited in Rachel’s Hazardous
Waste News #328. An update on two key incinerator
battles. March 11, 1993.
13. Jenkins v. U.S. Environmental Protection Agency, 92CAA-6 (ALJ Dec. 14, 1992). Available online, http://
www.oalj.dol.gov/public/wblower/decsn/92caa06a.htm.
14. http://www.pmac.net/pesticide_wars.html. Pesticide
wars: the troubling story of Omar Shafey.
15. http://www.web.amnesty.org/ai.nsf/Index/
AMR250281997?OpenDocument&of=COUNTRIES%5CCUBA.
Dr Desi MENDOZA Rivero, Cuba.
16. http://www.web.amnesty.org/ai.nsf/index/
EUR460421996 (1996) Federal Security Service (FSB)
versus Prisoner of Conscience Aleksandr Nikitin: Return
to Soviet Practices.
17. www.rsf.fr/uk/html/europe/cplp01/lp01/261201.html Reporters wtihout Borders, Dec 2001 http://
www.internews.ru/ngo/pasko/indexe.html (1997)
Grigory Pasko sentenced to four years in prison.
18. Martin B (1993) Stamping Out Dissent. Newsweek. 26
April: 49-50.
19. Pesticide & Toxic Chemical News (1981) Dr. Mel
Reuber, Pathologist, Gets Sharp Censure, Warning from
his Supervisor (25 April): 22-23.
20. Watterson A (1998). Occupational health and safety
policy in the UK: who is driving the agenda with the
“greylisting” of health and safety activists? Int J Occup
Environ Health 4(4):275-77.
21. Coulter, J (1980) Memorandum re use of ethylene oxide
as a sterilant in the S.P.F. laboratory at Northfield, 16
April.
22. van den Bosch, R. (1978) The Pesticide Conspiracy.
Garden City, NY: Doubleday.
23. http://home.cwru.edu/activism/kern.html. Conflicts of
interest: the case of David Kern.
24. Westin, JB and Richter, ED (1990) The Israeli breastcancer anomaly. Ann NY Acad Sci 609:269-279.
25. Byers, RK and Lord, EE (1943) Late Effects of Lead
Poisoning on Mental Development, American Journal of
Diseases of Children. Vol. 6: 471-494.
26. Smythe, L (1980) letter, New Scientist, 85: 775-776. As
cited in reference Martin, B (reference 1).
27. http://www.whistleblower.org/www/998news.htm#Oil.
Big oil and the battle to protect the environment.
Government Accountability Project, Fall 1998.
28. Sternberg, S (1996) An Unusual Case: Law Overseas
173
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Myron Mehlman’s case against Mobil Corp. attempts to
apply New Jersey statute to incident in Japan. The
Scientist 10(1).
De Blas, A. (1994) Environmental Effects of Mount
Lyell Operations on Macquarie Harbour and Strahan.
Sydney: Australian Centre for Independent Journalism,
University of Technology, Sydney.
http://www.jpost.com/Editions/2000/03/20/News/
News.4270.html. Siegel-Itzkovich, J (2000).
Environmental group aiding expert in libel suit.
Jerusalem Post Online Edition. March 20, 2000
http://www.iseepi.org/ethguide.htm#victimized. ISEE
support for victimized colleagues.
Abraham, J (1995). Science, Politics and the
Pharmaceutical Industry: Controversy and Bias in Drug
Regulation. London: ICL Press.
Alibek, KW and Handelman S (2000). Biohazard. Dell
Publishing Company.
Nigg, HN, Radulesc G (1994). Scientific Misconduct in
Environmental Science and Toxicology. JAMA
272:168-170.
LaFollette MC. Stealing Into Print: Fraud, Plagiarism,
and Misconduct in Scientific Publishing. Berkeley:
University of California Press; 1992:294.
American College of Epidemiology (2000). Ethics
Guidelines. Ann Epidemiol. 10: 487-497.
http://ori.dhhs.gov/html/policies/oripolicies.asp. Office
of Research Integrity.
Ong, EK and Glantz, SA (2001). Constructing “sound
science” and “good epidemiology”: tobacco, lawyers,
and public relations firms. Am J Public Health
91(11):1749-57
174
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Richter, et al., Whistleblowers in Environmental Science
ISEE Ethics Committee Epidemiologist Whistleblower/Messenger Questionnaire:
1. Personal status
• ISEE Member? Y/N _____; ISEA Member? Y/N_____; Age _____
• Gender M/F _____
• Personal Status: M, S, D, W _____
• Children (Give no___ )
Education
Undergrad
MD
MPH/MSc/MS/MA
PhD/DPH
Post Doc
Residency Spec
From
____
____
____
____
____
____
To
____
____
____
____
____
____
Where* (see Code)
__________
__________
__________
__________
__________
__________
Code:
America: NA, CA, LA
Euro pe: WestE, Med,
EastE Mideast: ME
Africa: WA, EA, SA Asia:
Ind, CentAsia, Jp, Ch,
SEA Oceania: Aus, PI
2. Currently Employed
Where? _________________________ see code above
By: Govt Civilian Military Police (Circle one)
Level: National Regional/Province/District/Municipal (Circle one)
University/College ________________________
Independent research institute
Foundation _____
Trade Union NGO Self Employed
Industry/Corporation: If yes? ____________________
Multinational Y/N _____
Other_______
3. Tenured or permanent? Y/N
Rank (Univ): Prof Sr Lect/Lect__ Asst__Other_____
4. Research/salary funded by: (Circle correct answer)
Government
Industry
Foundation Other
No funding
5. Harassment: Following publication of research findings on health risks from environmental exposures,
have you ever experienced:
Ostracism Y/N
Demotion Y/N
Criminal investigation
/Prosecution/Trial Y/N
Confiscation of data Y/N
Loss of job Y/N
Physical threats Y/N
Threat of loss of job Y/N
Threats of lawsuits Y/N
Physical Attack Y/N
Transfer Y/N
Lawsuits Y/N
Imprisonment Y/N
Other
175
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
How many other co-researchers were there? ____ Did they experience any of the responses? Y/N
6. Research on specific problem which lead to episode(s) of harassment or threat or abuse:
during which research carried out: From______To_______________
Years
Was this research on the hazard/risk published in:
Peer reviewed journal (sited in SCI CIT INDEX)___
Masters thesis___
Doctorate___
Internal document of organization in which you were then employed/studied?
Professional society___
Non peer-reviewed journal____
Other____
Date of publication?_______ Would you be able to provide the Citation? Leave blank if you wish
___________________________________________
7. Response
7a. Did you receive assistance after being subject to any of the above problems? Yes___
No____
7b. If yes, from: Individual colleagues____ _ Superiors____ Professional societies ___ NGO's
inside country_____ Journalists/Media_____ Lawyers or legal aid g roups____Colleagues
outside country____ NGO’s outside country_______Family______Other__________
8. Publication If findings were not published, were you prevented from submitting findings on health
risks on a hazardous exposure/risk for publication in a peer reviewed journal? Yes___ No____
OPTIONAL____________________________________________________________________
9. F indings: Could you summarize the findings you discovered/reported for which you were harassed?
Study design
Outcome
RR/OR
Reference
(Cohort, CC, Prev,
Pop(s) / N
Exposure(s)
TS, Other)
_______________
__________ __________ __________ __________ __________
_______________
__________ __________ __________ __________ __________
10. In retrospect, were your findings: understated?____a proper assessment?____overstated?_____ For
further information: http//:www.iseepi.org/ethguide.htm
176
II. Teaching
6. Training in the Responsible Conduct of Research
Influencing the Moral Dimensions of Professional Practice:
Implications for Teaching and Assessing for Research Integrity
Muriel J. Bebeau, Center for the Study of Ethical Development, University of Minnesota,
USA
Key Words: Assessment, Moral development, Professional ethics, Research ethics
This paper will present implications for teaching and assessing for research integrity from 20 years of
experience designing and assessing ethical development in the dental profession. Data sources for the
implications include: 1) pretest/posttest data for 18 cohorts of dental students who completed a wellvalidated ethics program; 2) pre/post assessments of 28 practitioners referred by a licensing Board1
for individualized ethics instruction because they violated the State Dental Practice Act; and 3) efforts
in several professions to influence moral judgment development.
After pointing out some of the features of the Minnesota ethics program, the program’s
theoretical foundations (e.g., the processes of morality) are described. Each process suggests research
questions that motivate inquiry and assessment methods that were developed or used to investigate
the research questions and to gather evidence on program effectiveness. The paper continues with a
summary of data supporting the independence of the component processes and a discussion of the
ongoing search for behavioral indicators that could provide the “acid test” for the model. The paper
concludes with a discussion of the implications for the teaching and assessing for research integrity.
Special features2 of the curriculum include: 1) 43 contact hours distributed over four years;
2) required attendance and participation; 3) small group instruction—using dilemma discussion and
role-play; 4) an emphasis on student performance, self-assessment and personalized feedback; 5) use
of validated assessment methods that are checked for reliability; 6) involvement of high status
professionals (in measurement development and feedback); and 7) involvement of faculty in the
teaching. Thus, the curriculum isn’t a one-shot intervention, nor is it the isolated property of one
instructor.
Theoretical Foundations
The ethics curriculum, for students and referred practitioners, is designed to promote functional
processes that give rise to morality: 1) ethical sensitivity; 2) moral reasoning; 3) moral motivation
and commitment; and 4) ethical implementation (1). Moral failing is conceptualized as the result of
deficiencies in one or more of the processes. Rest’s Four Component Model of Morality,
operationally defined below, is a substantial departure from much of the work in psychology that
arbitrarily divides moral functioning into affects, cognitions, and behaviors (2).
The Four Component Model of Morality
Early in the cognitive developmental research program initiated by Kohlberg, he noted that, in
Corresponding author: Muriel J. Bebeau, Department of Preventive Sciences, School of Dentistry, University of Minnesota,
515 Delaware S.E., Minneapolis, MN 55455, 612-625-4633 (voice), 612-626-6096 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
addition to moral judgments, other processes
were important to the production of moral
behavior (3). Rest made these processes more
explicit in what he called the Four Component
Model of Morality (1). Starting from the
question: how does moral behavior come about,
Rest suggested that the literature supports at least
four component processes, all of which must be
activated in order for moral behavior to occur.
These include:
1. Moral sensitivity (interpreting the
situation as moral)
This process highlights the idea that moral
behavior can only occur if the individual codes
the situation as moral. Specifically, Component
1 focuses on the various actions that are available
and how each action might affect the self and
others.
2. Moral judgment (judging which of the
available actions are most justified)
This is the process that Kohlberg emphasized.
Here the focus is on judging which of the various
options are the most ethically justified. Further,
the job of a psychologist and educator is in
sketching out how the justification process
develops and under what conditions these
processes inform real-world choices.
3. Moral motivation (prioritizing the
moral over other significant concerns)
Less understood than the other processes, the
main concern of Component 3 is, “why be
moral.” The model acknowledges that
individuals have a number of legitimate concerns
that may not be compatible with the moral
choice: for instance, career pressures, established
relationships, idiosyncratic personal concerns,
among many others. Some of the most notable
lapses of ethical behavior in the professions can
be attributed to low priority placed on the moral,
even when the moral choice is very well
understood.
4. Moral character (being able to
construct and implement actions that service
the moral choice)
Component 4 represents the processes by which
one constructs an appropriate course of action,
avoids distractions, and maintains the courage to
continue.
It is important to notice that the model is not
conceived as a linear problem-solving model.
For example, moral motivation may impact
moral sensitivity, and moral character may
constrain moral motivation. In fact, Rest (1)
makes clear the interactive nature of the
components. Further, the Four Component
Model assumes that cognition and affect cooccur in all areas of moral functioning. Thus,
moral action is not simply the result of separate
affective and cognitive processes operating in
interaction, as suggested by traditional models of
moral function that focus on three domains—
cognitions, affects and behavior (4, 5). Instead,
each of the four components are mixes of
affective and cognitive processes that contribute
to the component’s primary function (e.g.,
identifying a situation as moral). Bebeau, Rest,
& Narvaez suggest that researchers focus
attention on identifying processes as they
contribute to moral action, rather than attempting
to understand moral actions from a starting point
defined by arbitrarily dividing moral functioning
into affect, cognitions, and behavior (2).
The debate on the usefulness of a
psychological theory of morality, that has its
foundation in the work of Lawrence Kohlberg, is
addressed in “Postconventional Moral Thinking”
(6). This paper presents a theory of moral
judgment development that is not grounded in a
particularistic moral theory—as was
Kohlberg’s—but is grounded in empirical
evidence illustrating that as individuals develop,
so do the basic understandings they bring to
resolving complex moral problems. Such
findings are of importance to ethics education in
general, as the goal of ethics education is, simply
put, to promote ethical development. The
authors contend that their findings will be of
particular importance to research ethics educators
because of their interest in promoting critical
thinking about responsible research conduct (6).
In the past, ethicists working in the professions
questioned the usefulness of a moral
development theory (and related measures) that
favored a particular moral theory, observing that
practitioners working on real problems often
developed well-reasoned solutions without
regard to a particular theory or even to
principlism as a way of arriving at moral
judgments (7).
By amending a theory of moral judgment
development to make it congruent with advances
in moral philosophy, the authors hope to counter
current views of the obsolescence of moral
psychology and support more interdisciplinary
collaboration in the design and evaluation of
moral education programs. Further, a more
enlightened view of the role of tests of moral
judgment development should enable educators
180
–––––––––––––––––––––––––––––––––––––––––––––––– Bebeau, Influencing the Moral Dimensions of Professional Practice
to put such tests to more appropriate use.
Besides drawing attention to a broader
conception of postconventional moral thinking,
the authors direct the reader’s attention to a
broader conception of morality, one that
encompasses moral judgment, but that also
addresses other aspects of moral functioning,
including moral sensitivity, motivation, character,
and competence. The Four Component Model of
Morality has been a centerpiece for research
activities at the Center for the Study of Ethical
Development for nearly 20 years.
Educational Interventions Assessed in
Terms of the Four Components
A program of research and educational
development to investigate the usefulness of the
model was initiated by Jim Rest and the author in
the early 80s. Variations on these research
questions motivated the inquiry: Can ethical
sensitivity (or any of the other components) be
reliably assessed? Do students differ in ethical
sensitivity (or other components)? Can
sensitivity (or other components) be enhanced?
And, is ethical sensitivity distinct from other
components?
The Four Component Model offers unique
information and direction for educational
development. First, it suggests profitable areas
for measurement development. To claim that a
program is effective in a broad sense, it seems
reasonable to expect changes within each of the
four components. For the dental curriculum,
measures of each component were designed and
validated, and data from them helped identify
deficiencies to consider as the curriculum was
designed. There are measurement models and
methods for assessing each of the components (2,
8). These can be used as templates for
assessment in various contexts.
Second, the model provided direction for
instructional design for groups, as well as for
individual referrals. For referred practitioners,
deficiencies were noted in various components
and were associated with particular moral
weaknesses (9). Targeting specific deficiencies
in an individualized instructional program proved
to be an effective intervention strategy, resulting
in substantially enhanced posttest performance.
Measures for the Components of Morality
Five measures are used to assess performance in
the Dental Ethics Curriculum. A brief
description of each measure and the findings are
summarized as follows:
Component I: Ethical Sensitivity
The Dental Ethical Sensitivity Test (DEST)
The DEST (Form A or B) (10, 11) assesses the
ability to recognize the ethical issues hidden
within the professional problems dentists
encounter in practice. Students’ verbal responses
to four audio-taped dramas are recorded and
transcribed, and provided to the student and to a
practicing dentist, who each apply the DEST
coding scheme, then meet for personalized
feedback. The validity and reliability of the
DEST are reported in several studies,
summarized in Bebeau (8) and Fravel and
Bebeau (12). Briefly, the results support these
conclusions: 1) Ethical sensitivity can be
reliably assessed. Calibrated raters achieved item
agreement ranging from 84.7 percent to 88
percent. Reliability estimates for individual
cases ranged from .83 to .92; 2) Students and
practitioners vary in sensitivity to ethical issues.
Students at different levels of education in
medicine and dentistry (physicians vs.
technicians or dentists vs. hygienists) differed
significantly, such that those with longer
preparation showed higher levels of sensitivity.
Further, the DEST is sensitive to institutional
differences; 3) Women have a slight edge over
men in recognizing ethical issues, but differences
were not attributed to differential recognition of
the care and justice issues; 4) Ethical sensitivity
can be enhanced through instruction; 5) Ethical
sensitivity is distinct from moral reasoning
abilities. Correlations between the DEST and
Defining Issues Test (DIT) posttest are
consistently low (see later section for more
detail); 6) Despite the stressful nature of the
DEST assessment—responding on the spot to
complex cases, having responses taped,
transcribed, and sent to a practicing professional
is a high-stakes examination—students value the
assessment and feedback experience.
Component II: Moral Reasoning and
Judgment
In this section, two measures are described: a
well-established measure (DIT) and a newlydevised, context-specific test of ethical reasoning
and judgment (Dental Ethical Reasoning and
Judgment Test [DERJT]). In the case of the DIT,
181
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
the discussion will include findings from new
analyses with new indices for three of the recent
cohorts of dental students.
The Defining Issues Test
The DIT measures life-span development of
moral reasoning and judgment (13). The DIT is
the most widely used test of moral judgment
development and is often used as an outcome
measure for intervention studies, because it has
an exceptional validation history.3 Students read
dilemmas, and then rate and rank the importance
of each of 12 arguments to support their position.
Confirmatory factor analysis of a mega-sample
of over 44,000 subjects shows that items
(arguments) cluster around three general moral
schemas: Personal Interest, Maintaining Norms,
and Postconventional schemas (14). Typically,
researchers have reported scores in terms of the P
score—the proportion of items selected that
appeal to Postconventional moral frameworks for
making decisions. The average adult selects
postconventional moral arguments about 40
percent of the time, the average Ph.D. candidate
in moral philosophy or political science about
65.2 percent of the time, the average graduate
student 53.5, with the average college graduate at
42, and the average high school student at 31.8
percent.
Progress in moral judgment is
developmental, and development proceeds as
long as an individual is in an environment that
stimulates moral thinking. College has a
powerful effect on moral judgment development.
McNeel’s meta analysis of 22 longitudinal
studies of liberal arts students estimates first year
college students at 36, seniors at 46, estimating
an effect size of .80 (15). Effect sizes of about
0.80 are among the largest effect sizes for many
college impact variables that have been studied.
In fact, effect sizes are higher for moral judgment
than for the many cognitive and affective college
outcome variables that have been studied (16).
Yet professional schools (e.g., Veterinary
Medicine, Medicine, Dentistry, and Accounting)
are programs where one does not typically see
gains associated with the educational program,
unless the program has a specially-designed
ethics curriculum (17). Further, for some
students and some professions, programs actually
seem to inhibit growth (18, 19).
Change in moral judgment can be attributed
to the ethics curriculum (18). The average
entering Minnesota dental student scores 46
(with cohorts ranging from 42 to 49 across the 15
classes tested). The average graduate selects
postconventional arguments 51 percent of the
time (with cohorts ranging from 47 to 55). Effect
sizes vary across classes, with a range of .12 to
.78, with an average of .43. For each cohort,
scores tend to be normally distributed. For
entering students, as many as 35 percent are not
using postconventional moral schemas as often
as the average adult, with about seven percent
above the mean of philosophy and political
science graduate students. Although we see an
upward shift in the distribution at posttest, with
16 percent lower than the mean of the average
adult, and 20 percent above the mean of
philosophy and political science graduates; of
particular interest are the proportion of students
who showed no change or regressed from pretest
to posttest. By classifying students’ change
scores into categories defined by the standard
error of measurement (18), Bebeau reported that
44 percent of the 1,229 students who participated
in the curriculum made moderate to highly
significant gains, 40 percent showed no change,
and 16 percent regressed on the P score (20).
New Indices and New Analyses of DIT
Scores
Observations of what appeared to be regression
in postconventional reasoning in our intervention
studies prompted the validation studies, including
development of an alternate form of the DIT and
a reanalysis of moral education interventions that
attended to several moral cognition variables
derived from DIT scores (6, 14, 21, 22, 23, 24).
Moral Schema Profiles. Instead of relying
only on the P score as a measure of pretest to
posttest change, a profile showing the proportion
of times a student rates was constructed to
illustrate important items for each of three
general schema: a Personal Interests schema
(Kohlbergian Stage 2 and 3 items); a Maintaining
Norms schema (Stage 4 items): and a
Postconventional schema (Stage 5 and 6 items).
Figure 1 illustrates how two profiles with similar
P scores can reflect differing levels of moral
judgment development. Examining profiles from
students who did not show gains in DIT P scores
from pretest to posttest (20) illustrates a
substantial reduction on the Personal Interest
schema coupled with an increase on the
Maintaining Norms schema, without significant
change on the Postconventional schema score. In
fact, when the statistically significant pretest/
posttest change for the 18 cohorts of students that
participated in the dental curriculum was
182
–––––––––––––––––––––––––––––––––––––––––––––––– Bebeau, Influencing the Moral Dimensions of Professional Practice
50
45
PI
MN
P
Type 4
40
35
30
25
20
15
10
5
0
Transition
Transition
Type
3
Type 3
Consolidation
Consolidation
Type 4
Figure 1. Moral judgment profiles illustrating similar P
scores, but differences on other moral cognition variables.
PI = Personal Interests Schema
MN = Maintaining Norms Schema
P = Postconventional Moral Schema
reanalyzed, the reduction in the Personal Interests
schema score appeared much greater and more
consistent across cohorts than changes in P score.
By focusing only on the P score, researchers may
be missing change that is quite significant.
Consolidation/Transition. Figure 1
illustrates another variable to consider in
describing change. When there is little evidence
of discrimination among the schema-typed items,
students are classified as transitional. A flat
profile is viewed as a marker of developmental
disequilibrium, or transition, since there is no
evidence of a schema preference. A further
discussion of this topic is addressed by Thoma
and Rest (22). A pretest/posttest analysis of
consolidation/transition status was conducted for
222 dental students (20), showing that nearly half
the students (46.9%) were in a transitional status
at pretest, whereas only 27.1 percent exhibited
the transitional status at posttest.
Type. Profiles can further be classified by
type (22), where type reflects both the
predominant schema and the extent of its use.
By reexamining several intervention studies
reported in the literature, Yeap showed that Type
provided a more illuminating description of
change that occurred as a result of an
intervention than relying simply on the P Score
(24). A pretest/posttest analysis of six Types was
also conducted for the 222 students reported
above. Whereas the pretest responses were
distributed among Types 3, 4, 5, and 6, 61.2
percent were classified at Types 5 and 6
(postconventional types), with the distribution
peaking at Type 6. For the posttest responses,
75.8 percent were classified at Types 5 and 6,
with 59.9 percent at Type 6. By way of
comparison, Yeap reported college student
samples peaked at Type 3.
These new analytical procedures may help to
unravel some of the puzzles researchers have
cited, where professional groups like Accounting
and Auditing (19) seem to regress on moral
judgment as a result of an educational program.
Such analysis may clarify McNeel’s findings that
programs that are too careerist (focus narrowly
on technicalities of beginning job performance)
or too dogmatic (in closing off questioning and
inquiry) inhibit growth in reasoning (15). Such
findings would have implications for developing
research integrity. Courses that focus narrowly
on the rules of research conduct may focus
attention on the minimal (legal) standards, rather
than on aspirational standards for research
integrity.
Tests like the DIT are valuable for assessing
general reasoning that is a critical element of
professional ethical development, but they may
not be sensitive to the specific concepts taught in
a professional ethics course—or indeed, in a
research ethics course. The question (for
educators) is often whether to teach specifically
to the codes or policy manuals, or to teach
concepts particular to a discipline—informed
consent, intellectual property, conflict of interest,
etc.
The Dental Ethical Reasoning and
Judgment Test (DERJT)
The DERJT is a first effort to test application of
context-specific concepts (taught in ethics
courses) to real cases (25). The test is similar to
the DIT, in that cases are presented followed by
lists of action choices and justifications. The
action and justification choices for each problem
were generated by a group of Minnesota dental
faculty and residents. The scoring key was
developed by a group of “dental ethical experts.”
When taking the test, a respondent rates each
action or justification, then selects the two best
and two worst action choices, and the three best
and two worst justifications. Scores are
determined by calculating the proportion of times
a respondent selects action choices and
justifications consistent with “expert judgment.”
In validation studies, Bebeau and Thoma have
seen clear expert novice differences (25).
Further, scores for students, practitioners, and
referrals appear to be normally distributed. In a
183
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
study comparing our graduates’ responses to
familiar vs. unfamiliar problems presented on the
test, it appears that a good grasp of
postconventional moral schemas is a necessary
condition for transfer to new problems.
Component III: Motivation and
Commitment
The Professional Role Orientation
Inventory (PROI)
The PROI assesses commitment to privilege
professional values over personal values (26, 27).
Likert scales assess dimensions of
professionalism that are theoretically linked to
models of professionalism described in the
professional ethics literature. The PROI scales,
in particular the responsibility and authority
scales, have been shown to consistently
differentiate beginning and advanced student
groups and practitioner groups expected to differ
in role concept. By plotting responses of a
cohort group on a two dimensional grid, four
distinctly different views of professionalism are
observed (26) and, if applied, would favor
different decisions about the extent of
responsibility to others. In comparing practicing
dentists with entering students and graduates, our
graduates consistently express a significantly
greater sense of responsibility to others than
entering students and practicing dentists from the
region. This finding has been replicated for five
cohorts of graduates (n = 379). Additionally, the
graduates’ mean score was not significantly
different from a group of 48 dentists, who
demonstrated special commitment to
professionalism by volunteering to participate in
a national seminar to train ethics seminar leaders.
A recent comparison of pretest/posttest scores for
the Classes of 1997-1999 (20) indicates
significant change (p < .0001) from pretest to
posttest. Cross-sectional studies of differences
between pre and posttest scores for a comparable
dental program suggests that ethics instruction
accounts for change.
To provide students or practitioners with
individualized feedback on their role concept, an
interpretive guide is provided enabling a
respondent to sum his or her own scores on each
scale, plot them on the two dimensional grid (one
grid is provided for the authority and
responsibility scales, one for the agency and
autonomy scales), and then compare responses to
their cohort. Descriptions of each of the models
of professionalism are included to stimulate
thinking about the model of professionalism that
appears to be dominant for the individual. When
the scales and interpretive guide are used in an
educational setting, participants can compare and
discuss items and challenge each other’s
thinking.
Developing a concept of role appears to
require instruction and opportunities for
reflection. At entry to professional school,
Minnesota dental students do not illustrate a
good understanding of key concepts of
professionalism like service to society, or the
priority of patient well-being, or the duty to selfregulation (8). But, even after participation in an
instructional program in which students write an
essay describing their perception of their
professional role (the program is of demonstrated
effectiveness and includes generous amounts of
practice and feedback on performance), key
concepts like self-regulation, service to society,
and the basic duty to place patient’s rights before
self-interest are still frequently omitted or
miscommunicated by as many as 20 percent of
the students. The literature on concept learning
has helped us see that when students have no
functional schema for a particular concept,
several educational experiences are required to
instill a clear concept of the professional’s role.
Whether instilling a clear idea of the
professional’s role will motivate students to place
moral values over personal ones is a key
question. The most direct evidence of a
relationship between role concept and
professionalism comes from the study of
performance of the 28 members of the practicing
community, referred for courses in dental ethics
because of violations of the dental practice act.
Although the practitioners varied considerably on
measures of ethical sensitivity, reasoning, and
ethical implementation, 27 of 28 were unable to
clearly articulate role expectations for a
professional (9).
Component IV: Moral Implementation
(character and competence)
Shifting to the last component, character and
competence, the authors have observed that
guided practice changes the expectation of
efficacy that is likely to change behavior. Roleplaying builds competence and confidence in
resolving thorny ethical problems, and skills in
184
–––––––––––––––––––––––––––––––––––––––––––––––– Bebeau, Influencing the Moral Dimensions of Professional Practice
communication and negotiation are necessary
requisites of this competence.
A Professional Problem Solving Index
Problem-solving and role-playing performance
scores are calculated for eight complex cases that
present difficult human interaction problems (8,
20). Students are directed to prepare 1) an
interpretation of the facts that must be addressed
if the problem is to be resolved efficiently; 2) an
action plan; and 3) a verbatim dialog to illustrate
the implementation of the action plan. A
checklist, prepared for each case, assures some
uniformity in judging responses. Each response
is reviewed by a peer and by the course instructor
who provide written comments identifying the
strengths and shortcomings of the assignment.
As with other measures, scores are normally
distributed and cohort differences are observed.
Independence of the Components of
Morality
Rest’s Four Component Model predicts the
independence of the components (1). Prior
studies have typically reported low to very low
correlations between ethical sensitivity and moral
judgment, but correlations among the other
components have varied from very low to an
occasional moderate correlation. Often sample
sizes have been low, challenging the reliability of
the estimates. Recently, Bebeau reported
correlations between components for a larger
sample (230 students) (20). Except for the
expected moderate correlations (.46) between the
DIT Pretest and Posttest and between the PROI
Pretest and Posttest scales (.38), each measure
appears to provide unique information about
ethical decision making competence. Consistent
with earlier studies, correlations are consistently
very low between the DEST and the DIT, and
between the DEST and other component
measures (8). The exception is between the
DEST and the DERJT justification score, where
there appears to be some overlap between the
two tests (r = .28). Also consistent with earlier
reports (27), there appears to be some low to
moderately-low relationship between the PROI
Responsibility Scales and the DEST and DIT.
The Continuing Search for Behavioral
Indicators
Several attempts have been made to show the
contributions of each of the components to
meaningful behavioral indicators. Although
moral judgment is linked to a wide range of prosocial behaviors (28), including clinical
performance ratings for nurses (29, 30),
physicians (31) and dentists (8), and to
preferences for the more altruistic law disciplines
for law students (32), the search for behavioral
measures to examine the relative contribution of
each component to the behavioral outcomes has
been a frustrating one. The author’s most recent
effort (20) has been to calculate a productivity
index that reflects students’ success in interacting
effectively with patients to achieve acceptance
and completion of treatment recommendations.
To meet competency requirements, the student
must achieve an average monthly index (over all
months of clinical practice) of .75 or above.
Although there was considerable range in
productivity from .67 to 1.19, since students must
meet a .75 overall average in order to graduate,
the productivity index, while identifying highly
effective students, also produces a highly skewed
distribution (Mean = .80, S.D. = .08). In the
analysis, productivity, like Grade Point Average,
was not related to any of the measures of
morality.
The explanatory power of the Four
Component Model is observed, taking a
somewhat different approach, i.e., working
backward from disciplinary action to examining
deficiencies in the components. Baldwin
observed a relationship between the number of
malpractice claims and moral judgment scores,
noting that a high DIT score had a kind of
protective effect, insulating one from claims (33).
For dental practitioners referred for ethics
instruction, disciplinary actions were directly tied
to significant deficits in one or more of the
components (8, 9). Further, one consistent
observation, in addition to a deficiency in either
sensitivity, reasoning or implementation, is the
difficulty 27 of the 28 referrals had in articulating
the expectations of the profession. After targeted
instruction, directed toward role concept
development and remediation of one or more
other deficiencies, we observed measurable
improvements in performance, coupled with
documented changes in the behaviors that gave
rise to the disciplinary action. Further, to date,
there have been no cases of recidivism.4
Examining case studies bolsters the
understanding of the connection between the
components and behavior, and provides direction
for education.
185
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Conclusions
Analyzing data from the sources cited indicates:
1) striking individual differences among students
and practicing professionals on each of the
measures; 2) that competence on one of the
processes does not predict competence on
another; 3) that curricula of rather modest
duration can influence performance in
measurable ways (our curriculum consists of 43
contact hours); and 4) that strengths and
weaknesses in each of the processes are linked to
real-life ethical behavior. The findings described
in this paper support Rest’s contention that moral
failings can result from deficiencies in one or
more of the processes. Findings also support the
importance of attending to each when designing
curriculum. Further, whether a curriculum
promotes ethical development depends on
whether that curriculum incorporates the
elements of effective instruction.
Implications for Teaching and Assessing
for Research Integrity
If the objective is to develop thoughtful and
responsible scientists who act with integrity and
have broad understanding of their role and a
commitment to integrity in science, it is
important to do more than teach the rules and
policies that apply to the conduct of research.
Before engaging in case discussions, research
ethics teachers need to address the expectations
of a scientist. Students cannot be expected to
intuit the norms and values that undergird the
research enterprise. And, it is not clear that they
can “pick them up” from role models. The
expectations need to be explicitly taught and
formally assessed, preferably in writing. By
asking students to express the concepts in their
own words, and in writing, misperceptions can
be identified and addressed before they become
an issue. Once the expectations of the scientist
are clear, engage students in active learning
(using cases, if possible) to facilitate ethical
sensitivity, reasoning and problem solving.
When designing case materials, careful thought
should be given to the particular process that is
of concern. Too often, cases are written and
participants are asked: What should the
protagonist do? Such a question focuses on
problem solving, rather than problem
identification or moral reasoning. Certainly a
skilled facilitator can redirect attention to
reasoning or problem identification, but it is
sometimes much more difficult.
The author’s experience suggests that for
novice ethics teachers (which most of us are)
focusing on sensitivity, reasoning, and role
concept independently of one another will more
efficiently develop the skill needed for effective
problem solving. Ethics teachers should not
expect that carefully targeted courses will
develop the more advanced skills in ethical
reasoning that might result from courses in moral
philosophy. Yet, problem-based practice (using
cases) can be especially effective in helping
students recognize and subsequently avoid
personal interest arguments while strengthening
awareness and adherence to the rules of
responsible research conduct.
Notes
1. The referrals from the State Board came about because
some of the Board members have been involved in the
undergraduate curriculum for students. They wondered
whether violations of the Dental Practice Act reflected
ethical deficiencies that could be remediated by the
kinds of experiences we provided for students.
2. For a detailed account of the undergraduate dental ethics
curriculum, see Bebeau (1994).
3. There is extensive literature on the construct validity of
the DIT. See Rest, Narvaez, Bebeau, & Thoma (1999)
for a summary and references to the 400 published
studies using the DIT.
4. It is important to note that none of the practitioners
referred for remediation involved problems with
impulse control, substance abuse, mental illness, or
significant personality disorders.
Bibliography
1. Rest J. Morality. In: Mussen PH, series editor, Flavell J,
Markman E, volume editors. Handbook of child
psychology: vol. 3, cognitive development. 4th ed. New
York: Wiley; 1983. p. 556-629.
2. Bebeau MJ, Rest JR, Narváez DF. Beyond the promise:
a perspective for research in moral education.
Educational Researcher 1999;28 (4):18-26.
3. Kohlberg L. Stage and sequence: the cognitive
developmental approach to socialization. In: Goslin DA,
editor. Handbook of socialization theory. Chicago: Rand
McNally; 1969. p. 347-80.
4. Eisenberg N. Altruistic emotion, cognition and behavior.
Hillsdale (NJ): Erlbaum; 1986.
5. Lickona T. Educating for character. New York: Bantam;
1991.
6. Rest J, Narvaez D, Bebeau MJ, Thoma SJ.
Postconventional moral thinking: a neo-Kohlbergian
approach. Mahwah (NJ): Erlbaum; 1999.
7. Beauchamp TL, Childress JF. Principles of biomedical
ethics. 4th ed. New York: Oxford University Press;
1994.
186
–––––––––––––––––––––––––––––––––––––––––––––––– Bebeau, Influencing the Moral Dimensions of Professional Practice
8. Bebeau MJ. Influencing the moral dimensions of dental
practice. In: Rest JR, Narvaez D, editors. Moral
development in the professions: psychology and applied
ethics. Hillsdale (NJ): Erlbaum; 1994. p. 121-46.
9. Bebeau MJ. Continuing ethics education for
remediation. Paper presented at the meeting of the
American Association of Dental Schools; Orlando (FL);
1997.
10. Bebeau MJ, Rest JR, Yamoor CM. Measuring the ethical
sensitivity of dental students. J Dent Educ 1985;49
(4):225-35.
11. Bebeau MJ, Rest JR. The Dental Ethical Sensitivity
Test. Minneapolis (MN): Center for the Study of Ethical
Development; 1990.
12. Fravel BJ, Bebeau MJ. The role of the American
College of Dentists in assessing ethical sensitivity.
Poster session presented at the 25th annual meeting of
the Association of Moral Education; Minneapolis (MN);
1999 Nov.
13. Rest J. Development in judging moral issues.
Minneapolis (MN): University of Minnesota Press;
1979.
14. Rest J, Thoma S, Edwards L. Designing and validating a
measure of moral judgment: stage preference and stage
consistency approaches. Journal of Educational
Psychology 1997;89 (1):5-28.
15. McNeel SP. College teaching and student moral
development. In: Rest JR, Narvaez DF, editors. Moral
development in the professions: psychology and applied
ethics. Hillsdale (NJ): Erlbaum; 1994. p. 27-50.
16. Pascarella ET, Terenzini PT. Moral development. In:
How college affects students: findings and insights
from twenty years of research. San Francisco (CA):
Jossey-Bass; 1991. p. 335-68.
17. Rest JR, Narvaez D, editors. Moral development in the
professions: psychology and applied ethics. Hillsdale
(NJ): Erlbaum; 1994.
18. Bebeau MJ, Thoma SJ. The impact of a dental ethics
curriculum on moral reasoning. J Dent Educ 1994;58
(9):684-92.
19. Ponemon L, Gabhart D. Ethical reasoning research in
the accounting and auditing professions. In: Rest JR,
Narvaez D, editor. Moral development in the
professions: psychology and applied ethics. Hillsdale
(NJ): Erlbaum; 1994. p. 101-19.
20. Bebeau MJ. Influencing the moral dimensions of dental
practice: implications for the professions. Invited
address. American Educational Research Association;
New Orleans (LA); 2000 Apr 24-28.
21. Rest JR, Thoma SJ, Narvaez D, Bebeau MJ. Alchemy
and beyond: indexing the Defining Issues Test. Journal
of Educational Psychology 1997;89 (3), 498-507.
22. Thoma SJ, Rest JR. The relationship between moral
decision-making and patterns of consolidation and
transition in moral judgment development.
Developmental Psychology 1999;35:323-33.
23. Rest J, Narvaez D, Thoma SJ, Bebeau MJ. DIT2:
devising and testing a revised instrument of moral
judgment. Journal of Educational Psychology 1999;91
(4):644-59.
24. Yeap CH. An analysis of the effects of moral education
interventions on the development of moral cognition
[dissertation]. Minneapolis (MN): University of
Minnesota; 1999.
25. Bebeau MJ, Thoma SJ. “Intermediate” concepts and the
connection to moral education. Educational Psychology
Review 1999;11 (4):343-60.
26. Bebeau MJ, Born DO, Ozar DT. The development of a
Professional Role Orientation Inventory. Journal of the
American College of Dentists 1993;60 (2):27-33.
27. Thoma SJ, Bebeau MJ, Born DO. Further analysis of the
Professional Role Orientation Inventory [abstract 116]. J
Dent Res 1998 Mar;77 (Special Issue):120.
28. Thoma S. Moral judgments and moral action. In: Rest
JR, Narvaez D, editors. Moral development in the
professions: psychology and applied ethics. Hillsdale
(NJ): Erlbaum; 1994. p. 199-212.
29. Duckett LJ, Ryden MB. Education for ethical nursing
practice. In: Rest JR, Narvaez D, editors. Moral
development in the professions: psychology and applied
ethics. Hillsdale (NJ): Erlbaum; 1994. p. 51-70.
30. Krichbaum K, Rowan M, Duckett L, Ryden MB, Savik
K. The clinical evaluation tool: a measure of the quality
of clinical performance of baccalaureate nursing
students. J Nurs Educ 1994;33 (9):395-404.
31. Sheehan TJ, Husted SD, Candee D, Cook CD, Bargen
M. Moral judgment as a predictor of clinicial
performance. Evaluation and the Health Professions
1980;3:393-404.
32. Landsman M, McNeel SP. Moral judgment and
preference for public interest law practices among
beginning law students. Paper presented at the Annual
Meeting of the Association for Moral Education;
Glasgow, Scotland; 2000 July 7-11.
33. Self DJ, Baldwin DC Jr. Moral reasoning in medicine.
In: Rest JR, Narvaez D, editors. Moral development in
the professions: psychology and applied ethics.
Hillsdale (NJ): Erlbaum; 1994. p. 147-62.
187
Research Ethics in US Medical Education: An Analysis of Ethics Course
Syllabi
James M. DuBois, Center for Health Care Ethics, Saint Louis University, USA
Jill E. Ciesla, Center for Health Care Ethics, Saint Louis University, USA
Kevin E. Voss, Center for Health Care Ethics, Saint Louis University, USA
Keywords: Curriculum, Ethics, Medical education, Medical ethics, Medical schools
Medical education trains future physicians as medical practitioners. For this reason ethics education
for medical students has traditionally focused on themes revolving around the patient-physician
relationship: veracity, informed consent, fidelity, confidentiality, non-maleficence, and the like (1-3).
While many of these themes overlap with themes in research ethics, these ethics courses may be
inadequate for those future physicians who will engage in research of any kind – including clinical
trials, patient surveys, or program assessments (4-7). Research ethics introduces new and important
themes related to experimental design, interaction with communities, and the dissemination of
information (8,9). The well being of patients, physicians, and research institutions is at stake when
physicians fail to abide by rules for ethical research (9,10).
Recent, highly publicized failures to follow protocol at major medical centers reinforce the idea
that Institutional Review Boards (IRBs) are inadequate to ensure ethical research behavior. These
facts give rise to an important research question: To what extent is research ethics incorporated into
the ethics curriculum at medical schools in the United States (US), where future clinical researchers
are trained? This question takes on additional significance when one considers that medical students
may be engaged in clinical research in various forms even before completing undergraduate medical
studies (5,11,12).
This study builds upon a larger study that the first two authors of this paper conducted on the
ethics curriculum in US medical schools. DuBois and Ciesla analyzed syllabi from required ethics
courses in US medical schools with the aim of identifying and rank-ordering course objectives,
teaching methods, course content, and methods of student assessment (13). (The term “ethics course”
is used here to refer broadly either to a self-standing course or to a formal educational unit within a
larger course.) The present study analyzes in detail the content of the research ethics portion of
required ethics courses in the 4-year medical doctor (MD) curriculum at US medical schools. It
makes no attempt to describe responsible conduct of research (RCR) education at medical schools as
a whole, which frequently house graduate and postgraduate programs in the biomedical sciences, and
accordingly offer more extensive RCR courses outside of their MD programs.
Methods
This study was presented to the Institutional Review Board of Saint Louis University. It was approved
Corresponding author: James M. DuBois, PhD, DSc, Center for Health Care Ethics, Saint Louis University, Salus Center,
4th Floor, St. Louis MO 63104, 314-977-6660 (voice), 314-268 5150 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
as an exempt study given guarantees that
participation would be voluntary, subjects would
be adults, and confidentiality would be
maintained by publishing only aggregated data.
Instrument and Participants
The American Association of Medical Colleges
(AAMC) provided mailing labels for all
curriculum directors of 4-year medical colleges
in the US (N=121). A 1-page survey was sent to
all curriculum directors asking whether ethics is
taught as a formal required component, as an
elective, or not at all. It also inquired into the
year or years in which ethics is taught. The
survey further requested course syllabi for all
formal ethics components in the 4-year medical
curriculum.
Analysis
In the larger study, two researchers read all
syllabi using an open coding method to produce a
comprehensive list of all elements found in the
syllabi that fell into one of four generic
categories: (1) course objectives, (2) teaching
methods, (3) course content, and (4) student
assessment methods. All other statements (e.g.,
pertaining to class times, locations, and
instructors) were ignored. The specific elements
of the syllabi were then placed into categories.
These categories were used to create variables in
a SPSS database. Schools, rather than syllabi,
constituted cases in the database: if a school had
more than one required ethics component, data
from all required course syllabi were entered into
that case. Data from 10 syllabi (17%) were
entered by two researchers to establish interrater
reliability.
The present study identified those syllabi that
included content on research ethics.
The research ethics sections of syllabi were
read using an open-coding method to generate a
comprehensive list of research ethics content.
The results of this open-coding process were then
placed into general categories. These categories
were entered into an expanded SPSS database.
Statistical analysis aimed above all to provide
descriptive data on the frequency of various
research ethics content. Pearson’s r was used to
test whether the mean number of content areas
covered was significantly correlated with either
class size or tuition cost.
Results
Surveys were returned by 72% of the schools
(n=87). Seventy-nine percent (n=69) of these
schools claimed to require a formal ethics course.
Of these schools, 84% (n=58) provided ethics
course syllabi. The two raters categorized items
the same in 90% of the cases. In the predecessor
study, analysis and codification of all syllabi
identified 10 course objectives, 8 teaching
methods, 39 content areas, and 6 methods of
student assessment. The mean for individual
schools was 3 objectives, 4 teaching methods, 13
content areas, and 2 methods of assessment.
Among the 39 different content areas,
research ethics ranked 11th. Twenty-three of the
58 syllabi (39.6%) addressed research ethics in
some fashion. Analysis of the research ethics
sections of these syllabi revealed 82 specific
themes that fall under 17 different general
categories.
Table I (below) presents these 17 general
categories in rank order, along with the specific
themes that fall under each category. It further
indicates where the categories and specific
themes overlap with the US Public Health
Service’s (PHS) “Core Instruction Areas” for
courses on the Responsible Conduct of Research
(RCR) (14). (This policy of December 1, 2000
was suspended by the Bush administration in
February 2001 pending further study. This paper
refers to the policy because it continues to serve
as a model for many institutions and it remains
under discussion among legislators and policy
makers.)
The average number of general research
ethics topics addressed in these 23 syllabi is 6,
with individual schools covering anywhere from
1 to 11 topics. Only six topics were covered by
more than half of those syllabi that address
research ethics. In rank order these are: clinical
trials; informed consent; general ethics of human
subject research; government committees and
regulations; history and background to research
ethics; and protecting vulnerable populations. No
research ethics topic was covered by more than
21% of the 87 participating schools. The number
of research ethics topics covered did not correlate
significantly with either school enrollment
(r=.10, p<.45) or tuition costs (r=.10, p<.43).
Discussion
While Mastroianni and Kahn conducted a useful
and informative pilot study of NIH grantee
institutions’ training efforts in RCR, this study is
the first to examine comprehensively the RCR
curriculum in US medical programs. Our study
190
–––––––––––––––––––––––––––––––––––––––––––––––––––––––– DuBois, et al., Research Ethics in US Medical Education
exposes two possible causes for concern. First,
too few medical schools teach research ethics in
any fashion within their MD program. No topic
in research ethics – including clinical trials – is
covered by more than 21% of all medical
schools. The topic of Institutional Review Boards
is covered by less than 13% of medical schools,
despite the fact that medical researchers are most
likely to work precisely with human subjects.
Second, it appears that important topics are
wholly missing even in those programs that teach
research ethics. This becomes clear when
comparing the specific research ethics topics
covered within medical ethics syllabi to the
“Core Instruction Areas” PHS identified for RCR
education (14). For example, the first five of nine
core areas PHS identifies (data acquisition,
management, sharing, and ownership; mentor /
trainee responsibilities; publication practices and
responsible authorship; peer review; and
collaborative science) seem wholly missing from
these syllabi. (The only possible exception is one
syllabus that mentions industry/university
relationships.)
It is, of course, possible that some of these
topics are covered under other general headings
(e.g. ‘collaborative research’ might be discussed
under ‘clinical trials’). This is one limitation of
the method used: a topic is identified only if it
explicitly appears on the course syllabus. This
means that syllabi using only very general
headings will be shortchanged. Nevertheless, a
course syllabus should be a reliable statement of
the objectives and content of a course, and most
syllabi were quite detailed (as the larger study
demonstrated). Thus, it seems safe to conclude
both that very few MD programs discuss research
ethics and that those that do ignore at least half of
the topics PHS wants to see addressed.
However, the significance of these findings
cannot be firmly established until other questions
are answered:
• To what extent are medical students participating in clinical research?
• Are current requirements for RCR instruction
likely to be successful in targeting future
physicians who are funded by private
industry?
• To what extent do clinical researchers encounter special ethical topics that are not
covered in general RCR courses?
These questions remain unanswered. Literature
in academic medicine has addressed the roles of
undergraduate medical students in research
(5,11,12). However, the prevalence and extent of
students’ roles and whether they are specifically
listed in study protocols remains unknown. Thus,
it is difficult to know whether education in RCR
is a pressing need for medical students, or
whether these years might be viewed simply as a
convenient time to introduce education in RCR.
Research has shown that private industry is
now funding more research than is the
government (15). Government requirements
regarding RCR instruction pertain only to
government-funded research, and according to at
least one study, two-thirds of NIH grantee
institutions require RCR instruction only to the
extent that the government mandates it (16).
These facts suggest that a “blanket” approach to
educating future physicians would be the safest
route to ensuring RCR instruction for clinical
researchers. However, given the scope of recent
government requirements, such a blanket
approach would have to be initiated by a
professional institution like the AAMC.
Finally, it is difficult to anticipate how well
the RCR programs that are currently being
mandated will address the specific ethical
concerns that arise in clinical, medical research.
This study has shown that 13 of our 17 categories
could easily be subsumed under just one PHS
Core Area: #6, Human Subjects. This suggests
that typical RCR instruction aims to cover a
broad range of issues that arise in research (such
as authorship, peer review and the treatment of
animals), whereas physicians feel the need for a
highly focused and intensive treatment of human
subject research. The years of medical school
may be the best or only time to provide this sort
of special-tailored education in RCR.
While this study has provided new answers
to questions about the current educational
training of medical students in RCR, it has also
managed to bring new questions to the fore. Only
after these questions are answered, will the
significance of this study’s findings be properly
understood.
Acknowledgements
The authors thank Doris Rubio, Ph.D., for
assistance in study design. They thank all
participating medical schools for their
cooperation. They thank Gerard Magill, Ph.D.
and Dennis Daly, SJ for securing funding for this
project. This study was funded by the Marchetti
Fund at Saint Louis University.
191
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Bibliography
1. Miles SH, Lane LW, Bickel J, Walker RM, Cassel CK.
Medical ethics education. Coming of age. Acad Med
1989;64:705-14.
2. Fox E, Arnold RM, Brody B. Medical ethics education:
past, present and future. Acad Med 1995;70:761-9.
3. Musick DW. Teaching medical ethics. A review of the
literature from North American medical schools with
emphasis on education. Medicine, Health Care, and
Philosophy 1999;2:239-54.
4. Mastroianni AC, Kahn JP. The importance of expanding
current training in the responsible conduct of research.
Acad Med 1998;73:1249-54.
5. Taylor I, Baum M, Cooper A, Johnston IDA. Dilemmas
facing surgical research in the ‘90s. Ann R Coll Surg
Engl 1991;73 Suppl 4:70-2.
6. Friedman PJ. Research ethics: a teaching agenda for
academic medicine. Acad Med 1990;65:32-3.
7. Jonsen AR. Clinical research ethics. Journal of Medical
Education 1975;50: 918-9.
8. Brody B. The ethics of biomedical research. An
international perspective. New York: Oxford University
Press; 1998.
9. Levine RJ. Ethics and regulation of clinical research.
Baltimore: Urban & Schwarzenberg; 1986.
10. Spicker SF. The use of human beings in research: with
special reference to clinical trials. Boston: Kluwer;
1988.
11. Christakis N. Do medical student research subjects need
special protection? IRB: A Review of Human Subject
Research 1985;7:1-4.
12. Makarushka JL, Lally J. Medical schools, clinical
research, and ethical leadership. Journal of Medical
Education 1974;49:411-8.
13. DuBois, JM, Ciesla, JE, Voss, KE. Ethics education in
US medical schools. A study of syllabi. Academic
Medicine. In press.
14. Office of Research Integrity and Public Health Service.
PHS policy on instruction in the responsible conduct of
research. Adopted and published on December 1, 2000.
Available from: URL: http://www.ori.dhhs.gov
15. Cho MK, Shohara R, Schissel A. Policies on faculty
conflicts of interest at US universities. JAMA 2000;
284: 2203-8.
16. Mastroianni AC, Kahn JP. Encouraging accountability in
research: a pilot assessment of training efforts.
Accountability in Research 1999; 7: 85-100.
192
–––––––––––––––––––––––––––––––––––––––––––––––––––––––– DuBois, et al., Research Ethics in US Medical Education
Table I: Rank Order and Content of the Research Ethics Categories
An asterisk * followed by a number indicates that the general category or specific topic overlaps with a PHS Policy “Core
Instructional Area.” The number indicates which of nine instructional areas it overlaps with.
‘Percent valid’ indicates how often a research ethics topic is included in those syllabi from the 23 schools that actually teach
research ethics.
‘Percent all’ indicates how often a research ethics topic is included among all participating schools (i.e., the 87 schools that
returned a survey).
1. CLINICAL TRIALS (*6) – 78% of valid / 21% of all
•
•
•
•
•
•
•
•
•
•
•
Therapeutic vs. non-therapeutic research
Person as patient vs. research subject
Physician as clinician vs. physician as scientist
Selection of subjects for clinical trials
Randomization
Patient as research subject vs. health research subject
Ethics of medical students’ roles in clinical research
Drug testing and the role of the FDA
Whether scientific methods provides sole criterion for treatment efficacy
Industry / university relationships (*possibly 5 & 9)
Types of clinical trials
2. INFORMED CONSENT (*6) – 70% of valid / 18% of all
•
•
•
•
•
•
Informed consent in clinical vs. research setting
Sample consent form for adults
Emergency waiver of informed consent
Coercion
Deception – active and passive
Placebos
3. GENERAL ETHICS OF HUMAN SUBJECT RESEARCH (*6) – 65% of valid / 17% of all
• Ethics of human experimentation
• Justification of research involving human subject
• Challenges to human subject protections
4. GOVERNMENT COMMITTEES & REGULATIONS (*6 & others) – 61% of valid / 16% of all
• Belmont report
• President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research (197983)
• National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research (1974-78)
[Published Belmont Report]
• Federal regulations
• National Bioethics Advisory Committee
• Declaration of Helsinki
• Practice and regulations
• OPRR reports, Protection of Human Subjects
• Title 45, Code of Federal Regulations, part 46 (1994)
• Nuremberg Code (as living document)
5. HISTORY AND BACKGROUND OF RESEARCH ETHICS – 57% of valid / 15% of all
•
•
•
•
•
•
•
•
•
•
Nazi experimentation / Holocaust (awareness of attitudes toward)
Nuremberg Code (as historical document)
Tuskegee study of syphilis (awareness and attitudes toward)
Abuses and errors of early eugenics
“Frankenstein”
Sloan-Kettering experiments
Willowbrook experiments
Henry Beecher revisited (article by DJ Rothman)
Introduction to sulfonamides revisited (articles by BH Lerner)
Research in the Hippocratic Oath (i.e., the fact that it is not addressed therein)
193
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
6. PROTECTING VULNERABLE POPULATIONS (*6) – 52% of valid / 14% of all
•
•
•
•
•
•
Minorities
Newborns, Infants, Children
Soldiers
Prisoners
Mentally ill
AIDS patients
7. IRB (*6) – 48% of valid / 13% of all
• IRB issues
• Definition of research / Novel therapy vs. research
8. RESEARCH INTEGRITY & MISCONDUCT (*8 & 9) – 39% of valid / 10% of all
•
•
•
•
•
•
•
•
Accuracy of published data
Research fraud (*8)
Appearance of impropriety
Scientific misconduct (*8)
Scientific integrity
Appropriate credentials
Research quality guidelines for both academic and non-academic environments
Conflicts of interest (*9)
9. ETHICAL PRINCIPLES IN HUMAN SUBJECT RESEARCH (*6) – 39% of valid / 10% of all
•
•
•
•
•
•
Respect autonomy
Do good (beneficence)
Fairness / justice
Avoid harm to subjects (non-maleficence)
Justify level of risk
Apply process of ethical decision making to research ethics
10. ANIMAL EXPERIMENTATION (*7) – 30% of valid / 8% of all
• Animal rights
• Use of animals for research
• Poor living conditions for research animals
11. GENETIC RESEARCH AND THERAPY (*6) – 26% of valid / 7% of all
•
•
•
•
•
•
Genetic research
Germ-line therapy
Somatic cell genetic therapy
National Human Genome Research Institute
Genetic information and privacy
Cystic fibrosis research
12. RESEARCH AND THE SOCIAL GOOD (*6) – 22% of valid / 6% of all
•
•
•
•
•
Medicine and the goals of society
Research in the international context
Social utility of research
Relationship between ethics, science, and technology
Balancing society’s mandates, competing pressures to innovate
13. MINIMIZING RISKS (*6) – 22% of valid / 6% of all
• Establishing gold standard
• Asking whether risk is proportionate to benefit
14. SUBJECT SELECTION (*6) – 13% of valid / 3% of all
• Ensuring the inclusion of women, children and minorities (a concern of justice, rather than protection)
15. EMBRYO AND FETAL RESEARCH (*6) – 9% of valid / 2% of all
• Stem cell research
• Research on live-born fetuses
16. EPIDEMIOLOGY (*6) – 4% of valid / 1% of all
• Ethics of epidemiology
17. MILITARY RESEARCH ETHICS (*6) – 4% of valid / 1% of all
• Experiments related to weaponry
• Using compounds not fully tested in a wartime situation
194
Teaching Ethics in Biomedical Science: Effects on Moral Reasoning Skills
Elizabeth Heitman, University of Texas School of Public Health, Houston, Texas, USA
Patricia J. Salis, University of Texas School of Public Health, Houston, Texas, USA
Ruth Ellen Bulger, Uniformed Services University of the Health Sciences, Bethesda,
Maryland, USA
Key words: Defining Issues Test, DIT, Ethics education, Evaluation, Responsible conduct of research
Academic institutions that train professionals play an important role in ensuring that trainees learn the
ethical norms of their respective disciplines, and that they learn to behave ethically from the start of
their professional lives. The National Institutes of Health requirement that funded research training
programs include education in scientific integrity has made formal courses on the responsible conduct
of research increasingly common in academic medical centers and research universities.
There is still no consensus on what constitutes the most appropriate subject matter, format,
methods, or faculty for teaching the responsible conduct of research. The objectives of general
courses on the responsible conduct of research and scientific integrity typically include increasing
students’ understanding of the norms of scientific practice, their recognition of ethically problematic
situations in science, and their ability to analyze and respond to such situations in a morally mature
manner. Courses vary in the specific content, the number of contact hours, the format (lecture, smallgroup discussion, video or web-based tutorials), and the instructors’ professional background and
ethical expertise. The effectiveness of available courses probably also varies. Studies of how students
are affected by formal ethics courses in such disciplines as engineering, law, dentistry, medicine,
nursing, journalism, accounting, veterinary medicine, and social work have found that course design
influences the extent to which students’ ethical reasoning skills change during the courses (1-3). Such
evaluation in the area of scientific integrity, however, is still in its infancy.
The syllabi of courses on the responsible conduct of research in several institutions suggest that
such courses present at least three different kinds of instruction to students. The first is the “how-to”
of science, in which the practical, procedural dimensions of science, rather than its ethical
dimensions, are the focus: how to devise an experiment, give a talk, or write a manuscript. The
second kind of instruction relates to the rules, regulations, and professional norms articulated by the
organizations in which scientists work, their professional societies, and/or the government: how to
make experimental data available for use, how to address suspected research misconduct, and how to
deal ethically with animal and human subjects. Ethical considerations are often addressed as an
aspect of these practical issues. Lecture and individual reading assignments are effective mechanisms
for teaching both of these traditional types of subject matter, and students’ understanding and
retention can be evaluated by an objective written (including computerized) or oral exam.
Corresponding author: Elizabeth Heitman, PhD, University of Mississippi Medical Center, 2500 North State Street, Jackson,
MS 39216, 601-815-1134 (voice), 601-366-3262 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
The third type of instruction presented by
these courses relates to students’ ability to
recognize the ethical aspects of problems that
they encounter in their research, and their ability
to address these issues in a considered way. This
instruction involves their developing moral
reasoning skills rather than simply
comprehending information, and it frequently
uses case discussion or problem-based learning.
Two decades ago the Hastings Center Project on
the Teaching of Ethics proposed three criteria for
evaluating the effectiveness of such instruction:
1) whether the student understands the central
concepts; 2) whether the student can make
cogent oral and written ethical arguments; and
3) whether the student can recognize ethical
problems and examine them rationally (4). This
evaluation is typically conducted through a more
subjective examination using actual case
analysis, possibly in a written or oral exam, but
ideally in a more interactive setting.
The Hastings Center Project emphasized that
helping students develop skills to recognize and
analyze ethical issues and stimulating their moral
imagination are fundamental to the effective
teaching of ethics. The Association of American
Medical Colleges handbook, Teaching the
Responsible Conduct of Research through a Case
Study Approach (5), has also stressed the need to
enhance students’ ethical awareness and
problem-solving skills in formal education on the
responsible conduct of research. Ideally, the
courses should have a positive effect on students’
actual and future behavior, helping individuals
avoid ethically problematic behavior and
enhancing their ability to resolve unfamiliar
ethical conflict appropriately.
After several years of teaching a formal
course on the responsible conduct of research at
the University of Texas Health Science Center at
Houston, the course’s organizers sought to assess
its effects and to determine what outcomes could
be evaluated formally. The course, The Ethical
Dimensions of the Biomedical Sciences,
originated in 1984 as an institutional response to
an incident with a foreign graduate student that
would have been considered plagiarism for a
student schooled in the United States (6, 7).
Consideration of the case highlighted the
administration’s and faculty’s need to articulate
the university’s ethical expectations and to teach
U.S. academic and professional standards to all
students. The primary objectives of the course
subsequently developed by Dr. Ruth Bulger, and
later continued by Drs. Stanley Reiser and
Elizabeth Heitman, have been to encourage
students’ interest in the ethical development and
goals of science, and to teach students to prevent,
recognize, analyze, and resolve ethical conflicts
in the daily conduct of their work (8).
From the beginning, the course has used a
combination of formal reading assignments,
didactic lecture, and small-group case discussion
to address a wide variety of issues in the
responsible conduct of research. Its faculty have
always included both ethicists and bench and
clinical researchers from various disciplines, both
as lecturers and as discussion leaders. Most are
senior faculty. Since 1988, the course has been a
requirement for graduation from the Graduate
School of Biomedical Sciences, and it is an
elective for graduate students in the School of
Public Health. For the past four years,
approximately 120 students have enrolled in the
course each fall, including 90+ from the
Graduate School of Biomedical Sciences’ 22
degree programs, 10-15 students from the School
of Public Health’s 11 degree programs, and
several post-doctoral fellows from the UT
Medical School and MD Anderson Cancer
Center. Students in biomedical sciences typically
take the course in their first semester, while
others often enroll in the second half of their
formal graduate study.
Objective written examinations demonstrated
that the course effectively enhanced students’
knowledge and understanding of both the
practical how-to of science and the rules,
regulations, and professional norms of research
that the course addressed. Written analysis in the
final exam demonstrated students’ ability to
identify and consider ethical issues. Students’
course evaluations also confirmed that most of
them found the course valuable to their
professional development. However, the faculty
wanted to assess the more comprehensive effects
of the course on students’ professional attitudes
and behaviors.
To affect students’ current behavior and
shape their future action, instructors of courses in
the responsible conduct of research must have
three things: 1) an effective way to teach desired
behaviors; 2) an effective way to motivate
students to adopt these behaviors; and 3) a
reliable way to measure behavior change. In a
broad literature review, we found no clearly
identifiable, successful method for teaching
ethical behavior or motivating students to act
196
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Heitman, et al., Teaching Ethics in Biomedical Science
ethically. While there has been work on how best
to evaluate students’ comprehension and
retention of information related to ethical
conduct, we found no generally accepted way to
measure the presumed beneficial effect of ethics
courses on behavior.
In the absence of accepted measures of
behavior change and future practice, surrogate
measures of the effectiveness of courses on the
responsible conduct of research are needed.
Bebeau (9) and her colleagues have developed a
set of teaching materials for education in the
responsible conduct of research that considers
four psychological processes in the decision to
act ethically: moral sensitivity (the ability to
interpret a moral situation and the effects of
various courses of action on the parties
involved); moral reasoning (judgment about
which course of action is right); moral
commitment (intention to do what is right) and
moral perseverance (the ability to follow through
with ethical behavior). Their method of
evaluating the effectiveness of courses that use
the group’s instructional materials assesses the
essential components of students’ moral
discernment and moral reasoning.
Efforts to define, implement, and assess
education in the responsible conduct of research
in graduate science programs have parallels in
medical education, where considerable work has
been done on the teaching of professional ethics
and the evaluation of such teaching. The effects
of ethics courses on medical students’ moral
reasoning skills have been studied since the late
1970s (10). Such evaluations have linked
different types of ethics education with changes
in students’ moral reasoning, and have suggested
that case-based discussion can significantly
increase students’ moral reasoning ability.
The Defining Issues Test (DIT) is the
instrument used most frequently to measure
moral reasoning skills and the effects of
education on moral reasoning. The DIT was
developed by James Rest and colleagues at the
University of Minnesota Center for the Study of
Ethical Development (11). The test is a
standardized, computer-scored test that is easily
administered to groups. It is based on Kohlberg’s
theory of cognitive moral development, which
considers the principle of justice as the highest
moral good. The DIT presents six morally
problematic scenarios; the subject ranks the
importance of various moral criteria for judging
how to act, then chooses a course of action.
Scores are reported in terms of a P%, which
measures the extent of “principled” reasoning
behind the individual’s assessment of the cases.
Cross-cultural applications have found that DIT
scores increase consistently with subjects’ age
and education level.
This study explored whether two offerings of
our course on The Ethical Dimensions of the
Biomedical Sciences had an effect on students’
principled moral reasoning, as measured by the
DIT.
Methods
Following an IRB-approved protocol, a total of
215 graduate students who were enrolled in The
Ethical Dimensions of the Biomedical Sciences
course were asked to complete the DIT at the
beginning (before-course) and the end (aftercourse) of the 1997 and 1998 classes. Use of
individual codes protected students’
confidentiality. Computerized scoring by the
University of Minnesota Center for the Study of
Ethical Development generated P% scores.* The
analyses used students’ change scores — the
after-course test score minus the before-course
test score — as the data. A preliminary analysis
of differences in change scores between the 1997
and 1998 classes (t-test, independent samples)
was performed to determine whether it was
possible to combine the data from the two
classes. Next the effectiveness of the course in
improving students’ principled judgment by was
tested directly analyzing whether their change
scores differed significantly from zero (t-test,
matched pairs). Finally, an analysis of variance
(ANOVA) test was run to determine whether
students’ gender or country of undergraduate
education (US or non-US) was related to
differential change scores.
Results
One hundred seventy-two students (80% of the
original 215 students) completed both a beforecourse and an after-course test, 95 students in
1997 (87% of 109) and 77 in 1998 (73% of 106)
(Table 1). One or both tests from 14 of these
172 subjects were excluded from analysis based
on scoring criteria used by the University of
Minnesota Center for the Study of Ethical
Development. The final sample therefore
contained 158 students who had valid scores for
both the before-course and the after-course tests.
Change scores did not differ significantly
between the 1997 and 1998 classes (t=-0.88,
197
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
p=0.38), so a combined analysis of the two
classes was possible.
The primary analysis assessed the course’s
effect on principled judgment: It revealed that
the students showed no significant after-course
improvement in principled judgment, as
measured by the DIT P% score (Figure 1, Table
2). Indeed, the pattern in six of the eight subgroups (Figure 2) was for after-course scores to
drop slightly.
Follow-up analyses of the influence on
change scores of students’ gender and location
of undergraduate schooling indicated that neither
gender nor location of education had a significant
effect for the combined 1997 and 1998 courses
(Table 3), for the 1997 students alone (Table 4),
or for the 1998 students alone (Table 5). For the
combined group and the 1997 group, there was
no significant interaction between the gender
factor and the location-of-schooling factor, but
this interaction was significant in the 1998 group
(Table 5). The 1998 data in Figure 2 suggest that
this result arose from the distinctive pattern
among men educated in the U.S. Their aftercourse scores declined somewhat, while those of
both groups of women and of men not educated
in the U.S. either improved very slightly or
stayed essentially the same.
Conclusions
The finding that no significant change had
occurred in P% scores after the course on the
Combined
1997 & 1998
Classes
215
1997
Class
1998
Class
109
106
No. people who took 2 tests
172
( 80 % of 215)
95
(87% of 109)
77
(73% of 106)
No. test pairs sent for scoring
No. test pairs not used
Final no. people or test pairs
172
14
158
(92% of 172)
(73% of 215)
95
11
84
(88% of 95)
(77% of 109)
77
3
74
(96% of 77)
(70% of 106)
No. people who took at least 1 test
Table 1. Composition of final study sample.
* one or both tests in pair purged by scorers for invalidity; one of pair purged by us due to absence of valid pairmate; scorers failed to process test pair
Group
Change
Score
Mean
(SD)
t Value *
p Value
1997 & 1998 Combined
-1.27
(11.75)
-1.3 6
0.17
1997
-2.05
(11.64)
-1.6 1
0.11
1998
-0.40
(11.89)
-0.2 9
0.78
Table 2. Statistical evaluation of course’s effect on DIT P% scores: t-tests
(matched pairs) of change scores (after-course minus before-course) in
combined classes and in each class alone.
198
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Heitman, et al., Teaching Ethics in Biomedical Science
Source of Variance
Degrees of Freedom
Gender
1
Country of education
1
Gender X country interaction
1
Error
154
Total
157
F Value
0.21
0.54
0.90
p Value
0.65
0.46
0.34
Table 3. Statistical evaluation of effect of gender and location-of-schooling on DIT P% scores: analysis of
variance of change scores (after-course minus before-course) in combined classes.
Source of Variance
Gender
Country of education
Gender X country interaction
Error
Total
Degrees of Freedom
1
1
1
80
83
F Value
0.16
0.09
0.25
p Value
0.69
0.77
0.62
Table 4. Statistical evaluation of effect of gender and location-of-schooling on DIT P% scores: Analysis of
variance of change scores (after-course minus before-course) in 1997 class.
Source of Variance
Gender
County of education
Gender X country interaction
Error
Total
Degrees of Freedom
1
1
1
70
73
F Value
1.75
0.97
4.86
p Value
0.19
0.33
0.03
Table 5. Statistical evaluation of effect of gender and location-of-schooling on DIT P% scores: Analysis of
variance of change scores (after-course minus before-course) in 1998 class.
responsible conduct of research was a surprising
and frustrating outcome, given the course’s
perceived value within the university and the
number of studies that report significant changes
in students’ moral reasoning skills after similar
courses in professional ethics. Even more
perplexing was that students in most sub-groups
actually showed slight declines in P% scores
after the course.
Upon reflection, the authors concluded that
principled moral reasoning is only one of a
number of skills and concepts that we hope to
teach and foster in our course. Much of the
material and related discussion in the course
focuses on common conflicts and practical
ethical strategies in research and collegial
interaction. Rest and colleagues (12) noted in
1999 that Kohlberg’s theories, and thus the DIT,
address formal ethical structures of society, what
they call macromorality, and do not illuminate
the micromoral phenomena of personal, face-toface interactions in everyday life. Thus these
null findings suggest that it is essential to ask
different questions or use different methods to
evaluate the complex issue of the outcomes of
the course.
The establishment and ultimate success of
education in the responsible conduct of research
will require effective means of assessing the
impact of such programs on students’ knowledge,
awareness, and moral reasoning. Under the most
recent proposal requiring such education in all
Public Health Service-funded institutions, a wide
variety of formats appear to satisfy the new
credentialing standards. Suggested options range
from semester-long academic courses to day-long
workshops to hour-long web-based tutorials, to
self-study reading programs. As academic
research institutions develop the expertise needed
to provide education in the responsible conduct
of research, mechanisms must also be developed
to assess the extent to which these different
formats are effective in enhancing participants’
moral reasoning skills. Recent observations
reported by Bebeau and colleagues suggest that
some apparently unchanged DIT scores may
mask important differences in moral sensitivity
and reasoning (13). Expanded use of the DIT
should strive to uncover all significant changes in
moral reasoning in order that academic courses
can target their educational intervention
appropriately.
However, if the objective of education in the
199
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
responsible conduct of research is to shape the
behavior of researchers and to reform the culture
of research, methods for evaluating such change
must be developed, and instructors must learn
how to present the rules, regulations, and
professional norms of science in a way that
motivates researchers to adhere to them.
Note
* The Center generated the P% scores using its
new system of validity checks, which should be
considered when comparing these results to those
of older studies.
Bibliography
1. Schlaefli A, Rest JR, Thoma SJ. Does moral education
improve moral judgment? A meta-analysis of
intervention studies using the Defining Issues Test.
Review of Educational Research 1985; 55: 319-20.
2. Self DJ, Baldwin, DC Jr., Wolinsky FD. Evaluation of
teaching medical ethics by an assessment of moral
reasoning. Medical Education 1992; 26: 178-184.
3. Self DJ, Wolinsky FD, Baldwin, DC Jr. The effect of
teaching medical ethics on medical students’ moral
reasoning. Academic Medicine 1989; 64: 755-759.
4. Callahan D, Bok S. Hastings Center project on the
teaching of ethics. In: Callahan D, Bok S, eds., Ethics
teaching in higher education. New York: Plenum Press,
1980.
5. Association of American Medical Colleges. Teaching
the responsible conduct of research through a case study
approach. Washington, DC: AAMC, 1994.
6. Bulger RE, Reiser SJ, Studying science in the context of
ethics. Academic Medicine 1993; (Suppl. 3): 55-59.
7. Reiser SJ, Heitman E. Creating a course on ethics in the
biological sciences. Academic Medicine 1993, 68: 876879.
8. Bulger RE, Heitman E, Reiser, SJ. The ethical
dimensions of the biological sciences. New York:
Cambridge University Press, 1993.
9. Bebeau M. Moral reasoning in scientific research. Cases
for teaching and assessment. Bloomington, IN: Poynter
Center for the Study of Ethics and Assessment, 1995.
10. Self DJ, Baldwin DC Jr. Moral reasoning in medicine.
In Moral development in the professions: Psychology
and applied ethics. In: J Rest & D. Navares, eds.,
Hillsdale, NH: Lawrence Erlbaum Associates, 1994.
11. Rest JR. Development in judging moral issues.
Minneapolis, MN: University of Minnesota Press, 1979.
12. Rest J, Narvaez D, Bebeau MJ, Thoma SJ.
Postconventional moral thinking: A neo-Kohlbergian
approach. Mahwah, NJ: Lawrence Earlbaum Associates,
1999: 15.
13. Bebeau MJ. Influencing the moral dimension of
professional practice: Implications for teaching and
assessing for research integrity (abstract). A Research
Conference on Research Integrity, Bethesda. MD,
November 19, 2000.
200
201
0
10
20
30
40
50
60
70
47.13
Women, US
47.58
Before course
Figure 1.
43.01
Women, nonUS
43.93
After course
36.91
Men, US
40.39
40.17
Men, nonUS
40.04
Mean DIT P% Scores Before and After C
Combined 1997 and 1998 Classes
* Data derived with new
validity checks; norms
derived with old validity
checks.
< Junior high 20.00
< Senior high 31.03
< BA graduates 44.85
< College 43.19
< Moral philosophers 65.1
NORMS*
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Heitman, et al., Teaching Ethics in Biomedical Science
202
0
10
20
30
40
50
60
70
47
44 43
38 37
39
38
After course
US
nonUS
US
nonUS
Women
Men
1997
50
Before course
47
44 43
44
36
41
43
US
nonUS
US
nonUS
Women
Men
1998
45
Figure 2. Mean DIT P% Scores Before and After C
1997 Class and 1998 Class
* Data derived with new
validity checks; norms derived
with old validity checks.
< Junior high 20
< Senior high 31
< BA graduates 45
< College 43.19
< Moral philosophers 65
NORMS*
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Fostering Research Integrity through Educational Programs: Lessons
Learned at the University of Minnesota
Jeffrey P. Kahn, Center for Bioethics, University of Minnesota, USA
Peggy A. Sundermeyer, Office of the Vice President for Research, University of Minnesota,
USA
Muriel J. Bebeau, Department of Preventive Sciences, University of Minnesota, USA
Virginia S. Seybold, Department of Neuroscience, University of Minnesota, USA
Keywords: RCR curriculum development, Research integrity education programs, Responsible conduct of
research education
The implementation of a Public Health Service (PHS) policy on Instruction in the Responsible
Conduct of Research (RCR) would be a significant challenge to universities because of its broad
inclusion of personnel involved in research. The University of Minnesota is already meeting this
challenge with the delivery of a comprehensive educational program to over 2,000 faculty and
principal investigators (PIs) in calendar year 2000.
The University of Minnesota is a large, land-grant institution. The intellectual diversity of the
institution is reflected in its 21 collegiate units, 3,000 tenure and tenure-track faculty, and 10,000
graduate students enrolled in 150 masters and doctoral programs. The foundation of our educational
programming in RCR developed centrally, early in the 1990’s, to support the educational requirement
of training grants. These programs were expanded to faculty in the mid-90’s in response to growing
institutional and national concern about misconduct in research. The current curriculum is the result
of an institutional corrective action plan initiated by National Institutes of Health (NIH) in 1997.
Therefore, a unique set of circumstances required the University of Minnesota to implement a
comprehensive educational program in RCR before announcement of the PHS policy on Instruction
in RCR.
Our goal is to share the experience of our institution in order to aid others in the development of
programs to meet the requirements of the PHS policy. Points of discussion within the context of the
evolution of the educational program at Minnesota include 1) policy as framework for education,
2) development and delivery of the curriculum, 3) resources and financial investment, and 4)
evaluation.
Policy as Framework in Education
One strength of the educational initiative at the University of Minnesota is that the importance of
RCR is reflected in institutional policies. The Board of Regents, the administrative authority of the
Corresponding author: Jeffrey Kahn, Center for Bioethics, University of Minnesota, N504 Boynton, 410 Church St. SE,
Minneapolis, MN 55455-0346, 612-624-9440 (voice), 612-624-9108 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
University, passed the Code of Conduct in 1996.
This policy pertains to all members of the
University community and states that we will
“adhere to the highest ethical standards of
professional conduct and integrity.” While
affirming the common values of research and
scholarship, it is a clear demonstration of
institutional ownership of these values. In 1999,
the Board of Regents passed a revised policy on
Principal Investigator Eligibility on Sponsored
Projects. This policy requires PIs to complete a
required education in RCR before any awarded
funds are released for spending. The policy was
implemented March 1, 2001, preceding the PHS
policy by approximately two years and providing
the motivation for compliance with the
educational requirement. Both policies can be
viewed at http://www.ospa.umn.edu/policy/
respolcy.htm.
The University of Minnesota has a strong
tradition in faculty governance, so it is not
surprising that the faculty senate has also
promoted RCR. In 1999, the faculty senate
passed the policy on Education in Responsible
Conduct of Sponsored Research and Grants
Management (see http://www1.umn.edu/usenate/
policies/grantsmgmt.html). Whereas this policy
reiterates the expectation that PIs and project
personnel have the responsibility to behave in
accordance with the highest ethical standards, it
also defines the responsibility of the University
to provide individuals involved in research with
information and resources that support
responsible conduct. The policy describes the
framework for implementing educational
programs under the leadership of the Vice
President for Research and Dean of the Graduate
School. It outlines the formation of three
advisory committees, one for each major
constituency: Academic personnel (including
faculty and academic administrators), research
staff (including graduate and postdoctoral
trainees as well as project staff), and
administrative staff (including accounting and
secretarial support). The charge to each of these
committees is to define the educational needs of
the constituency, develop the curriculum,
recommend delivery formats for the curriculum,
propose appropriate recognition/accreditation,
and establish appropriate continuing education
requirements. The Vice President for Research
and Dean of the Graduate School is also charged
with the responsibility of maintaining a database
on meeting the educational requirements.
Development and Delivery of the
Curriculum
The development and delivery of the educational
program in RCR for investigators has been led by
the Faculty Education Advisory (FEA)
Committee. The FEA Committee is in its third
year of existence and is made up of faculty, with
senior administrators serving in ex officio
capacity. The Committee is staffed by personnel
from the Office of the Vice President for
Research. The Committee meets monthly and
has had remarkably consistent participation over
the three years. Members were added recently to
increase representation of disciplines within the
University.
Members of the FEA Committee are senior
and respected faculty and broadly represent the
diversity of the University’s colleges,
departments, and programs. The commitment of
faculty leaders, coupled with resources and
commitment from high-level University
administration, has been crucial to the success of
the FEA Committee’s effort. The Committee
has focused on three areas in RCR education:
(1) defining and identifying the target
populations; (2) identifying topic areas; and
(3) implementation.
Defining and identifying target populations
for RCR education and training
The initial focus of RCR educational
programming has been PIs, both because it
represents the largest group of faculty and staff
responsible for the performance of research, and
because the University has a system for
certification of PI status. This cohort represented
nearly 2,000 individuals, from across every
college and a diverse range of departments and
research areas.
This diversity led to a recognition that
education in RCR could not be successful as a
“one size fits all” program, and that we needed to
speak to the needs and interests of researchers
from outside biomedical research areas. But in
spite of the diversity of researchers’ needs, the
FEA Committee agreed on a need to achieve a
shared basic level of understanding for all
researchers on a core set of RCR issues. This is
based on the view that all researchers belong to
the University’s research community, and that
204
––––––––––––––––––––––––––––––––––––––––––– Kahn, et al., Fostering Research Integrity through Educational Programs
such membership brings with it certain
responsibilities, including basic familiarity with
the rule and issues in areas such as research that
involves human or animal subjects. So while
many researchers may never engage in human or
animal research, it is unacceptable for them to
pass it off as someone else’s problem. For those
researchers engaged in research involving human
or animal subjects, more in-depth education and
training in those areas is required. In addition to
both basic training for all and in-depth training
when appropriate, the FEA Committee is
developing recommendations for continuing
education in RCR.
Identifying topic areas
The FEA Committee’s second task was to
identify topic areas for curriculum development.
Since our efforts pre-dated the PHS/Office of
Research Intetrity (ORI) draft of final guidelines,
an initial list of topics was drawn from the list of
suggested topic areas in the ethics requirement
for NIH Training Grants (T32). The FEA
Committee then worked to make the list of topics
relevant to PIs. The current list of topics
includes:
Social Responsibility and Misconduct
Authorship and Peer Review
Data Management
Intellectual property
Conflict of Interest
Fiscal Responsibility
Human Subjects
Animal Subjects
Environmental health and Safety
After the PHS/ORI guidelines were issued,
we compared our list of topics to the guidelines
in an effort to assess what changes, if any, are
needed, and determined that we need to add
content on both collaborative science and
mentoring.
Implementation
After identifying the target population, and the
topic areas that would be covered, the FEA
Committee’s last task was to develop strategies
for implementation. Key components in our
effort include recruiting instructors with
appropriate expertise and experience, drawing
mostly from the ranks of the faculty; and a
commitment that face-to-face interaction be part
of the educational experience.
We have employed three separate formats for
instruction—classroom sessions totaling six
hours; web-based instruction for some financial
and grants management topics, followed by a 1.5
hour classroom session; and in-depth special
topic instruction involving a 1.5 hour classroom
session, web resources, and case studies.
Because of the number of hours of
instruction required and the diversity of
investigators who need to participate, a large and
diverse pool of instructors was recruited. We
have between four and six faculty who are
prepared to deliver one topic area; faculty are
paired with relevant professional staff for some
topics. These 37 instructors represent 13
colleges and 3 administrative units, and include 4
department heads, and 2 associate deans. While
all of the faculty agreed to teach in our RCR
efforts on a volunteer basis, the FEA
recommended and the University’s Vice
President for Research agreed that formal and
material acknowledgement of their efforts is
appropriate. To that end, funds were committed
to provide small professional development
awards to all faculty participating as instructors
in the RCR programs.
Resources & Financial Investment
A cornerstone of our program is faculty
involvement in the delivery of the curriculum.
Faculty are presenters or facilitators of discussion
for each topic. For some topics they are partnered
with staff who are available to answer more
technical questions. For example, faculty who
deliver the module on Intellectual Property are
paired with a staff member from the University
office of Patents and Technology Marketing.
Faculty are also involved in revising instructional
materials used in workshops and on the web, as
well as the curriculum itself.
The commitment of respected, senior faculty,
demonstrated by their leadership on committees
or their development of the curriculum, enabled
us to recruit other faculty for the delivery of the
curriculum. Another critical element for
recruitment was a detailed syllabus for each topic
of the curriculum. The syllabus includes learning
objectives, relevant policies, principles, issues for
discussion, reference materials, and case studies
for some topics.
One limitation of the curriculum was its biomedical flavor, particularly in case studies,
largely because of the disciplines represented on
the initial faculty advisory committee.
Recognizing this, we targeted faculty in
205
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
underrepresented disciplines to achieve greater
balance for delivery of the curriculum. Over 50
faculty from 34 departments are currently
involved in RCR curriculum development or
delivery. Besides enriching the curriculum, we
believed that faculty involvement throughout the
University would increase ownership and spread
commitment to the RCR. An unexpected
outcome of the diversity of disciplines has been
the high level of interest maintained by the
faculty as they see the issues in their topic take
on new dimensions and challenges from one
discipline to another.
Besides the demonstrated commitment of
faculty, a successful educational program in RCR
requires strong support services. Instructional
materials are revised and shared amongst
presenters. When the faculty audience asks
previously unanswered questions, the experts are
consulted. The answers are incorporated into
future workshops, and the curriculum and
instructional materials are revised as appropriate.
There are also numerous administrative tasks
associated with scheduling presenters, rooms,
and equipment; preparation of printed and web
based materials; registration and documentation
of attendance; tabulation of evaluations; and
feedback to and coaching of faculty presenters.
Although these activities happen mostly behind
the scenes, they are critical to the program.
Finally, communication is a critical support
service. Requirements and rationale must be
conveyed to the faculty and other research
personnel; progress must be reported to the
advisory committee (FEA), faculty senate
committees, administrative offices, and academic
administrators. All available communications
vehicles are used, including monthly newsletters
of the sponsored projects office and of colleges
as well as the University’s multiple publications;
printed brochures and flyers; web based home
pages and events calendars; meeting in person
with faculty committees, academic deans, and
special constituencies (IRB); and e-mailings from
the Vice President of Research, Deans, and
Department heads or chairs.
So what does all of this cost? The direct
expenses of the 62 workshops for 2,400
investigators over a 12-month period is the most
straight forward. Based on actual cost to date for
printing of materials, rental of rooms and
equipment, and similar expenses, these direct
expenses are projected to be $48,600, or $15.20
per person per session. This amount does not
include any compensation for the faculty
involved in the delivery. Based on the average
actual salaries of faculty involved in the
workshops, with an average of 1 – 2 hours
depending upon the topic, the value for delivery
would be an additional $32,300. This does not
include any estimate of faculty time for
preparation or involvement in discussions, via email or in person, of improvements or additions
to the materials, sharing of additional references,
or similar and recurring work. Although faculty
were recruited without any hint of monetary
reward, we were able to give those most involved
small professional development grants of
$1,000 – 2,000, for an expense of $24,000.
Direct administrative costs include the salary
and fringe benefits of 1.75 staff years: one full
time program coordinator, a 50% administrative
appointment of a faculty member acting as
program director; and an administrative fellow
(graduate student). However, the direct cost of
additional support services including design and
maintenance of web-based tutorials as well as
registration and recording keeping activities are
nearly impossible to tally since they are provided
by a number of centralized offices from the
graduate school administration to the human
resources offices.
Hardest yet to calculate are the cost of
faculty hours spent in participation. Since the
University of Minnesota has no formula for
faculty productivity or opportunity costs, one
simple estimate was based on salary. Applying
the composite faculty salaries for Category 1
universities in our region from the March 4,
2000, issue of Academe and the University of
Minnesota fringe benefits rate against the
estimate of 9,600 hours spent by faculty in
workshops or reading materials, we estimate the
cost of faculty participation at $425,000.
However, the benefit side of this equation is even
harder to estimate. Certainly the potential
liabilities exceed the total cost of the program,
including loss of faculty time.
Evaluation
Assessing for continuous course improvement
The RCR curriculum is currently offered in 2
parts of 3 hours each. At the end of each session,
participants are asked to complete a one-page
course evaluation form which asks 1) whether the
level of content for each topic is appropriate,
2) whether the information on each topic is
206
––––––––––––––––––––––––––––––––––––––––––– Kahn, et al., Fostering Research Integrity through Educational Programs
useful, 3) whether the session increased your
understanding, and 4) whether the materials and
resources were helpful. Finally, there is a place
for comments. Data from these forms have been
summarized and used to make course
improvements.
During the first 6 month period, 66% of the
participants (N=522) returned the evaluation for
part 1; 43% (N=1162) for part 2. In general, 80%
of the participants judged the material presented
to be appropriate. Lists of resources and
websites were considered the most useful
resources. Early on, criticisms outpaced
satisfactory remarks 3 to 1. Constructive
comments included: make the course more
interactive, provide readings ahead of time,
incorporate web based materials, and shorten the
length of time. Subsequent iterations of the
course adopted these suggestions. As a result,
the overall rating of usefulness improved, from
2.7 to 3.0 on a 4 point scale (with 4 being very
useful) for part 1 and from 2.5 to 2.9 for part 2.
In addition, there were fewer critical comments,
and the number of statements of praise increased.
Reflecting on the course evaluation data and
our efforts at course improvements, we have
identified the following contributors to
participant satisfaction:
Interactive programming. The more interactive
the program, the more it is viewed as useful.
Group size. Smaller groups are better received
than larger groups.
Presenters from diverse disciplines. Participants have been less satisfied when the
presenters are all from the same discipline.
Topic. Some topics seem to be inherently more
interesting than others. For example,
authorship seems to be rated as most interesting irrespective of who presents the
material. Other topics, like intellectual
property and conflict of interest typically get
lower ratings for usefulness. However, when
we have broadened the topic of intellectual
property to include more on copyright, there
were some improvements in rating. Staff
have speculated that in areas like intellectual
property and conflict of interest may be
inherently dissatisfying as it is seldom
possible for the presenter to give definitive
answers to questions.
Assessing promotion of responsible conduct
Documenting faculty participation in an initial
and on-going educational program in RCR
demonstrates compliance with a federally
mandated corrective action plan (e.g., the NIH
plan currently in effect for the University of
Minnesota). It does not, however, provide
evidence that the attitudes, values, and behaviors
that gave rise to the disciplinary action have
changed. Likewise, installing a model system for
financial accountability, such as the Electronic
Grants Management System (EGMS), can alert
an individual faculty member and his/her unit
head when a proposed action is not within the
bounds of sanctioned behavior. It does not,
however, assure that the moral climate in which
research is conducted is enhanced, or will it
necessarily improve the ability of investigators to
interpret ambiguous situations and identify better
choices. If we hope to provide evidence that we
have improved the integrity of the researcher and
climate of the institution, we need measures that
assess the more elusive outcomes of the research
ethics enterprise and that can be used to examine
the effectiveness of our educational programs
and compliance systems.
In Fall of 1999, a faculty group was
convened to identify opportunities for assessment
of outcomes. The following were identified:
Self-assessment questions in web-based
modules. Self assessment items have been
included in several topics: Fiscal Responsibility,
Intellectual Property, Conflict of Interest,
Informed Consent, Protecting Human Subjects.
Although self assessment items are included, we
have decided not to invest resources to assess
knowledge level outcomes.
University-wide climate surveys to track
perceptions of ethical research practices. The
last Faculty and Staff Climate Survey of the
University of Minnesota was conducted in 1997,
with a summary reported in 1999. Questions are
being prepared for the next survey. The purpose
will be to track perceptions of the extent to which
the University climate supports ethical conduct
generally. Questions would be directed toward
ethical research practices as well as issues of
academic integrity.
Narrative interviews of unit administrators.
In addition to eliciting their perceptions of the
norms of research conduct, interviews with unit
administrators is a way of identifying areas
needing attention.
Graduate student perceptions of the doctoral
experience. Melissa Anderson directs the
207
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Academic Life Project, funded by NSF, which
studies the normative experiences of doctoral
students (see paper by M. Anderson in these
proceedings for additional information on this
study).
Adaptation of measures of ethical reasoning
and role concept. One reason for the paucity of
information on assessment of instructional effects
in this area is the lack of well-validated outcome
measures. Measures must be grounded in a wellestablished theory of ethical development and be
sufficiently user friendly to enable their use for a
variety of purposes. We propose to develop two
outcome measures: (1) a measure of ethical
reasoning and judgment about common problems
arising in the research setting, and (2) a measure
of role concept, i.e., how the researcher
understands his/her role relative to other
researchers. The measures will assess two of the
four dimensions of competence described by
Rest’s Four Component Model of Morality (Rest,
1983). The areas are chosen because prior studies
support the usefulness of the methods for
outcome assessment and for demonstrating the
links between performance and day-to-day
ethical behavior. The two measures will be
modeled after existing measures designed for
assessing the outcomes of ethics education in
dentistry. (See paper by M. Bebeau in these
proceedings for additional information on these
approaches).
In summary, a national effort is required to
design outcome measures that can be used to
assess the effectiveness of institutional education
programs in RCR. Measures must wellgrounded theoretically, well validated, and
sufficiently user friendly to enable their use for a
variety of purposes. Such purposes may include:
1) determining the range of criteria that define
competence in topic areas among different
disciplines, 2) conducting a needs assessment to
identify areas where instructional resources
should placed, 3) identifying individual
differences or problems that require intervention
or remediation, 4) providing feedback to
individuals, departments, and institutions on
research ethics competence, 5) determining the
impact of current programs, and 7) studying the
relationship between competence and ethical
behavior.
Bibliography
1. Rest, J. Moraliaty. In Mussen PH (ser. ed.), Flavell J,
Markman E (vol. eds.), Handbook of Child Psychology,
Cognitive Development, Vol. 3, (4th ed). New York:
Wiley; 1983; p. 556-629.
208
Being a Scientist: Educating for Ethical Conduct
Chloe Doby Little, Department of Physical Therapy, Western Carolina University, USA
Katherine Le Guin White, Department of Physical Therapy (retired), Western Carolina
University, USA
Key Words: Ethical conduct, Problem-based learning curriculum, Reflection-in-action skills, Reflection-inexperimentation, RCR
This project is predicated on a reflective way of life for being a scientist as the epistemological
foundation for educating health professions students in the ethical conduct essential for scientific
integrity and progress. Thus, being a scientist exemplifies a reflective way of life; and educating
health professions students for ethical conduct embodies the reflective practitioner epistemology
explicated by Schon in his books, The Reflective Practitioner and Educating the Reflective
Practitioner (1, 2). Schon (1) challenges traditional professional curricula and educators that
continue to implement course content based on the positivist, technical-rational epistemology of the
nineteenth and twentieth centuries. The reflection-in-action epistemology Schon (2) pioneered offers
health professions educators and practitioners a theoretical system of knowledge for helping faculty in
science-based professions education update curricula.
The thesis of this project is that a transitional problem-based learning (PBL) curriculum in the
allied health professions provides an excellent framework for education of reflective practitioners.
Reflective practitioners are problem solvers and ethical scientists. Faculties who are themselves
exemplary reflective researchers and teachers can teach ethics through successful PBL experiences
that guide health professions students in development of ethical conduct as the foundation for their
way of life as science-based, reflective practitioners.
A transitional PBL curriculum in the health professions is structured to guide students from
acquisition of new information and knowledge through application of that knowledge in solving
clinically-based problems to reflection-in-action as practitioners. Put another way, the transitional
PBL curriculum helps health professions students progress from information gathering and
knowledge warehousing to practitioners who know through reflection-in-action and are therefore wise
clinicians rather than master technicians.
Faculties, who are science-based, reflective practitioners and instructors, integrate scientific
research, scholarship, and teaching. Successful implementation of reflection-in-action epistemology
in health professions curricula depends in large measure on the participation of wise, dedicated
faculty whose ethical conduct as scholars and as teachers is manifested in their successful
participation in those reflective dimensions of problem-based learning experiences.
Introduction
Keith-Spiegel, et al., (3) report that scientific misconduct is socialized during undergraduate years
Corresponding author: Chloe D. Little, PW Box 2145, Cullowhee, NC 28723, 828-227-2288 (voice), 828-227-7071 (fax),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
with students believing that significant results
will earn them better grades. Recent research by
Davidson, et al., (4) lends additional support to
these findings. One also can speculate that
scientific misconduct reflects the attitudes of
society. Dishonesty and misrepresentations have
become commonplace and acceptable in the
absence of social sanctions against these
behaviors and also as a result of increased
professional competition and increased pressure
to produce. Since the 1940’s the incidence of
self-reported academic misconduct by college
students has risen 55-78 percent. (5) Other
examples of misconduct include medical school
faculty applicants misrepresenting research
citations, (6) ethics committees endorsing
unnecessary research, (7) peer-reviewed journals
editors misappropriating authorship, (8) and
researchers faking data in experiments or failing
to report unfavorable results (9). Some
researchers suggest that there has been a
“reorientation away from traditional values,”
especially in scientific inquiry (10) . Others
speculate that fraud and dishonesty in scientific
research are the inception rather than the rule
(11).
Regardless, scientists and institutions must
maintain quality and integrity in scientific
research if progress and public support are to be
sustained. To promote responsible research,
college and university faculties must sensitize
future scientists to the critical issues in research
ethics and guidelines. Also, the National
Institutes of Health requirements mandate all
institutions participating in training grants show
they provide instructions to faculty and students
in the principles of scientific integrity (12).
Additionally, the final report of the Commission
on Research Integrity noted the importance of
providing “formal and informal educational
opportunities to sensitize both junior and senior
scientists to critical issues in research ethics and
their institution’s guidelines” (13, p.16).
Although expecting college and university
faculties to single-handedly prevent research
misconduct is unrealistic, faculties can create
informal learning environments to promote high
standards by engaging students in open
discussions of ethical and unethical research
practices, carefully supervising and mentoring
student research, encouraging responsible data
management, and modeling ethical behaviors.
Faculties also can create formal methods for
integrating the study of scientific values and
responsible conduct in the academic courses.
This project presents informal and formal
methodologies to encourage health professions
graduate students to develop reflection-in-action
skills and values that foster ethical practice in
health professions services and clinical research.
The ultimate goal is to describe a curriculum for
promoting active student learning throughout a
series of scientific research courses.
Implementing Problem-Based Learning
Curriculum in Scientific Research for
Graduate Health Professions Students
First semester course content includes three casebased problems for students to study and discuss:
university-specific guidelines for conduct of
scientific research, how to construct a research
project, and the virtues of ethical research.
Second semester course content is focused on
student implementation of the research project
constructed during the first semester. In
subsequent semesters, students reflectively
examine with faculty mentors their completed
student projects for ethical integrity.
Learning issues in the first case-based
problem explored in semester one focused on
defining scientific misconduct through
differentiating negligence from deliberate
dishonesty and examining institutional research
policies, especially distinguishing human and
non-human research, confidentiality, and the
obligations of scientific researchers. Students
complete an institutional review board proposal
for their subsequent projects. The second
problem progresses students to application of
those skills and behaviors learned in the first
case-based problem on the rudiments of
responsible scientific conduct. Learning issues
for this case include practicing ethical data
management and examining the ethical content
of published research studies. The third problem
is structured to concentrate student learning on
management of conflicting interests,
determination of criteria for multiple authorship,
reporting scientific misconduct, and the process
by which research grants are awarded.
Second semester learning issues arise from
reflection on students’ performances as they
begin to conduct their research projects,
structured during the first semester. Throughout
this course faculty and student reflection-inaction and faculty mentoring become critically
important. Learning experiences during this
semester are more informal than those structured
210
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Little & White, Being a Scientist
for the first course. Students complete their
projects in subsequent semesters, varying from
one to three. Equally critical throughout these
times is informal student-faculty discussions,
supervision, and reflection that occurs during
regularly scheduled small group or individual
meetings.
Benefits of Problem-Based Curriculum
and Learning Experiencesfor Faculty and
Students
Students and faculty alike are beneficiaries of
PBL experiences and curricula. Students develop
problem-solving skills through student-directed
discussions and information gathering
assignments. They also learn to become selfdirected and independent learners, habits that
equip them for lifelong learning in practice
communities, even in those remote settings
where colleagues and library resources may be
scarce. As they become more independent
learners, students begin to actively demonstrate
increasingly critical, creative thinking.
Assessment of one’s peers during PBL
experience is an essential dimension of PBL that
requires active participation of all students in a
learning group. To that end, students must learn
to assess themselves and their colleagues in
honest, thorough, deep, and sincere ways.
Learning to work critically in this manner helps
students reach greater depths of understanding
the importance of frequently and realistically
evaluating their performance as team members
and learners; they also become skilled in
applying the same sensitivities to evaluating the
participation and performance of their peers in
learning groups. These assessment skills and
values also relate to other aspects of PBL:
information management, creation of measurable
knowledge bases for solving problems, and
assessing peers, social and ethical skills,
communication effectiveness, and the ability to
work effectively as a team member.
Finally, development of leadership skills is
fostered through revolving, shared group
leadership. For each problem-solving session,
students select a group leader, facilitator, and
recorder. All group members serve in each
capacity throughout a semester.
If PBL is to be successful, faculties must
become models and coaches, relinquishing their
traditional roles as lecturers and purveyors of
information. In this role, faculties develop skills
that monitor student learning during a problem-
solving session and throughout the curriculum.
To properly monitor student learning, faculties
must become proficient in classroom reflective
behaviors that probe and challenge students’
thinking conclusions and processes, keep
students involved throughout exploration of the
problem, adjust levels of challenge to students,
and manage group dynamics so that processes
move toward constructive resolution of the
problem. Development of learning materials and
writing comprehensive clinical problems that
challenge students demand faculty creativity and
planning that exceed those faculty demands
imposed by a curriculum predicated on
traditional technical-rational epistemology.
Faculties relinquish the resident expert status to
become guides for student learning that is
independent and self-directed. Faculty expertise
in asking rather than telling, planning and
guiding rather than showing is essential for
successful discussions and problem solving
sessions.
Formal and Informal Methodology
Designs
Problem-based learning methodologies presented
here are designed to encourage first-semester
health professions graduate students to develop
reflection-in-action skills and values for ethical
practice as clinicians and as researchers. The
ultimate goal of the methodology is to promote
active student learning in the education of future
scientists who will consistently demonstrate
ethical scientific research behaviors.
As with the previously discussed benefits of
PBL for students and faculty alike, effective PBL
methodology design occurs only when faculties
and students participate successfully in the
process. At a minimum, faculties must openly
discuss with students during learning group
sessions those ethical and unethical behaviors in
scientific research reported in the literature and in
the faculty member’s experience as a scholarresearcher. Faculties also must carefully and
continuously supervise student research activities
while mentoring student development as novice
researchers. To be credible leaders for
development of ethical behaviors in students,
faculties must be personally engaged in ongoing
and successful scientific research and
scholarship.
Student involvement in design of PBL
methodology requires full participation of all
group members in researching the literature
211
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
available on ethical and unethical practices in
scientific research. Students also must learn to
engage faculty and student peers in reflective
discussions throughout the problem solving
group experiences.
Finally, students must demonstrate learned
ethical behaviors in their own student research
projects completed after their first semester.
Formal faculty and student responsibilities
for methodology design and successful
implementation are focused on scientifically
rigorous planning and participation guidelines.
Faculties are charged with responsibility for
developing curriculum materials that include a
series of complex, real world, “ill-structured”
problems to stimulate learning, integration and
organization of learned information that ensure
application of past learning to future problems.
Curricular materials include learning objectives
for each PBL problem, definition of PBL
techniques, appointment of small groups of 5-7
student learners, identification and instruction of
tutors, guidelines for student leadership process
and responsibilities during group learning
sessions, and development of assessment tools.
Beyond these process design matters, the
essential faculty responsibility is creating
multiple cases that form the bases for student
learning. Without solid, reality-based clinical
cases, the process cannot proceed as a valid or
effective learning experience. As stated earlier,
faculty also must model the values promoted as
ethical conduct for scientists. They must
consistently demonstrate their ability to reflectin-action as they participate in the group learning
experiences.
Students likewise have many formal
responsibilities for achieving successful PBL.
Students must learn to formulate hypotheses as
individuals and as learning team members. They
must learn to participate effectively and
responsibly as group members for many
outcomes, including designing a plan to solve the
problem, researching available and pertinent
information, justifying individual and group
decisions and conclusions, recognizing multiple
acceptable solutions to a given problem,
evaluating the performance of themselves, their
peers, and their tutors, and demonstrating novice
reflection-in-action skills and values.
Discussion and Conclusion
encourages independent learning during which
students develop depth of understanding of
content (14). Through PBL students become
more involved in and responsible for their own
learning. The objectives of PBL are to assist the
process of active learning by students as they
develop effective clinical reasoning skills, such
as critical appraisal, decision making,
collaboration, and self-directed learning habits in
order to participate effectively and actively in the
small group discussions during the problem
solving of cases. (15, 16) Each problem should
be designed to provoke critical inquiry, to
encourage independent access to multiple and
diverse learning resources, and to generate lively,
focused, and pertinent small group discussions.
Reflection-in-action during and after completion
of a problem promotes transfer of learning as
well as generation of new concepts (16). Recent
research findings suggest PBL curricula are
effective methods of learning and that students
successfully transfer knowledge and skills in
timely and meaningful ways (17, 18, 19).
Researchers have shown PBL promotes
higher order thinking skills (16). PBL is a
curriculum approach that places students in the
active role of problem solver during the process
of constructing meaning from case-based
problems that mirror real-world situations.
Throughout the process students develop
problem-solving and information gathering
strategies, reflection skills, and disciplinespecific knowledge bases. In the absence of
actual clinical experiences during problem
solving discussions, students learn to make
judgments based on facts, information, logic, and
rationalization alone, they must use higher
thinking orders to justify decisions based on
application of learned principles. Nevertheless,
the defining measurement of learning during an
academic course is the quality of research
produced by the student, an outcome that may
not be evident throughout the span of the course.
Therefore, continued supervision and mentoring
of a student’s future research activities beyond
the first semester is essential for facilitating
ethical development. The authors believe that
through PBL students will exhibit reflection-inexperiment skills that will culminate ultimately
in reflection-in-action skills1 as they complete
their student research projects and move toward
mastery as scientific researchers.
Problem-based learning, based on small group
discussion and clinically-based problems,
212
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Little & White, Being a Scientist
Bibliography
1. Schön DA. The reflective practitioner: How
professionals think in action. NY, NY: Basic Book, A
Division of Harper Collins Publishing; 1984.
2. Schön DA. Educating the reflective practitioner:
Toward a new design for teaching and learning in the
professions. San Francisco: Jossey-Bass Publishers;
1987.
3. Keith-Spiegel P, Lee B, Spiegel D, Zinn-Monroe M.
Research described in Scientific Misconduct: An
annotated bibliography by Keith-Spiegel, P., Aronson,
K., & Bowman, M. 1994. Available from: URL: http://
www.lemyne.edu/OTRP.
4. Davidson EW, Cate HE, Lewis CM, Hunter M. Data
manipulation in the undergraduate laboratory: What are
we teaching? First ORI Research Conference of
Research Integrity; 2000 November 19-21; Bethesda,
MD. 2000.
5. Educational Testing Service. Cheating is a personal foul:
Academic cheating background. 1999. Available from:
URL: http://www.nocheating.org/adcouncil/research/
cheatingbackgrounder.
6. Goe LC, Herrera AM, Mower WR. Misrepresentation
of research citations among medical school faculty
applicants. Acad Med, 1998; 73: 1183-1186.
7. Savulescu J, Chalmers I, Blunt J. Are research ethics
committees behaving unethically? Some suggestions for
improving performance and accountability. BMJ, 1996;
313(7069): 1390-1393.
8. Flanagin A, Carey LA, Fontanarosa PB, Phillips SG,
Pace BP, Lundberg GD, Rennie D. Prevalence of
articles with honorary authors and ghost authors in peerreviewed medical journals. JAMA, 1998; 280(3): 222224.
9. Friedberg M, Saffran B, Stinson T, Nelson W, Bennett
C. Evaluation of conflict of interest analysis of new
drugs used in oncology. JAMA, 1999; 282(15): 14531457.
10. Bechtel HK, Pearson W. Deviant scientists and
scientific deviance. Deviant Behav, 1985; 6: 237-252.
11. Knight JA. Exploring the compromise of ethical
principles in science. Perspectives Biol Med, 1984; 27:
432-441.
12. National Institute of Health. Requirements for programs
on the responsible conduct of research in National
Research Service Award Institutional Training
Programs. NIH Guide for Grants and Contracts, 1989;
19: 1-35.
13. U. S. Department of Health and Human Services,
Commission on Research Integrity. Integrity and
misconduct in research. Washington, DC: Office of
Research Integrity; 1995.
14. Schmidt HG. Problem-based learning: Rationale and
description. Med Educ, 1983; 17: 11-16.
15. Dolmans DH, Schmidt H. What drives the student in
problem-based learning? Med Educ, 1994; 28: 372-380.
16. Barrows HS. (1998). The essentials of problem-based
learning. J Dent Educ, 1998; 62(9): 630-633.
17. Borduas F, Carrier R, Drouin D, Deslauriers D, Trenblay
G. An interactive workshop: An effective means of
integrating the Canadian Cardiovascular Society clinical
practice guidelines on congestive heart failure into
Canadian family physicians’ practice. Can J Cardiol,
1998; 14: 911-916.
18. Dolmans DH, Schmidt H. The advantages of problembased curricula. Postgrad Med J, 1996; 72: 535-538.
19. Dolmans DH, Gijselaers WH, Schmidt HG, van der
Meers SB. Problem effectiveness in a course using
problem-based learning. Acad Med, 1993; 68: 207-213.
213
Encouraging Accountability in Research: A Pilot Assessment of Training
Efforts*
Anna C. Mastroianni, School of Law and Institute for Public Health Genetics, University of
Washington, USA
Jeffrey P. Kahn, Center for Bioethics, University of Minnesota, USA
Keywords: Pilot assessment, RCR education and training, Training grants
The objective of this pilot assessment was to describe the response of a sample of grantee institutions
to the federally-mandated training requirement in the responsible conduct of research that is part of
NIH Training Grant (T32) funding. Materials collected by the Department of Health and Human
Services (DHHS) were reviewed and described with the following five research goals:
• describe the target audience for training programs
• describe the locus of instructional responsibility for training programs
• describe whether all trainees at an institution participate in the same training program
• describe the program approaches, materials used and program contents
• create a source of baseline information for planning evaluations of future training programs
• identify areas for further research and analysis
Methods
The sample consisted of a collection of materials assembled by DHHS. These included syllabi, course
outlines, case studies, reading lists, institutional research policies, and other information provided by
training grant recipient institutions about their research ethics programs. In June 1996, the Office of
Science Policy, Office of the Assistant Secretary for Planning and Evaluation, DHHS, sought to create
“a library of course materials that are being used by T32 grantees.” A letter was sent to a stratified
sample of T32 grantees requesting “any training materials currently used to instruct trainees in
research integrity and misconduct” (1). The stated goal of collecting this library of information was
to provide an understanding of training programs in the responsible conduct of research, including the
range of institutional approaches for meeting the training grant requirement. This information was
not collected as part of assessing regulatory compliance or as part of any oversight effort, but to
create a resource and a source of baselines information for planning evaluations of future training
programs (2).1 This sample served as a convenient and best available sample for this review.
Excerpted from: Mastroianni A, Kahn JP. Encouraging accountability in research: A pilot assessment of training efforts.
Accountability in Research 1999;7:85-100. Some policy implications of the results presented here are also discussed in
Mastroianni, A.C. and Kahn, J.P. The importance of expanding current training in the responsibleconduct of research. Acad
Med 1998; 73(12):1249-1254.
Corresponding author: Anna Mastroianni, University of Washington School of Law and Institute for Public Health Genetics
1100 NE Campus Parkway, Seattle, WA 98105-6617, 206-616-3482 (voice), 206-616-3480 (fax),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DHHS contacted awardees at 50 of the 210
institutions that held training grants as of October
1995. (3) DHHS selected these 50 based on
number of training grants, geographical location,
status as public or private institution, and number
of T32 trainees at the institution. For those
institutions with multiple training grants,
individual grants were selected for inclusion in
the sample in order to obtain diverse
representation. Selection factors included: the
number of trainees, the distribution of pre- and
post-doctoral students, and representation of
clinical and basic research.
DHHS contacted Principal Investigators by
telephone and follow-up letter, and requested that
they provide “any training materials currently
used to instruct trainees in research integrity and
misconduct [including] materials such as the
syllabi, course outlines, case studies, reading
lists, institutional codes of conduct in research,
etc., [and] any information [that] readily . . .
describes the context in which such materials are
introduced to students and the method of
training” (4). Respondents from 45 of the 50
institutions contacted provided information
concerning a total of 75 training grants.
Access to and copying of these publicly
available materials was provided by the Office of
Science Policy, Office of the Assistant Secretary
for Planning and Evaluation, DHHS, in
November 1996.
Approach
A coding form was developed as a method to
collect and summarize information from the
sample. Descriptive statistics were calculated
using standard statistical software.
The characteristics of the sample were
described at the level of either the institution
(n=45), or the responsible conduct of research
training program (n=75). In order to understand
whether institutions shared characteristics based
on number of training grants, the sample of
institutions was stratified into thirds by number
of training grants. For this purpose, these
groupings were categorized as: “low-density”
institutions (14/45 [31.1%] of the institutions)
which held four or fewer training grants;
“medium-density” institutions (15/45 [33.3%] of
the institutions) which held from five through
nine training grants; and, “high-density”
institutions (16/45 [35.6%] of the institutions)
which held ten or more training grants.
Institutions also could have been grouped by
total number of trainees. In examining total
number of trainees and number of T32s against
other variables, each was found to be a proxy for
the other. Variables, where appropriate, are
grouped by numbers of T32s only.
Results
There were 45 institutions in the sample
representing 660 T32s (number of T32s at each
institution ranges from 1 to 60, with a median of
6) and 4,883 trainees (number of T32 trainees at
each institution ranges from 3 to 507, with a
median of 38). Responses concerning 75 training
grants were represented in the sample.
Of the 45 institutions, 25 [55.6%] were
public educational institutions, 17 [37.8%] were
private educational institutions, and 3 [6.7%]
were non-academic institutions (i.e., a
professional organization, a non-profit service
provider, and an independent research
organization).
Institutional Characteristics
The sample was reviewed to determine the target
audience for the training programs. Two-thirds of
institutions represented in the sample required
that only T32 trainees receive training in the
responsible conduct of research. In this sample,
this result was not affected by the number of
training grants held by the institution: 9/14
[64.3%] of low-density, 10/15 [66.7%] of
medium-density, and 11/16 [68.8%] of highdensity institutions required training only for T32
trainees. Over one-quarter of all of the
institutions, however, required much broader
participation of either all trainees in the school or
college, all graduate students or all trainees in the
institution.
In half (23/45 [51.2%]) of the institutions
represented in the sample, the responsibility for
the responsible conduct of research training
program was located at the departmental or
Principal Investigator level. Another quarter
located the responsibility at the institutional
level. In the materials submitted, 4 [8.9%] of the
institutions placed responsibility for the program
in their ethics faculty. The institutions that
placed responsibility for the program in their
ethics faculty were among the highest-density
institutions in the sample. They each had 18 or
more training grants, and represented the top
quarter of the sample by number of training
grants. The majority of low-density and
medium-density institutions had the locus of
216
–––––––––––––––––––––––––––––––––––––––––––––––––––– Mastroianni & Kahn, Encouraging Accountability in Research
program responsibility at the department level
[64% and 66%, respectively], while the majority
of high-density institutions had the locus of
program responsibility above the department
level [75%].
For those 41 institutions with more than one
NIH training grant, 24 [58.5%] used the same
responsible conduct of research program for all
those required to receive training in the
responsible conduct of research. As the number
of training grants at an institution increased, the
proportion of institutions utilizing the same
responsible conduct of research training program
decreased. Seven of the 10 [70%] low-density, 9
of the 15 [60%] medium-density, and 8 of the 16
[50%] high-density institutions used the same
program for all trainees.
Program Characteristics
The material from the 45 institutions in the
sample included information from 75 training
grants. Depending on the characteristic being
examined, the following analyses were based on
either the number of institutions (n=45) or the
number of programs (n=75). The denominator is
noted in each case.
two institutions were among the highest-density
institutions.
Lecture was the most popular method of
instruction represented in the sample (53/75
[70.7%]). (Table I) To examine whether
programs relied solely on lectures to satisfy the
requirement, the frequency of lecture format in
combination with other methods of instruction
was determined. (Table II) For those programs
that used lectures as a method of instruction, only
a small proportion (4/53 [7.5%]) did not
supplement lectures with some less didactic
method or methods of instruction that provide
opportunities for greater interaction. It is
interesting to note that the materials indicated
that there was very little use of “brown bag”
discussions to satisfy the requirement.
Contact hours could be determined for 42 of
the 75 [56%] programs for which information
was received. The median number of contact
hours for these programs was 10 hours. The
range was from 4 to 72 contact hours.
Method of Instruction*
Lecture
Case study
Small group
Seminar
Student presentation
Mentor
Brown bag
Computer
Program approach
Submitted materials indicated that one-quarter
of the programs specifically tailored training to
the trainee population, with either disciplinespecific focus or both general and disciplinespecific material.
Of the 45 institutions, 28 [62.2%] had a
formal course in place to satisfy the training
grant requirement. A greater proportion of
medium-density and high-density institutions
utilized a formal course than did low-density
institutions: 5 of the 14 [35.7%] low-density
institutions, 13 of the 15 [86.6%] mediumdensity institutions, and 10 of the 16 [62.5%]
high-density institutions had a formal training
course in place.
Fourteen [31.1%] of the institutions
represented in the sample had programs that
indicated the availability of ethics training that
could be taken to supplement the course or
training offered to satisfy the training grant
requirement.
Only two institutions indicated that formal
training was provided to faculty who then carried
out the required responsible conduct of research
training—a “train the trainer” approach. These
# [%]
53 [70.7]
42 [56.0]
36 [48.0]
21 [28.0]
11 [14.7]
9 [12.0]
1 [1.3]
0 [0]
Table 1. Method of program instruction. n=75
* programs could have more than one method of
instruction
Methods of Instruction
Lecture only
Lecture + seminar
Lecture + small group
Lecture + case studies
Lecture + small group + case
studies
Lecture + seminar + small
group
Lecture + seminar + small
group + case studies
Lecture + brown bag + small
group
# [%]
4 [7.5]
3 [5.7]
11 [20.8]
16 [30.2]
14 [26.4]
3 [5.7]
1 [1.9]
1 [1.9]
Table 2.Combination of methods of program
instruction with lectures. Fifty-three programs used
lecture as part of their instructional format. n= 53
217
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Program Contents
Rank
Content Area
# [%]
Material from the 75 training
1
Authorship
65 [86.7]
grants was reviewed to
2
Data Management
56 [74.7]
determine whether course
3
Human Subjects
53 [70.7]
content included the five topic
4
Animal Use
51 [68.0]
areas recommended by NIH in
5
Conflict of Interest
49 [65.3]
the NRSA policy—conflict of
6
Institutional Policy
45 [60.0]
interest, responsible authorship
7
Skills-Writing
44 [58.7]
(including issues of peer-review,
7
Confidentiality
44 [58.7]
plagiarism, and research
9
Skills-Publishing
38 [50.7]
reporting), policies for handling
10
Intellectual Property
37 [49.3]
misconduct (including
11
Mentor/Mentee
35 [46.7]
institutional policies, federal
12
Information Sharing
24 [32.0]
policies, whistleblowing and
13
Whistleblowing and Reporting Misconduct
22 [29.3]
reporting misconduct), policies
14
Moral Reasoning
21 [28.0]
regarding the use of human and
15
Other Content
20 [26.7]
animal subjects, and data
16
Federal Policies
16 [21.3]
management (including
16
Grants Management
16 [21.3]
fabrication, falsification,
18
Skills-Grant Preparation
15 [20.0]
handling research data,
19
Organizational Structure
14 [18.7]
materials and information, and
20
Skills-Oral Presentation
11 [14.7]
data and objectivity).
21
Science and Society
10 [13.3]
Fifty-one [68%] of the T32
22
Laboratory Safety
9 [12.0]
programs covered four or five
23
Skills-Teaching
6 [8.0]
of the NIH recommended
24
Skills-Tenure
4 [5.3]
program content areas while 24
24
Skills-Funding
4 [5.3]
[32%] of the T32 programs
26
Skills-Jobs
3 [4.0]
covered three or fewer of the
categories. The top five ranked
Table 3. Ranking of program content areas. N = 75; programs can have more
categories fell within the five
than one content category.
NIH recommended program
Thirty-six of the 75 [48%] programs
content areas, and the top ten ranked categories
provided syllabi or other similar program
were addressed by at least half of the T32
materials in the information sent in response to
programs. (Table 3)
the DHHS request. Of those, 6 [16.7%]
Content issues that were identified by fewer
identified goals and objectives for the responsible
than half the programs include:
conduct of research training program. Based on
• whistleblowing and reporting misconduct (22
this limited information, few programs set forth
of 75 programs)
traditional goals and objectives for their
• the more theoretical issues encompassed in a
educational efforts.
category we labeled “moral reasoning” (21
of 75 programs)
Training Materials
• social issues encompassed in a category we
The information submitted was reviewed to
labeled “science and society” (10 of 75
identify the most frequently noted training
programs)
materials used by programs. The top
• development of certain skills necessary for
three referenced training materials were:
becoming a productive scientist, e.g. grants
1) institutional policies concerning the
preparation and funding, job hunting, oral
responsible conduct of research (45/75 [60%]);
communication, tenure, teaching, etc., (3 to
2) Korenman et al., Teaching the Responsible
Conduct of Research through a Case Study
15 programs).
Approach: A Handbook for Instructors (5) (30/75
General skills related to publishing and writing
[40%]); and, 3) the National Academy of
received greater attention, with 38 and 44
Science’s On Being a Scientist (6) (24/75 [32%]).
programs addressing them, respectively.
While the institutional policies are specific to
218
–––––––––––––––––––––––––––––––––––––––––––––––––––– Mastroianni & Kahn, Encouraging Accountability in Research
each institution, Korenman et al. (5) and NAS (6)
are prepared and widely distributed by
professional societies. Of the materials
referenced in the sample, the following four each
are marketed as offering complete training
materials in the responsible conduct of research
without need to supplement: Korenman et al. (5);
Macrina, Scientific Integrity: An Introductory
Text with Cases (7); the American Association for
the Advancement of Science’s Integrity in
Scientific Research: Five Video Vignettes (8); and
Bulger et al., The Ethical Dimensions of the
Biological Sciences. (9) Forty-three [57.3%] of
the programs used one or more of these four
materials. A greater proportion of high-density
institutions (12/16 [75%]) used at least one of
these four “ready-to-use” training materials, than
did low- or medium-density institutions (7/14
[50%]; and 7/15 [46.6%] respectively).
Discussion
In this sample, training in the responsible
conduct of research in response to the NIH
requirement was most often directed at T32
trainees. While the NIH policy encourages
expanding training to others, it requires that only
T32 trainees receive such training. If this result
is representative of institutional commitment to
training in the responsible conduct of research,
future scientists’ exposure to responsible conduct
of research will largely depend on their source of
funding. The characteristics of the minority of
institutions that make a broader commitment to
responsible conduct of research education and
training for its trainees deserve further
exploration.
The T32 recipient institutions in the sample
employed a diversity of approaches to satisfying
the training grant requirement. Approaches
varied both among and within institutions.
Further, the number of T32s held at the
institution had some impact on how the training
grant requirement was met.
Locating program responsibility at the
departmental or Principal Investigator level, as
did about half of the institutions in the sample,
may offer ethics training that is more tailored to
the trainees’ disciplines. In the materials
reviewed, a quarter of the programs offered some
discipline-specific training. Further research is
necessary to determine whether a relationship
exists between discipline-specific training and
location of program responsibility within an
institution.
The finding that a greater proportion of highdensity institutions placed program responsibility
above the departmental level may indicate that as
institutional demand for responsible conduct of
research training programs increases, more
shared institutional resources are sought.
However, based on T32 density, those institutions
with the highest density had the smallest
proportion that utilized the same responsible
conduct of research training program for all
trainees. This finding may be attributable to more
diverse training programs for which different
approaches are used, even if some institutional
resources are shared. Perhaps the administrative
level at which the ethics training decision is
made affects the institutional approach. Future
research might focus on examining this question,
and the sharing of institutional resources
regardless of any differences in program
approach.
The small number of institutions that placed
responsibility for teaching in their ethics faculty
may be a reflection of the fact that institutions
with greater numbers of training grants are more
likely to have ethics faculty—it would be
interesting to compare the characteristics of
institutions that have ethics faculty and place
program responsibility in them.
Contrary to the expectations of the authors,
lecture format alone was rarely used; nearly twothirds of the programs employed lectures plus
additional instructional approaches. Also
contrary to popular belief among teachers of
responsible conduct of research, brown bag
discussions were rarely identified as an approach
used to satisfy the training grant requirement.
The wide range of contact hours offered by
programs underscores the great diversity in the
implementation of the requirement.
The majority of programs (51/75 [68%])
specifically addressed four or five of the NIHrecommended subject categories. Either the
recommendations in the NIH policy have
influenced program content or the subject
categories are well-chosen and represent the
commonly accepted basic issues in the
responsible conduct of research.
Some variation in the subject matter covered
by programs may result from differences in the
needs of trainees in basic versus clinical research.
However, four of the five NIH-recommended
categories are relevant to all scientific research,
i.e, one category, human and animal research
issues, may not be relevant to all researchers.
219
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Therefore, one would expect a higher proportion
of programs than was observed to address at least
four of the categories.
Most educational efforts in other areas
typically identify goals and learning objectives as
a way of focusing teaching approaches and
assessing their success. In this sample, few of the
T32 programs (6/36) identified goals and
objectives. This would seem to imply that
programs do not approach training in the same
manner they would typically approach other
educational efforts.
“Ready made” materials and materials
sanctioned and made available by professional
organizations were a popular source of training
materials. This underscores the need to ensure
that these materials, which are so heavily relied
upon, are of high quality, complete, appropriately
tailored for the target audiences, and widely
available.
The most popularly used material,
institutional ethics policies, is critical for
trainees’ basic understanding of responsible
conduct of research. The proportion in the
sample who used these policies as a educational
tool could be viewed as unexpectedly low (45/75
[60%]).
Future Research
In addition to the findings discussed, this review
indicates the need for further research on
institutional approaches to education and training
in the responsible conduct of research. First,
additional research is needed on the
characteristics of training programs. A
description of primary and participating
instructors in training would be instructive,
particularly knowing the extent to which an
institution’s ethics faculty are involved in
program development, administration and
teaching. In addition, it would be useful to
understand the differences in approach and
content of training provided for trainees in
different disciplines, particularly in the clinical
sciences as compared to the basic sciences. This
information would point to differences in the
perceived needs for subgroups of trainees, and
could aid development of appropriate materials
and programs, for example, the use of core
programs with tailored components.
Second, research is needed on the
effectiveness of training initiatives. Evaluation of
a variety of programs and their approaches would
be particularly useful. Some target areas for
program evaluation include:
• the use of a core program plus tailored
discipline- and skill-specific components
• resource utilization-sharing by multiple
programs within and among institutions
• skill-based training programs, with assessment of trainee competencies
• the importance of goals and objectives of
programs as away to focus the educational
effort
• resource needs for “train the trainer” approaches
• the effectiveness of stand-alone training
materials
• the effectiveness of one-day programs compared to series of sessions
There is also a need to identify how to broaden
current training efforts to ensure that all
scientists-in-training are prepared to address
ethical dilemmas in their professional careers,
regardless of the source of funding for their
training. Such initiatives might include education
of institutional administrators about the
importance of responsible conduct of research
training beyond T32 trainees and the enlisting of
institutional commitments for broadened training
efforts. In addition, there is a need for improved
dissemination of effective approaches to
responsible conduct of research training in the
relevant professional literature.
The results of this review should not be
viewed as representative of responses to the NIH
mandate at either the programmatic or
institutional level because of the sample’s
limitations. The way the sample was selected
and the generality of the government’s request
for materials may have had some impact on the
results. Since the materials were collected
independently from this review, a targeted
questionnaire would provide more detailed
information. However, the results of this review
are a valuable first step in describing how
institutions and investigators meet the mandate
for training in responsible conduct of research
Conclusion
The intent of this pilot assessment was to
describe for the first time how institutions and
investigators are responding to the NIH mandate
for training in the responsible conduct of research
that is part of NIH Training Grant (T32) funding.
The results provide a snapshot of the variety of
220
–––––––––––––––––––––––––––––––––––––––––––––––––––– Mastroianni & Kahn, Encouraging Accountability in Research
approaches used in programs across the country.
Understanding the range of approaches taken
in the education and training in the responsible
conduct of research is a crucial part of any effort
to encourage accountability in research, on the
part of trainees, researchers, institutions, and
funders. Those engaged in training and
education can gain important insights for further
study given the diversity of approaches seen in
this review, while at the same time pointing to
the need for some consistency of training
content. Further, education and training in the
responsible conduct of research should be part of
all the training of all scientists and not a function
of the source of funding for training. Only by
assuring the highest standard of research conduct,
can we be confident that the trust the American
people continue to place in biomedical research
is truly deserved.
5. Korenman SG, Shipp AC, AAMC Ad Hoc Committee
on Misconduct and Conflict of Interest in Research
Subcommittee on Teaching Research Ethics. Teaching
the Responsible Conduct of Research through a Case
Study Approach: A Handbook for Instructors.
Washington, D.C.: Association of American Medical
Colleges, Division of Biomedical Research; 1994.
6. National Academy of Sciences. On Being a Scientist
(2nd Edition). Washington, D.C.: National Academy
Press; 1995.
7. Macrina FL. Scientific Integrity: An Introductory Text
with Cases. Washington, DC: ASM Press; 1995.
8. American Association for the Advancement of Science.
Integrity in Scientific Research: Five Video Vignettes.
Washington, D.C.: AAAS; 1996.
9. Bulger RE, Heitman E, Reiser SJ, editors. The Ethical
Dimensions of the Biological Sciences. Cambridge:
Cambridge University Press; 1993.
Acknowledgements
Thanks to William Raub and Lily Engstrom at
the U.S. Department of Health and Human
Services for providing access to the sample of
materials, and to Wayne Hartley of the Center for
Bioethics at the University of Minnesota for his
assistance in statistical compilation.
Notes
1. DHHS staff selected this approach to the collection of
resources because their primary purpose was to gain
insights into the scope and character of the materials
being used to teach responsible conduct of research, and
in a way that minimized the reporting burden for the
cooperating institutions. They recognized from the
outset that this approach would enable only qualitative
characterization at best, and unlike a formal survey,
would not yield readily analyzable data. (DHHS, 1997)
Bibliography
1. DHHS. Raub WF. Office Communication. Director of
the Office of Science Policy, Office of the Assistant
Secretary for Planning and Evaluation, DHHS, to a
sample of T32 recipient institutions, June 25 [on file
with authors]; 1996a.
2. DHHS. Research integrity project. [Unpublished
memorandum prepared by Preya Sharma on file with
authors]. 1996b Aug 15.
3. DHHS. Raub WF. Personal Communication, DHHS;
1997 Oct 24.
4. National Institutes of Health. NIH program
announcement PAR-98-006: Mentored scientist
development award in research ethics. NIH Guide
Grants Contracts 1997a; 26(37): Nov 7.
221
A Plea for Pursuing New Dimensions of Assessment in the Teaching and
Learning of Research Integrity
Carl Mitcham, Liberal Arts & International Studies, Colorado School of Mines, USA
Barbara M. Olds, Liberal Arts & International Studies, Colorado School of Mines, USA
Ronald L. Miller, Chemical Engineering & Petroleum Refining, Colorado School of Mines,
USA
Keywords: Engineering ethics, Ethics in science, Learning assessment, Responsible conduct of research
education, RCR, Scientific integrity, Science policy, Teaching assessment
Our basic thesis is simple: There are abundant research opportunities involved with the need to assess
the teaching and learning of research integrity. In one sense this thesis is a cliché. Research
opportunities are abundant everywhere; more research can be conducted on almost anything and
everything–even in quite narrowly defined areas such as the quantitative assessment of teaching and
learning about research integrity.
It is nevertheless possible to interpret our thesis in a broader and more provocative sense and to
argue for breaking out of a restricting if well established, four-sided system of constraints. The
teaching and learning of research integrity is, after all, concerned with integrity–from the Latin
integritas, which signifies not only purity or correctness but also and more fundamentally soundness
or completeness, the undiminished or unimpaired wholeness of a thing. Integrity is related to
integritas, bringing together. There is more to ethics than what has been going on in research ethics,
and research ethics will profit from more extensive connections than heretofore pursued.
Before making an effort to move beyond the constraints, it will be useful to describe in slightly
greater detail the two-dimensional box in which this issue of assessing the teaching and learning of
research integrity is currently confined.
Narrow Interpretations of RCR Education
It is increasingly common at research universities to teach courses or modules on research integrity or
the responsible conduct of research (RCR)–as is now required by National Institutes of Health and
Public Health Service grant award guidelines, and as has been reported more generally in Michael
Davis (1). To date, however, efforts to measure the effectiveness of RCR curricula have been limited
if not anecdotal. Nicholas Steneck’s bibliographic background report for the present proceedings
volume begins to identify such limits (2), although he is not as critical as we are of the present state of
affairs.
Constituting a first restriction, the whole literature on research integrity is highly concentrated in
the biomedical field. There are modest exceptions, but the most prominent instances of teaching and
Corresponding author: Carl Mitcham, Liberal Arts & International Studies, Stratton Hall 301, 1005 14th Street, Colorado
School of Mines, Golden, CO 80401, 303-273-3648 (voice), 303-273-3751 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
learning about RCR–and thus of possibilities for
RCR assessment–are found in the health care
fields, from general medicine to dentistry and
diverse medical research specialities. Given the
emphasis on informed consent issues in both
research and clinical practice, and the public
profile of regulations related to the treatment of
animals in research, this is perhaps to be
expected. It need not, however, be accepted
without question.
A second restriction is that research ethics
teaching focuses heavily on what may be termed
internalist over externalist issues. Issues
concerned with doing things right crowd out all
discussions about what might be the right things
to do; process overshadows substance.
Questions of precisely how to handle data
management, treat human and animal subjects,
pursue publication, deal with conflicts of interest,
and mentoring protocols dominate, at the expense
of critical reflection on the proper ends to pursue
with these methods (see the NIH Bioethics
Resources on the Web at nih.gov/sigs/bioethics/
researchethics.html, especially the NIH
supported link to Resources for Teaching
Research Ethics at medicine.ucsd.edu/research/
ethics/resources).
Still a third restriction is that although formal
RCR instruction obviously raises questions about
whether such teaching makes a difference–
whether it reduces research misconduct–
confirming evidence remains slight. In fact,
there is scant agreement even on the immediate
goals of RCR teaching and learning, thus making
it difficult to decide what would count as
evidence for or against short- or long-term
success. In consequence, many assessments of
RCR education have produced ambiguous
results.
Finally, a fourth restriction is that what
unambiguous assessment results do exist have
relied almost exclusively on the utilization and
adaptation of two specific instruments, the
Defining Issues Test (DIT) developed by James
Rest (3) and the Sociomoral Reflection Measure
(SRM) developed by John Gibbs (4), both of
whom had studied with, and in their work
attempted to more readily operationalize moral
development theorist Lawrence Kohlberg’s
Moral Judgment Interview (MJI). A clutch of
studies generated by Muriel Beabeau at the
University of Minnesota and her colleagues (5-7)
and Donnie Self at Texas A&M University and
his colleagues (8-10) all observe measurable if
modest correlations between ethics education and
moral reasoning skills, and some possible
implications for attitudes or behaviors. Michael
Kalichman and colleagues at the University of
California at San Diego (11, 12) have developed
an independent instrument that shows similar
correlations, although other studies (13) raise
doubts about the full significance of such
correlations.
No doubt partly as a result of the restrictions
in, if not the inconclusiveness of, existing
assessments, it has been argued that the goals of
ethics education should not be attitudes or
behaviors at all but simply skills and knowledge
(14). Indeed, the most common classroom
assessments of research ethics teaching
emphasize solely the learning of ethical
reasoning skills, with little attempt to gauge the
potential for long-term changes in behavior.
Arguments have even been made to the effect
that much more effective than RCR teaching in
the promotion of scientific integrity would be the
establishment of clear behavioral guidelines
followed by some form of monitoring such as
data audits (15). When education fails, try social
control.
Broader Interpretations of RCR
Education
Quantitative assessment of teaching and learning
about research integrity in the academic
classroom is thus boxed in on four sides. Such
constraint reflects the analytic and reductionist
strategy of modern scientific methodology, which
is based on the demand for and promise of
metrical results; this is a strategy that must
continue to be pursued. At the same time, there
is no need to completely restrict approaches to
such a flat plane. Indeed, especially given the
wealth of issues associated with moral education,
there are grounds for stepping beyond such
constraints–that is, for expanding our horizons in
the assessment of the teaching and learning of
research integrity.
First, boundaries may be extendend slightly
by recognizing the limits of particular
instruments such as the DIT and SRM. One
modest movement in this direction would be to
consider the relevance of other instruments for
assessing cognitive or intellectual development
such as the Reflective Judgment (RJ) scale
developed by Patricia King and Karen Kitchener
(16) on the basis of the work of William G.
Perry, Jr. (17). It may be noted, for instance, that
224
–––––––––––––––––––––––––––––––––––––––––––––––– Mitcham, et al., A Plea for Pursuing New Dimensions of Assessment
the RJ instrument has been pioneered at the
Colorado School of Mines (18) and repeated at
Pennsylvania State University (19) as a tool to
assess the intellectual development of
engineering students. Although not focused on
moral development, RJ has potential implications
for ethics learning that deserve exploration.
Second, precisely because RCR education
raises research questions about long- as well as
short-term effectiveness, the goals of the teaching
and learning about research integrity should
themselves become themes for research. This
would constitute, as it were, an initial step off the
flat place of quantitative research. Research into
goals, as opposed to research on the effective
implementation of goals, calls for more than
quantitative or empirical study. It calls for
historical and philosophical analysis and
reflection. It may be noted, for instance, that
current assessment strategies tend to carry
forward, more or less uncritically, the applied
ethics movement that arose during the 1980s.
At the very beginning of this revival Daniel
Callahan (20) proposed five goals for the
teaching of ethics in higher education: (a)
stimulating the moral imagination,
(b) recognizing ethical issues, (c) eliciting a
sense of moral obligation, (d) developing analytic
skills, and (e) tolerating and reducing
disagreement and ambiguity. Viewed against the
background of the analytic meta-ethics dominant
at that time, these were all worthy and even
modestly revolutionary goals. Historically,
however, the focus has increasingly narrowed to
simply developing analytic skills. The teaching
and assessment of research ethics has largely
accepted this narrow inheritance, as is reflected
in the very terminological emphasis on
“responsible conduct of research.”
Philosophically, there are even deeper historical
issues to be raised if RCR education is examined
in the light of such classic reflections on the
moral life as those present in the works of Plato,
Aristotle, and Augustine, not to mention the
Upanishads, the Sutras, the Torah, or the
Gospels.
Third, reflective reconsideration of the goals
of teaching and learning about research integrity
may stimulate recognition that as much if not
more pertinent teaching and learning goes on
outside the classroom as well as within it. This
recognition may, in turn, promote a search for
ways to assess meta-classroom learning. One
meta-classroom context is the professional
association. Yet lack of assessment is also
common among scientific professional societies.
Although most societies have codes of ethics that
clearly bear on research integrity, Mark Frankel,
director of the Scientific Freedom, Responsibility
and Law Program at the American Association
for the Advancement of Science (AAAS), has
concluded that few scientific societies are able to
tell whether these codes are working (21, 22).
Finally, extending such reflection even
further, it reasonably may be argued that such
internalist issues as data management, the
treatment of human and animal subjects,
publication protocols, conflicts of interest, and
mentoring standards cannot in reality be
separated from the focused externalist issues of
science and technology policy. Indeed,
international recognition of the immoral behavior
of some members of the medical research
establishment during World War II stimulated
adoption of the Nuremburg Code for free and
informed consent in human subjects research;
political concern in the United States during the
1980s about the improper behavior of scientists
using public funds has been one of the primary
drivers to promote RCR education. Surely both
of these historical points deserve to be taught
along with the norms of data management and
peer review.
Three (Intentionally Provocative)
Suggestions
Without attempting to draw definitive
conclusions from this four-fold unsystematic
expansion of the RCR educational context, we
would like to pose three summary pleas for the
pursuit of new dimensions in assessing the
teaching and learning of research integrity. In
this way we seek to make common cause with
others such as J. Andre (23) who have also called
for not limiting professional ethics courses to
moral reasoning analyses.
First, in light of the public policy roots of
RCR education and the larger philosophical and
religious traditions of ethics, is it appropriate to
focus on reasoning or analytic skills in ways that
slight attitudes and behavior? Would it not be
possible to develop, for instance, an instrument
for assessing cynicism and idealism among
students, and indeed to attempt to counteract a
too common passive cynicism? Social idealism
is an honorable heritage of the scientific
tradition, as exhibited by scientific leaders from
Francis Bacon to Albert Einstein. In a talk to
225
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
scientists and engineers at the California Institute
of Technology in 1931, for instance, Einstein
argued that
Concern for man himself and his fate must
always form the chief interest of all technical
endeavors . . . in order that the creations of our
mind shall be a blessing and not a curse to
mankind. Never forget this in the midst of your
diagrams and equations (24).
Contemporary witnesses to this tradition of
idealistic science can be found in the public
interest activism of International Pugwash
founding member and Nobel Peace Prize winner
Joseph Rotblat (25) as well as SunMicrosystems
co-founder Bill Joy (26). Introduction to such
moral heros of what may be termed scientific
social idealism should not be slighted to carve
out time for parsing moral dilemmas in conflict
of interest or authorship adjudication, as
important as these may well be.
Second, does research ethics need to be
conceptualized as distinct from engineering
ethics, as it has been so far? Does the
engineering/science separation not perpetuate
stereotypes of the pure scientist versus the
applied engineer–images at odds with reality in a
world in which virtually all science is dependent
on complex technological instrumentation?
Moreover, is it not the case that scientists have
something to learn from engineers regarding
ethics? Long before scientists, engineers
formulated ethics codes at the beginning of the
20th century; they also began taking them into
the classroom well before scientists (26).
In the engineering education community
today, considerable attention currently is being
given to ABET Criteria 2000, the new set of
accreditation guidelines developed by the
Accreditation Board for Engineering and
Technology (available at www.abet.org).
Criterion 3, for instance, contains 11 attributes
that graduates should possess, including
“understanding of professional and ethical
responsibility.” Many engineering programs are
developing methods to assess student progress in
this area, including the use of such instruments as
the DIT. There are also unexplored possibilities
for assessing teaching and learning in
engineering ethics by correlating results from the
Fundamentals of Engineering (FE) and
Professional Engineering exams required of all
professional engineers.
Research integrity should not be separated
from academic integrity in the research
university setting. The practical RCR
educational potential of student honor codes–
some specific to schools of engineering–perhaps
deserves as much attention as relations to
engineering ethics codes.
Finally, does the assessment of teaching and
learning itself not also deserve some assessment.
An assessment of teaching and learning
assessment requires both community engagement
and critical analysis. The practice of any
assessment should be guided by the principles
developed by the Assessment Forum of the
American Association for Higher Education (28),
which include the following:
•Assessment is most effective when it reflects
an understanding of learning as multidimensional, integrated, and revealed in performance over time.
•Assessment works best when the programs it
seeks to improve have clear, explicitly stated
purposes.
•Assessment works best when it is ongoing.
It is our contention that assessing of the teaching
and learning of research integrity has only begun.
This is true not only in the narrow senses
associated with quantitative investigation of
RCR, but also in the much broader senses of
attempts to develop relations between RCR and
idealistic science activism, engineering ethics
and academic codes, and the reiterative
assessment of assessment itself.
Bibliography
1. Davis, M. Ethics and the university. New York:
Routledge; 1999.
2. Steneck, N. Assessing the integrity of publicly funded
research. Background report. A Research Conference
on Research Integrity, Bethesda, MD, Nov. 18-20, 2000.
3. Rest JR., Narváez D, Bebeau MJ, Thoma S.
Postconventional moral thinking: A neo-Kohlbergian
approach. Mahwah, NJ: Erlbaum; 1999.
4. Gibbs, John C., Karen S. Basinger, and Dick Fuller.
Moral maturity: Measuring the development of
sociomoral reflection. Hillsdale, NJ: Lawrence
Erlbaum; 1992.
5. Bebeau MJ. Designing an outcome-based ethics
curriculum for professional education: Strategies and
evidence of effectiveness. Journal of Moral Education
1993; 22(3): 313-26.
6. Bebeau MJ, Thoma SJ The impact of a dental ethics
curriculum on moral reasoning. J Dent Educ 1994;
58(9): 684-92.
7. Bebeau MJ, Rest JR, Narvaez D. Beyond the promise: A
perspective on research in moral education. Educational
Researcher 1999 May; 28(4): 18-26.
226
–––––––––––––––––––––––––––––––––––––––––––––––– Mitcham, et al., A Plea for Pursuing New Dimensions of Assessment
8. Self DJ, Wolinsky FD, Baldwin Jr DC. The effect of
teaching medical ethics on medical students= moral
reasoning. Acad Med 1989; 64: 755-9.
9. Self DJ, Schrader DE, Baldwin Jr DC, Root SK,
Wolinsky FD, Shadduck JA. Study of the influence of
veterinary medical education on the moral development
of veterinary students. J Am Vet Med Assoc 1991;
198(5): 782-7.
10. Baldwin Jr DC, Daugherty SR, Self DJ. Changes in
moral reasoning during medical school. Acad Med
1991; 66(9) Suppl: S1-S3.
11. Kalichman MW, Friedman PJ. A pilot study of
biomedical trainees= perceptions concerning research
ethics. Acad Med 1992; 67(11): 769-75.
12. Brown S, Kalichman MW. Effects of training in the
responsible conduct of research: A survey of graduate
students in experimental sciences. Science and
Engineering Ethics 1998; 4(4): 487-98.
13. Eastwood S, Derish P, Leash E, Ordway S. Ethical
issues in biomedical research: Perceptions and practices
of postdoctoral research fellows responding to a survey.
Science and Engineering Ethics 1996; 2(1): 89-114.
14. Elliott, D, Stern JE. Evaluating teaching and
students=learning of academic research ethics. Science
and Engineering Ethics 1996; 2(3): 345-66.
15. Shamoo AE, Dunigan CD. Ethics in research. Exp Bio
Med 2000; 224(4): 205-10.
16. King PM, Kitchener KS. Developing Reflective
Judgment. San Francisco: Jossey-Bass; 1994.
17. Perry Jr WG. Forms of Intellectual and Ethical
Development in the College Years: A Scheme. New
York: Holt Rinehard and Winston; 1970.
18. Pavelich, MJ, Moore WS. Measuring the effect of
experiential education using the Perry model. Journal of
Engineering Education October 1996 85(4): 287-92.
19. Palmer B, Marra RM, Wise JC, Litzinger TA. A
longitudinal study of intellectual development of
engineering students: What really counts in our
curriculum? 30th ASEE/IEEE Frontiers in Education
Conference, Proceedings Oct 18-21 2000: S3A-2-6.
20. Callahan D. Goals in the teaching of ethics. In:
Callahan D, Bok S, editors. Ethics Teaching in Higher
Education. New York: Plenum Press; 1980. p.61-80.
21. Brainard, J. 2000. Societies urged to fight research
misconduct. Chronicle of Higher Education April 21,
2000; 46(33): A38.
22. DuMez E. The role and activities of scientific societies
in promoting research integrity. Professional Ethics
Report 2000; 13(3): 1 and 7-8.
23. Andre J. Beyond moral reasoning: A wider view of the
professional ethics course. Teaching Philosophy 1991
Dec; 14(4): 359-73.
24. Einstein A. Einstein on Peace. Nathan O, Norden N,
editors. New York: Simon and Schuster; 1960.
25. Rotblat, J. Taking responsibility. Science 2000 Aug 4;
289: 729.
26. Joy, B. Why the future doesn’t need us. Wired 2000
April; 8.04: 238-62.
27. Mitcham, C. The achievement of “Technology and
Ethics”: A perspective from the United States. In:
Technology and Ethics. Brussels: In press.
28. Assessment Principles. ASEE Prism 1997 Oct; 7(2):
10.
227
7. Responsible Conduct of Research Courses
The Responsible Conduct of Animal Research
Lida Anestidou, University of Texas-Graduate School of Biomedical Sciences at Houston,
USA
Elizabeth Heitman, University of Texas-School of Public Health at Houston, USA
Key Words: Animal, Animal research, Curriculum, Responsible conduct of research, RCR
Students in graduate education in the basic sciences have a high probability of using live animals at
some point in their research training. Although animal rights are a volatile issue for public debate, the
use of animals in graduate science education raises little controversy among research trainees. Due to
a National Institutes of Health (NIH) mandate, most graduate science programs today offer
instruction in the responsible conduct of research that may include the ethics of experimentation with
animal subjects1. Similarly, federal requirements for animal research review committees include
provisions for the technical training of students and others conducting procedures with live animals2.
As part of their responsibilities for overseeing the housing and care of research animals and the
safe conduct of research, the veterinary staff of the University of Texas-Health Science Center at
Houston offers formal training sessions in the safe and humane handling of laboratory animals and
proper techniques for a variety of procedures. These sessions are offered regularly and are often
filled well in advance.
The University’s Institutional Animal Care and Use Committee (IACUC) and the veterinarians of
the Center for Laboratory Animal Medicine and Care (CLAMC) are justly proud of their record of
concern for animal welfare and the institution’s humane research practices. Nonetheless, faculty
involved in the required research ethics course at the University of Texas-Graduate School of
Biomedical Sciences at Houston routinely hear comments from first- and second-year students who
feel uncomfortable in their animal work, particularly in mastering routine procedures after the formal
training has ended. Often these comments, made in small group discussions, are about the value of
biomedical research with animals and questions about animal suffering. The same students typically
express unwillingness to ask for help or further instruction for fear of criticism from their faculty and/
or older peers. Nonetheless, many agree that more direct training in the handling and use of specific
research animals would improve their skills, confidence, and attitude toward the work, as well as
improve the quality of their research.
Research in medical education has demonstrated that trainees who ignore or discount their
emotional responses to patients and the pain that medical procedures may cause are at risk of
becoming emotionally stifled, cynical, and even punitive in response to the suffering of others. In
contrast, by including formal attention to the emotional dimensions of patient care, medical educators
have been shown to foster trainees’ compassion and personal satisfaction in their work3. Moreover,
by learning to identify and address their emotional responses directly, medical trainees have been
Corresponding author: Lida Anestidou, UT-Houston Mdical School, Deptartment of Integrative Biology, Pharmacology &
Physiology, 6431 Fannin, Houston, TX 77030, 713-500-5082 (voice), 713-500-7444 (fax),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
found to improve the accuracy of their diagnosis
and treatment. Parallel risks and opportunities
exist for researchers who use animals, and efforts
to address the emotional dimension of animal use
make a valuable addition to the institution’s
efforts to enhance the integrity of scientific
research.
In response to the perceived need for more
focused education and hands-on training for
graduate students in the biomedical sciences, the
authors organized a new intensive course entitled
“The Humane Use of Animals in Biomedical
Research.” The course offers a highly structured
and multidisciplinary approach to responsible
animal research. Its goal is to provide instruction
in the ethics and regulatory aspects of animal
research, approaches to the reduction of the
numbers of animals used in specific protocols,
including alternative research methods, and
extensive practical training tailored to the
individual animal model that each participant
expects to use. Using a combination of didactic
sessions, case discussions, and direct, hands-on
laboratory instruction under the close supervision
of institutional veterinarians, the course faculty
seek to enhance students’ theoretical knowledge
base, technical skills, practical compassion, and
professional confidence.
An aspect unique to this course is the
inclusion of structured group discussion intended
to help students address their personal
experiences, concerns, values, and attitudes
regarding their interaction with animals and the
demands of animal research. Faculty facilitators
help students recognize and prepare for the
personal and ethical challenges of live animal
experimentation using a modified version of the
Balint method, which has been used in medical
education to promote personal awareness and
effective, compassionate patient care4.
The course was offered to graduate students,
post-doctoral fellows, research associates and
technicians across the University for the first
time in July 2000. The course schedule,
including topics, instructors, and format appears
in Table 1. The list of assigned readings for the
course appears in the Appendix.
Evaluation (Students’, Instructors’,
Course Coordinators’)
As part of the wrap-up on the last day of class,
students were encouraged to provide a
comprehensive evaluation of the course, with
particular attention to the aspects of reading
assignments, class structure and timing, and the
integration of theoretical material and practical
skills. One week following the end of the course,
the instructors and course-coordinators held a
similar debriefing and evaluation session with a
special focus on potential changes for subsequent
course offerings. The following constructive
suggestions were made by course attendees:
Positive points
1. The readings were comprehensive and
challenging.
2. The practical aspects and methodologic
training were invaluable even to students not
working in laboratories.
3. Learning about regulations and IACUC
activities from IACUC members was very
enlightening about the practicalities of
researchers’ obligations and institutional review.
4. The information on alternative methods
to animal research was important to new
researchers considering a variety of techniques.
5. The presence, knowledge, and guidance
of veterinarians were a tremendous intellectual
and practical asset.
6. The variety of viewpoints presented by
interdisciplinary faculty and guest lectures was
useful in understanding the scope of animal
research and its ethical gray areas.
7. Discussion of the personal demands of
research was valuable for integrating
interdisciplinary issues and helpful for students
seeking to come to terms with the demands of
their work.
8. The intensive class format enhanced
rapport among students and faculty.
Drawbacks and obstacles
1. The time commitment in an intensive 2week format was extremely hard for students to
manage along with their regular daily schedules.
2. The summer offering made scheduling
faculty assignments difficult because of their
travel schedules and other special commitments.
3. The logistical complexity of organizing
multiple faculty in both classroom and laboratory
was very time consuming for the course
organizers.
4. More practical discussion of alternative
methodologies by practicing researchers was
needed.
5. Students in science are often
uncomfortable with ethical ambiguity and like
232
––––––––––––––––––––––––––––––––––––––––––––––––––– Anestidou & Heitman, Responsible Conduct of Animal Research
clear answers.
6. Faculty need to focus more on the links
between ethical debate, science policy, and
practical demands of research.
7. The costs of laboratory materials for a
larger enrollment are likely to be considerable
8. Students’ perception of the need for such
DATE
CLASS
Topic
Monday
07/17
Lecture
•
Tuesday
07/18
Lecture
•
Wednesday
07/19
Lecture
•
•
Thursday
07/20
Lecture
•
Friday
07/21
Lecture
•
•
•
•
•
•
Monday
07/24
Tuesday
07/25
Lecture
Lab
Lecture
Lab
Wednesday
07/26
Lecture
Lab
Thursday
07/27
Lab
•
Friday
07/28
Lecture
Discussion
•
•
•
•
•
•
•
•
a course is variable. Faculty need to identify and
address the multiple goals of different students in
different backgrounds throughout the class.
Conclusion
Evaluation by the student and faculty participants
and a critique of the course by the course
INSTRUCTOR
Historical uses of animals in
biomedical research
Ethical and regulatory
perspectives on animals in
biomedical research
Scientific approaches to refining
animal research (the three Rs)
Balint group discussion
IACUC: its function and
responsibilities
How to fill out animal protocol
forms
Alternatives to animal models
AAALAC and the Guide
Housing and standards of care for
laboratory animals- Facility tour
Balint group discussion
Heitman
Anestidou
Heitman
Anestidou
Smith
Heitman
Anestidou
Heitman
Anestidou
Bjerckey
Goodwin
Blasdel
Heitman
Anestidou
Mouse biology, care, and
management
Head
General anesthesia and pain
control; rodent-specific protocols;
Anesthesia matters (video)
Rodent anesthesia practicum
Monkey retirement facility
speaker
Balint group discussion
Smith
Disposition of animals after
research
Euthanasia
Wrap up course material
Evaluation
Griffin
Heitman
Anestidou
Blasdel
Head
Heitman
Anestidou
Table 1. The Human Use of Animals in Biomedical Research-Course Outline and Schedule
233
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
coordinators resulted in significant enthusiasm to
repeat it. The course will be offered again in the
summer 2001 term, using mostly the same
didactic methods and material, but in a less
intensive format. The course coordinators,
CLAMC veterinarians, IACUC members, and
the University’s administration hope that in the
next few years the course will be developed into
both an integrated part of many students’
education at the Graduate School and a
continuing education course available to
researchers and others from outside our
institution.
2. National Institutes of Health, Alcohol, Drug Abuse, and
Mental Health Administration. Reminder and update:
requirement for programs on the responsible conduct of
research in National Research Service Award
institutional training programs. NIH Guide 1990; 19
(30): 1.
3. Gorlin R, Zucker, HD. Physicians’ reactions to patients:
a key to teaching humanistic medicine. New England
Journal of Medicine 1983; 308: 1059-63.
4. Novack DH, Suchman AL, Clark, W Epstein RM,
Najberg E, Kaplan C. Calibrating the physician.
Personal awareness and effective patient care. JAMA
1997; 278: 502-09.
Bibliography
1. U.S. Department of Health and Human Services. Guide
for the care and use of laboratory animals (revised).
NIH Publication #86-23. Washington, DC: DHHS,
1985.
Appendix.
The Human Use of Animals in Biomedical Research-Course Readings
(by topic)
History of animals in biomedical research; ethical & regulatory perspectives on animals
in biomedical research
•
•
•
•
F. Barbara Orlans, “The Beginnings of Institutionalized Animal Experimentation” and “Current Attitudes and Ethical
Arguments” in In the Name of Science: Issues in Responsible Animal Experimentation, New York: Oxford University
Press, 1993: 3-34.
Caplan, Arthur, “Beastly Conduct: Ethical Issues in Animal Experimentation”, Science, 1983, 406: 159-169
Brody, Baruch “The Use of Animals in Research” in the Ethics of Biomedical Research: An International Perspective,
New York: Oxford University Press, 1998: 11-30.
National Association for Biomedical Research, “The Strict Regulations that Govern Research” Animal Research Facts,
http://www.fbresearch.org/research98.htm
Procurement of animals for research and education; Scientific approaches to refining
animal research (the three Rs)
•
•
•
•
•
•
•
•
•
•
F. Barbara Orlans, “The Source of Laboratory dogs and Cats: Pound versus Purpose-Bred Animals”, in In the Name of
Science: Issues in Responsible Animal Experimentation, New York: Oxford University Press, 1993, 209-220.
“Shelter Intake and Euthanasia Trends”, Animal Policy Report 2000, 14 (2): 2.
Judith Reitman, “From the Leash to the Laboratory”, Atlantic Monthly 2000, 286(1): 17-21.
“Pet Theft: Urban Myth Makes Useful Propaganda”, FBR Facts (Foundation for Biomedical Research), 2000, 7(2), 2
pages. http://www.fbresearch.org
Joanna Weiss, “Squid’s Fate: Science of Seafood”, Houston Chronicle (from Boston Globe), June 27, 2000, 3D.
W.M.S. Russell & R.L. Burch, “Introduction” in The Principles of Humane Experimental Technique. London: Methuen
& Co., 1959, 3-13.
Alan M. Goldberg, Joanne Zurlow, & Deborah Rudacille, “The Three Rs and Biomedical Research” (editorial), Science,
1996, 272: 1403.
Ruth Ellen Bulger, “Use of Animals in Experimental Research: A Scientist’s Perspective”, Anatomical Record, 1987,
219: 215-220.
National Association for Biomedical Research, “Animals in Research 1998”, Animal Research Facts, http://
www.fbresearch.org/research98.htm
Michael F.W. Festing, et al., “Reducing the Use of Laboratory Animals in Biomedical Research: Problems and Possible
Solutions” — The Report and Recommendations of ECVAM Workshop 29, ATLA 1998, 26: 283-301.
234
––––––––––––––––––––––––––––––––––––––––––––––––––– Anestidou & Heitman, Responsible Conduct of Animal Research
The function and responsibilities of the IACUC
•
•
•
•
•
Robert A. Whitney, Jr., “Animal Care and Use Committees: History and Current National Policies in the United States”,
Laboratory Animal Science 1987, Special Issue, 18-21.
Henry J. Baker, “Essential Functions of Animal Care and Use Committees”, Laboratory Animal Science 1987, Special
Issue, 30-33.
“Consensus Recommendations on Effective Institutional Animal Care and Use Committees” Laboratory Animal Science
1987, Special Issue, 11-13
F. Barbara Orlans, “What Does the Public Have a Right to Know?”, in The Human Use of Animals Biomedical
Research. New York: Oxford University Press: 103-117.
UT -Houston Health Science Center, Animal Welfare Committee, “Animal Protocol Review Form”, revised 11/1/96.
Alternatives to animal models
•
•
•
•
•
•
Michael Ballis, “Why is it Proving to be So Difficult to Replace Animal Tests?” Lab Animal 1998, 27 (5): 44-47.
Richard N. Hill & William S. Stokes, “Validation and RegulatoryAcceptance of Alternatives”, Cambridge Quarterly of
Healthcare Ethics 1999, 8, 73-79.
Jacques LeClaire & Odile De Silva, “Industry Experience with Alternative Methods”, Toxicology Letters 1998, 102-103:
575-579.
Seymour Levine & Arthur Saltzman, “An Alternative to Overnight Withholding of Food from Rats”, Contemporary
Topics (American Assn. for Laboratory Animal Science) 1998, 37: 59-60.
Sharron Kirchain & Robert P. Marini, “A Tissue Harvesting Program as a Method for Implementing the 3Rs of
Biomedical Research”, Lab Animal 1998, 27 (8): 37-39.
Adrian Smith, Richard Fosse, David Dewhurst, & Karina Smith, “Educational Simulation Models in the Biomedical
Sciences”, ILAR Journal 1997, 38 (2), 82-88.
Standards of care and housing for laboratory animals
•
National Research Council, Institute of Laboratory Animal Resources, Guide for the Care and Use of Laboratory
Animals, Washington, DC: National Academy Press, 1996.
Anesthesia and pain control
•
•
•
•
Lawrence R. Soma, “Assessment of Animal Pain in Experimental Animals”, Laboratory Animal Science 1987, Special
Issue, 71-74.
American College of Veterinary Anesthesiologists, “Position Paper on the Treatment of Pain in Animals”, JAVMA 1998,
213(5), 628-630.
American Association for Laboratory Animal Science, “Position Statement: Recognition and Alleviation of Pain and
Distress in Laboratory Animals”, http://aalas.org
American Association for Laboratory Animal Science, “Policy #12 — Consideration of Alternatives to Painful/
Distressful Procedures — June 21, 2000. http://aalas.org
Disposition of animals after research
•
•
American Veterinary Medical Association Panel on Euthanasia, “1993 Report of the AVMA Panel on Euthanasia”,
JAVMA 1993, 202: 229-249.
UT -Houston Health Science Center, Center for Laboratory Animal Medicine and Care”, “Animal Adoption Release
form”, revised 11/1/96.
Wrap up and evaluation
•
Diane J. Gaertner, Lele K. Riley, & Dale G. Martin, “Reflections on Future Needs in Research with Animals”, ILAR
Journal 1998, 39: 306-310.
235
An Effective Short Course on Research Integrity
Bernard S. Gerstman, Department of Physics, Florida International University, USA
Key words: Course, Ethics, Integrity, Interactive, Research, Responsible conduct of research, RCR, Training
Ethical conduct in research has always been considered of utmost importance within the research
community. Historically, it was assumed that scientific ethics did not require special training.
Instead, the ethical manner in which to carry out research was presumed to be learned by new
scientists automatically and unconsciously, as if by osmosis, as the technical aspects of the research
were carefully taught by their superiors. This was of course, never true. Mendel and Millikan may
have fudged their data, along with numerous others of less renown.
More recently, consideration has been given to developing methods for training scientists in
research ethics, rather than relying on osmosis (1). Part of the impetus for this change is that the
problems associated with unethical procedures in research have become especially visible to the
public when they occur in research in the health sciences (2). This paper reports on a course of short
duration that is designed to train students efficiently and effectively in the ethical conduct of research.
Design
The course is designed for graduate students and undergraduates who have shown an interest in a
career in science. There is no obvious reason why the course design would not be applicable to
students outside the sciences. At this time, all science majors at the home institution do not take the
course. The science undergraduates who are required to take the course are affiliated with special
programs such as the Research Experience for Undergraduates funded by the NSF as well as NIH
funded programs.
The course is designed to meet for one hour each week and to contain a maximum of 15 students.
If necessary, such as in summer sessions, the course can be compressed into a two-week period, but
some of its effectiveness is lost. This will be discussed later in this section when the reason for this
loss in effectiveness will be clear.
The initial course meetings are organized like a traditional class with the faculty member
explaining various aspects of research integrity and unethical behavior. This is best introduced by a
short (one hour) summary of the general principles of ethics in western society, which can then be
used as the basis for the principles of research integrity and ethics. It is important that this
explanation of ethics in general be presented as a summary. If it is presented in another form, such as
an “Introduction to Western Ethics” or any other form that does not convey immediate de facto
credibility, the course runs the danger of degenerating into a philosophy discussion on ethics in
general. Valuable time will then be taken from the specific goal of training the students in scientific
integrity and the course is likely to be neither short nor effective.
In addition to explaining the principles of research integrity, it also is important to be explicit
Corresponding author: Bernard S. Gerstman, Department of Physics, Florida International University, University Park,
Miami, FL 33199, 305-348-3115 (voice), 305-348-6700 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
about the importance of adhering to these
principles. Thus, the first few lectures of the
course should cover the following topics:
1) the general principles of Research Integrity
(1);
2) how scientific progress is enhanced by
adherence to integrity by all researchers;
3) how scientific progress is slowed by unethical behavior, or even the perception thereof;
and
4) the direct impact of ethical misconduct in
research:
i) wasted money by universities and funding
agencies,
ii) wasted time by researchers who trust the
results of others, and
iii) injury or death to patients (biomedical
research).
The middle part of the course shifts to a
preceptorial structure with faculty led discussions
of selected reading material on recent cases
concerning violations of research integrity.
These case studies summarize the accusations,
how they were investigated, the decisions that
were reached, and penalties imposed, if any.
These case studies can be found in the Annual
Report from the Office of Research Integrity of
the Department of Health and Human Services
(3).
These case studies supply concrete examples
of the topics discussed in the first part of the
course. The vast majority of cases involve data
fabrication and falsification. This also presents
the opportunity to discuss types of research
misconduct that are common but insidious:
sloppy data taking and self-deception (4). In
these instances, the researcher is not consciously
violating the principles of ethical behavior.
Unfortunately, because the misconduct is
unconscious, there is no chance for selfcorrection (5). The case studies are useful in
training the students against sloppy data taking
and self-deception, which can appear to be, or
easily become, data fabrication or falsification.
The case studies also present concrete
examples of a new topic — the penalties suffered
by researchers who are found to violate the
principles of research integrity. The usual
penalties(3) are disbarring from receiving federal
funding for 3 to 5 years, monitoring of a
researcher by the home institution, mandatory
retraction or correction of publications, and
occasionally dismissal. Students initially
consider these penalties too light and suggest
criminal prosecution. The faculty member at this
point can explain the severe ramifications of
these penalties for the researcher’s career.
The third, and last, part of the course is the
most important for successfully conveying the
principles of research integrity and the necessity
of adhering to these principles. It requires each
student to make a half-hour presentation to the
class about a case of suspected unethical
behavior in research that they have investigated
through a literature search. The students are
expected to use what they have learned in the
earlier parts of the course in discussing the
following points:
1) an explanation of what actions constituted
unethical behavior, entailing enough of an
explanation of the scientific research so that
other students can understand why the
behavior was unethical;
2) how the unethical behavior was uncovered;
3) what the motivation might have been for the
unethical behavior;
4) what, if any, penalties (real or intangible)
were suffered by the perpetrators; and
5) what penalties the student thinks would have
been appropriate.
Information for these presentations can be
obtained from books(6,7,8) on the subject,
science magazines such as Scientific American,
and with especially well-known and recent cases,
newspapers and general readership magazines.
Students are informed early in the course about
the presentation and are told to choose a case as
soon as possible. It is hoped that by giving the
students several weeks to prepare for their
presentation, they will use the time to follow a
meandering path in their literature search and
learn about several different cases. If two
students choose the same case, the second
student to notify the faculty member is instructed
to pick another case.
Results
The first two parts of the course give the students
a customary introduction to the issues of research
integrity. The third part of the course is crucially
important for consolidating these issues. The
students are enthusiastic about making their
presentation and peer pressure motivates them to
do a thorough job. The presentation forces the
238
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gerstman, Short Course on Research Ethics
students to “step into the mind” of a scientist
who is behaving unethically. This obliges them
to confront the temptations to behave unethically
and solidifies the need for self-vigilance.
Conclusion
A short course can be effective in conveying the
necessity of integrity in research and in training
the students on how to perform research in an
ethical manner. For the course to be effective,
the students must be required to take an active
role. A class presentation by each student is of
crucial importance and the most important
element of the course.
Bibliography
1. Marshall Thomsen, Editor, Ethical Issues in Physics:
Workshop Proceedings, July 17-18, 1993, Eastern
Michigan University, Ypsilanti, Michigan.
2. United States Congress House Committee on
Government Operations: nineteenth report. Are
scientific misconduct and conflicts of interest hazardous
to our health? Washington: U.S. G.P.O., House report /
101st Congress, 2d session ; 101-688, September 10,
1990.
3. Office of Research Integrity (US), Annual Report 1998,
Office of the Secretary, Office of Public Health and
Science, Department of Health and Human Services.
4. Lederberg J, Commentary: Sloppy Research Extracts a
Greater Toll Than Misconduct, The Scientist, February
20, 1995, page 13.
5. Thomsen M and Resnick D. The Effectiveness of the
erratum in avoiding error propagation in physics,
Science and Engineering Ethics, 1995, 1:3.
6. William Broad and Nicholas Wade. Betrayers of the
Truth, Simon and Schuster, New York, 1983.
7. Alexander Kohn, False Prophets: Fraud and Error on
Science and Medicine, Basil Blackwell, New York,
1986.
8. Robert Park, Voodoo Science: The Road from
Foolishness to Fraud, Oxford University Press, New
York, 2000.
239
Resources for Instruction in Responsible Conduct of Research
Michael W. Kalichman, Department of Pathology, University of California San Diego, USA
Francis L. Macrina, Philips Institute of Oral and Craniofacial Molecular Biology,Virginia
Commonwealth University, USA
Jeffrey P. Kahn, Center for Bioethics. University of Minnesota, USA
Key Words: Instruction, Internet, Research ethics, Responsible conduct of research, RCR, World Wide Web
In recent years it has become clear that, despite its importance, training in ethics, standards, and
responsible conduct is too frequently minimal or absent in academic science. This deficit is being
corrected in part by the requirement that fellows funded by National Institutes of Health (NIH)
training grants receive such instruction. This requirement has been important to the development of a
variety of outstanding texts now available (1-8) and a number of very effective, thoughtful programs
developed across the country. However, no network provides ready communication about the goals,
resources, tools, or methods for such programs. As a result, the design and implementation of a new
program in responsible conduct of research (RCR) training can be frustrating if not intimidating.
It can be difficult to pull together material for a new RCR program. Unfortunately, such effort is
frequently duplicated even within the same institution and the resulting RCR instruction is uneven in
quality, topics covered, and audience reached. In addition, it appears that the most likely audience for
these programs has been limited to only those NIH trainees required to take part. This is contrary to
the goal that such training is best met by a program that reaches the broad spectrum of the academic
community including staff, undergraduates, medical students, pre- and post-doctoral fellows, and both
junior and senior faculty. However, with the rapid changes in access to the Internet, the technology is
now available to make formats, examples, contacts, and resources immediately available to any
institution interested in providing effective RCR instruction.
The Internet is now being used for a variety of purposes relevant to RCR instruction (9-17). In
just the last couple of years, these resources have evolved rapidly in both form and content. Many
institutions have created web sites that provide considerable content as well as lists of links to other
sites (9-10), typically in the area of ethics. In addition, many universities now have course materials
posted on the web (11-13) and in some cases Internet-based courses, designed to be run without
traditional classroom meetings (14,15). Finally, web-based information is available on programs such
as the “Survival Skills and Ethics” (16) and “Teaching Research Ethics” (17) workshops for teaching
about the teaching of responsible conduct of research. All of these resources provide important
contributions, but diverse audiences, differences between disciplines, and the frequency of significant
new developments, all minimize the value of any one approach to RCR instruction. The proposed
alternative is a continually evolving web site.
Corresponding author: Michael Kalichman, Ph.D., Director, Research Ethics Program, 0003, University of California, San
Diego, La Jolla, CA 92093-0003, 858-822-2027 (voice), 858-534-4772 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
A web site dedicated to resources on
instruction in the responsible conduct of research
could provide national access to the most
effective programs, materials, and methods for
such training. The long-term goal would be to
improve the quality and extent of RCR
instruction. Such a site would not only make it
possible for virtually any institution to develop
an RCR program, but would also increase
general awareness about what is being done, and
what can be done, to enhance instruction in RCR.
It is intended that this site would complement,
not replace, other tools for RCR programs (1-17).
Given the ongoing NIH requirement for training
grants to include instruction in RCR and the
proposed extension of this requirement to all
research staff working on PHS-supported
projects, many institutions need help to either
extend limited existing programs or to develop
new programs. However, even in the absence of
federal requirements, it should be enough to
know that access to proven materials and
methods for RCR instruction can only help to
foster responsibility in the conduct of research.
Methods
The core of the web site was first assembled from
materials already available for courses taught at
the University of California in San Diego,
Francis Macrina’s course at Virginia
Commonwealth University and his book on
“Scientific Integrity,” and course materials under
development at the University of Minnesota.
The site was initially designed to cover nine
topic areas: (1) Getting started; (2) Defining the
goals of an RCR program; (3) Elements of an
RCR program; (4) Guidelines, requirements, and
procedures; (5) Resources; (6) Case studies; (7)
RCR programs; (8) Contacts; and (9) Evaluating
an RCR program. The plan was that these
primary divisions would be subdivided into
topics generally considered to be relevant to
responsible conduct of research (e.g., conflict of
interest, use of animals in research, and criteria
for authorship). Using this framework for the
content available in the authors’ institutions, the
initial goals were to design and implement a
framework for the web site, insert materials from
the authors’ institutions, and annotate the
resources.
After completion of the first steps of the
project, the web site was to be improved through
an iterative process, including three phases of
external reviews, plus soliciting of suggestions
for additional materials. For this review phase,
the primary goals were to solicit new materials
from other institutions, modify the framework of
the site as needed to accommodate the new
resources and reviewer suggestions, annotate the
resources, and publicize the site.
For evaluation of the web site, reviewers
were asked to rank various aspects of the site’s
form and content in a brief online form.
Numerical rankings were to be scored using a
scale of 1 to 5 (1=very low, 2=low, 3=average,
4=high, 5=very high). Additional questions
asked for specific suggestions to improve the
web site, including recommendations of material
to be added.
Results
The first phase of this project was to develop a
web site framework for presenting resources on
instruction in the responsible conduct of research.
Beginning in September of 1999, work on the
web site began at the University of California,
San Diego with ongoing assistance from
collaborators at Virginia Commonwealth
University and the University of Minnesota.
During the initial months, the web site evolved
through several different formats until a version
was considered ready for external review. In July
of 2000, the first phase of external review was
begun. The three planned phases of review were
completed by November 1, 2000.
The first external review was based on a
limited release of the web site to four reviewers
(two from government agencies and two from
non-governmental organizations). In a series of
questions about web site form and content, scores
averaged between 3.25 and 4.75 with medians
between 3 and 5. The lowest scores were
generally assigned to the appearance and
navigability of the web site. Several valuable
suggestions were made for future improvements,
but one–ease of navigation–was sufficiently
important to address before the next phase of
review. Based on this concern, the structure of
the web site was considerably modified to
provide the user with a linear arrangement of
topics. This and other changes were completed
by the beginning of August 2000.
For a second external review, 13 people were
asked to participate. One of the 13 did not
respond to the invitation, three declined because
of conflicting commitments, but two
recommended other choices for reviewers.
Ultimately, of nine who agreed to review the site,
242
––––––––––––––––––––––––––––––––––– Kalichman et al., Resources for Instruction in the Responsible Conduct of Research
QUESTIONS
1. CONTENT
A. How would you rate the choices of topics covered?
B. How would you rate the quality of information provided?
2. NAVIGATION
How would you rate the ease for navigating within the Web site?
3. APPEARANCE
How would you rate the appearance of the Web site?
4. OVERALL
A. How would you rate the likelihood you would recommend this
resource for someone developing a new training program?
B. How would you rate the likelihood you would recommend this
resource for someone improving an existing program?
AVERAGE
MEDIAN
5.0
4.6
5.0
5.0
4.2
5.0
3.8
4.0
4.6
5.0
4.8
5.0
Table 1. Second phase of external review (6 reviewers). Using a scale of 1-5 (1 = very low, 5 = very high), the reviewers
answered the following six questions.
three failed to meet the deadline. The six
reviewers who responded were from two public
universities, one private university, two
government agencies, and one non-governmental
organization.
A summary of the average and median of the
second phase of reviewer evaluations is provided
in Table 1. The reviewers were extremely
positive about the content of the web site
(averages of 4.6 to 5.0). Compared to the
previous round of review, these reviewers were
also more positive about navigation (4.2 vs. 3.253.75). Although considered acceptable, no
reviewer scored appearance of the web site as a
5. In addition, the reviewers offered many
practical suggestions for improvements in
content, navigation, and appearance.
A third external review was begun in
September of 2000. A total of 48 people were
asked to review the web site by early October; 31
responded that they had the time and would be
willing to do so. Of those, reviews were
completed by 23 reviewers (16 public
institutions, 4 private institutions, 2 government
agencies, 1 Canadian government agency).
A summary of the average and median of the
third phase of reviewer evaluations is provided in
Table 2. Evaluation rankings were generally in
the range of 4 to 5. Lowest scores were for the
appearance of the web site (average=3.8) and
highest scores were for the likelihood that the
reviewers would recommend this web site as a
resource for someone developing a new training
program (average=4.8). The reviewers were
again generally positive, but several made
QUESTIONS
1. CONTENT
A. How would you rate the choices of topics covered?
B. How would you rate the quality of information provided?
2. NAVIGATION
How would you rate the ease for navigating within the Web site?
3. APPEARANCE
How would you rate the appearance of the Web site?
4. OVERALL
A. How would you rate the likelihood you would recommend this
resource for someone developing a new training program?
B. How would you rate the likelihood you would recommend this
resource for someone improving an existing program?
AVERAGE
MEDIAN
4.4
4.1
4.0
4.0
4.0
4.0
3.8
4.0
4.8
5.0
4.5
5.0
Table 2. Third phase of external review (23 reviewers). Using a scale of 1-5 (1 = very low, 5 = very high), the reviewers
answered the following six questions.
243
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
excellent suggestions for changes in structure and
content to the site. Few comments were repeated
across reviewers. Major areas of criticism
included:
1. Content: One reviewer was looking for a
pre-packaged RCR course along the lines
of the web-based tutorials for training
researchers working with human subjects.
As this reviewer observed, this web site
does not provide such a course.
2. Format: The most frequently voiced
concern was that the background (a grid
similar to a lab notebook page) was distracting.
3. Audience: It wasn’t clear to some reviewers
who the audience (i.e., instructors of RCR
courses) was for this web site.
4. Structure: Several reviewers failed to find
key elements of the web site (e.g., the
examples of courses) and some pointed out
confusion about the structure of some of
the sections (esp. resources and cases).
Related to this problem, a couple of the
links did not work, or did not work as
expected.
Several of the reviewers were quite supportive of
the web site, for example:
“The choice of topics to be covered in teaching
research ethics is excellent. I particularly think
it is useful that ‘minimal instruction’ is defined
for each of the topics and that more advanced
versions of the units are also suggested. This
will be quite helpful to faculty who are just
beginning to teach RCR, and who want to know
what is the minimum level of instruction they
need to meet.”
“I think the site looks great. It is very well
organized. It will be especially useful for
newcomers.”
“Best collection of materials related to RCR I
have found. The logical progression of steps
should make it easy to develop or improve
courses without becoming overwhelmed by the
task at hand. Linked pages were relevant and
provide materials for inspiration and contrast.”
“This is a very strong site and I learned a lot
just skimming. The links for case studies and
analysis of instructional delivery options were
quite good.”
“This is a great program. I think its strongest
feature is the way it brings together a wealth of
material in a useful and usable form.”
Based on the reviewer comments, further
significant changes were made to the structure of
the web site. As of its release, the structure of the
web site was designed around five topic areas:
Goals (Goals for RCR instruction), Content
(Suggested RCR topics: Descriptions and reading
lists), Format (Formats for RCR instruction:
Descriptions and examples), Tools (Tools for
RCR instructors: Texts, cases, and contacts), and
Evaluation (Evaluation of RCR instruction:
Overview and examples). After checking that the
structure of the web site was consistent and that
all links were active and accurate, the web site
was released for public use on November 1,
2000.
Discussion
As proposed, a new web site was developed to
facilitate access to resources for instruction in the
responsible conduct of research. With the
support of constructive comments from external
reviewers, an initial version of the web site was
made available to the research community
beginning on November 1, 2000. Based on
reviewer comments, this web site will be of value
both to those first developing programs of RCR
instruction and also to those seeking to improve
on existing programs of instruction.
To achieve the long-term goals for this web
site, it will be necessary for the site to evolve
both in terms of content and format. For this
purpose, the authors intend to solicit the latest
information about content and format of existing
RCR programs nationwide. Further, it will be
important to include mechanisms for ongoing
evaluation of the merits of the resources listed on
the web site and the web site itself. During this
next phase, the primary goals will be to survey
existing programs in RCR, solicit new materials
from these institutions, continue modifying the
framework of the site as needed to accommodate
the new resources, and implement mechanisms
for evaluating the effectiveness of the web site
and the resources listed.
Acknowledgments
The Office of Research Integrity (ORI) and
particularly Alicia Dustira, Lawrence Rhoades,
and Chris Pascal, all of ORI, are thanked for their
encouragement and support for the initial
244
––––––––––––––––––––––––––––––––––– Kalichman et al., Resources for Instruction in the Responsible Conduct of Research
development of this web site. The University of
California, San Diego is thanked for its support
of the Project Director (MWK), helping to make
this project possible. In addition, the generous
contributions of time and suggestions of the 33
external reviewers are gratefully acknowledged.
This work was supported by a contract from the
ORI to UCSD (contract #99T07035001D)
Bibliography
1. Bulger RE, Heitman E, Reiser SJ. The Ethical
Dimensions of the Biological Sciences. New York:
Cambridge University Press; 1993.
2. Elliott D, Stern JE. Research Ethics: A Reader. Hanover,
New Hampshire: University Press of New England;
1997.
3. Grinnell F. The Scientific Attitude. Second edition. New
York: Guilford Press; 1992.
4. Korenman SG, Shipp AC. Teaching the Responsible
Conduct of Research through a Case Study Approach: A
Handbook for Instructors. Washington, D.C.:
Association of American Medical Colleges; 1994.
5. Macrina FL. Scientific Integrity: An Introductory Text
with Cases. 2nd edition, Washington, D.C.: American
Society for Microbiology Press; 2000.
6. National Academy of Sciences, National Academy of
Engineering, Institute of Medicine. On Being a
Scientist: Responsible Conduct in Research.
Washington, D.C.: National Academy Press; 1995 [also
available online at: http://www.nap.edu/readingroom/
books/obas]
7. Penslar RL. Research Ethics: Cases & Materials.
Bloomington, Indiana: Indiana University Press; 1995.
8. Stern JE, Elliott D. The Ethics of Scientific Research: A
Guidebook for Course Development. Hanover, New
Hampshire: University Press of New England; 1997.
9. Center for Bioethics, University of Pennsylvania: http://
bioethics.net
10. Online Ethics Center for Engineering and Science: http:/
/onlineethics.org
11. Research Ethics Program, University of California, San
Diego: http://ethics.ucsd.edu
12. Scientific Integrity, a one-semester course, Virginia
Commonwealth University: http://www.vcu.edu/
courses/rcr
13. Teaching Ethics for Research, Scholarship, & Practice,
University of Minnesota: http://www.research.umn.edu/
ethics
14. Research Ethics, an Internet-based course, University of
Nebraska Medical Center: http://www.unmc.edu/ethics
15. Scientific Integrity, an Internet-based course, University
of California, San Diego: http://ethics.ucsd.edu/courses/
integrity
16. Survival Skills and Ethics, University of Pittsburgh:
http://www.edc.gsph.pitt.edu/survival
17. Teaching Research Ethics, Indiana University: http://
www.indiana.edu/~poynter/tre.html
245
An Interactive Web Site for Ethics Training: http://storiesandquestions.com
Rudolph J. Marcus, Stories and Questions, Occidental, CA, USA
Keywords: Case studies, Computer-based instruction, Ethics training instrument, Moral advisor, Self-directed
learning, Work environment
This paper reports on the construction and use of a web site for ethics training. The paper is divided
into three parts. Various uses of the web site as a delivery vehicle for ethics training are outlined in
the first part. Practical advantages of an ethics training instrument usable by individuals at their own
pace, place, and time are discussed in the second part. The web site itself and its operation are
described in the third part located after the references. The paper ends with suggestions for further
work in adding more seminars to the web site, further measuring the web site’s effectiveness, and
developing guidelines for facilitators.
Ethics Training With The Web Site
Types
The computer-based ethics training instrument is conceived as a delivery
Codes
technique. What might such a training instrument contain?
After the apprentice model lost its effectiveness for ethics training it was
Case studies
replaced by a recital of the appropriate ethics codes (Table I). Discussion of case
Awareness
studies is used for ethics training that is more specialized and pointed toward
Table I: Types of
particular disciplines and tasks. Both training in ethics codes and in case studies
ethics training
can be delivered by this sequenced text-and-question technique. The present web
site adds a third category to the ethics training instrument, awareness training.
Experience with the web site has shown different results from bulk presentation and from
sequenced presentation. Bulk presentation,
where the whole story and all of the questions
Training
Message
are presented at one time, usually draws either
What are the limits?
no response or a “head trip” response. The
Codes
sequenced presentation of a part of the story at a
What can I get away with?
time or an exercise with the story, accompanied
by a single question, appears to encourage the
What are the limits?
Case studies
thoughtfulness and inner work that lead to real
What can I get away with?
attitude change. It is that experience that leads
What is going on here?
Awareness
What can I do about it?
What might be the right thing to do?
to the statement of the previous paragraph that
the necessary ethics training of the ethics codes
themselves, and of applicable case studies, can
be delivered by this sequenced text-and-question
Table II Messages from various types of ethics training
Corresponding author: Rudolph Marcus, 17010 Burl Lane, Occidental, CA 95465, 707-876-1959 (voice),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
technique.
The implication that trainees get from recital
of appropriate ethics codes or from discussion of
case studies is “What are the limits?” The
message conveyed by such training is “What can
I get away with?” (Table II).
By contrast, awareness deals with questions
such as “What is going on here?” and “What can
I do about it?” The message conveyed by
awareness training is “What might be the right
thing to do?”
The reader may sense that I find the “What
might be the right thing to do?” more significant
and necessary in ethics training because it goes
beyond the given guidelines. However,
knowledge of applicable ethics codes and of their
application is case studies is an essential and
equally necessary part of ethics training. I have
discussed the multiplicity of the overlapping
ethics codes that researchers are subject to and
some of the conflicts between those codes in an
encyclopedia chapter (1).
Results of ethics training by recital of
appropriate ethics codes usually are measured by
attendance at the training, or by recall of the
Training
Codes
Case studies
Awareness
Results reported as
Attendance or recall
Discussion of case
Attitude changes
Table III: Results from various types of ethics
training reported as
contents (Table III).
Results of ethics training with case studies
are measured by qualities of participants’
discussion of the cases. Results of awareness
training can be assessed by noting attitude
changes. The same question is asked near the
beginning and near the end of each seminar. In
experience so far with the web site, and in the
talks and seminars on which the web site is
modeled, the response after the second time the
question is asked is much better elaborated, is
more grounded in the story, and shows a better
grasp of how the story applies to the participant’s
own life, problems, and actions. Changes
become apparent in the perceived self image of
the seminar participant. The changes are
reinforced by the facilitator calling attention to
them or asking whether the participant has
noticed the changes.
Practical Advantages Of Web Site Ethics
Training
Self-directed learning: No travel or time off
from work
The web site is an alternative learning mode for
ethics training which permits study and testing at
each individual’s time, place, and pace. It
reduces or eliminates trying to get everyone
together at a given time and place, with resultant
schedule changes and resistance to “a waste of
time.”
“Valuing of self-directed learning” (2)
applies to academic institutions and to needs for
ethics training there as well as in industrial and
governmental laboratories. Gunzburger writes.
“The survey results indicate that most schools
have not established systems for the valuing of
curriculum time that include balancing values for
lecture time and values for self-directed
learning.”
One surprise of experience with the web site
was the amount of motivation which is
generated by the sequenced self-study. No
coaxing or arguments by the facilitator were
needed to get clients to complete a seminar. If
there was no response to a session within two
weeks, re-presentation of the segment was
sufficient to elicit a response, often with an
explanation of what the holdup had been.
Engendering an ethical environment in self
and in work place
An unspoken assumption seems to be that
infractions of an ethics code are deliberate; if not
deliberate, then the incident was an “accident”
(3). Based on that assumption, an infraction can
be stopped and the perpetrator exposed and/or
punished.
Inherent conflicts and inconsistencies
between the different kinds of ethics codes, as
well as in individual codes themselves, are
discussed in the encyclopedia article previously
referred to (1). There are few, if any, places
where those caught in a conflict between
different kinds of codes can get help and advice.
The encyclopedia article concludes with a
training program that I have designed for
situations like that, and this training program has
now been put on the web site described in this
paper. In brief, the seminar participant who uses
the web site ethics training learns where to look
for the advice needed to cope with overlapping
248
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Marcus, An Interactive Web Site for Ethics Training
and conflicting ethics codes
A different way of stating the problem is
found in a letter to a college student who asked
for advice on ethical behavior In the letter, C. G.
Jung describes codes of moral values as “general
and not specific,” and therefore “they don’t
exactly apply to individual situations.” Jung’s
advice to the student is “try to live as
consciously, as conscientiously, and as
completely as possible (italics are Jung’s) and
learn who you are and who or what it is that
ultimately decides” (4).
The problem also has been stated in a law
review article about ethics in government, which
applies equally well to ethics in science (5):
“Our current obsession with investigating and
prosecuting individual wrongdoing may actually
prove counterproductive in our efforts to promote
ethical [science], promoting instead public
cynicism about [science]. To counteract this
emphasis on individual wrongdoing, we need to
pay more attention to ensuring that [science]
institutions are designed to engender an ethical
environment.”
All three ways of stating the problem
contradict the beginning assumptin—the old
vocabulary—that infractions of ethics codes are
deliberate or an “accident.” All three statements
indicate that infractions of ethics codes are NOT
always deliberate and that the perpetrators may
not even be aware of their inadvertent and often
avoidable errors affecting research integrity.
Jung’s advice to “learn who you are” to
behave ethically is exactly the aim of the training
program that I described in the encyclopedia
article. The training is to live with the opposites
in resolving conflicts of various ethics codes
within which researchers have to work (1, 6). It
is that kind of training program that I have now
brought out of the lecture and workshop stage
and put into the web site for one-on-one work
with an experienced ethics consultant. It is a
self-contained course that meets a previously
unrecognized need.
Present Status And Future Work
At this time the web site contains four such
seminars. One of the seminars deals with
collegiality and civility in the work place or, to
see it from the other side, conflict resolution in
highly polarized situations (7). Another deals
with how to find a moral advisor in a
hierarchically structured work environment (8).
Both describe work environments that often lead
to alleged scientific misconduct and how to deal
with them creatively.
Material is at hand for expansion of the web
site to about 12 seminars during the coming year.
The immediate next additions to the web site will
be four seminars dealing with the origins of
science (9). Together, they show four successive
stages of scientists working “consciously,
conscientiously and as fully as possible.” I have
used that material about twenty times in national
tour lectures for the American Chemical Society
under the title of “Nature and Gods, Science and
Scientists.”
Further research aspects of this work consist
of:
1. Adding more seminars to the web site.
2. Assessing its effectiveness. Effectiveness
can be gauged by looking at changed opinions, feelings, or assessments of problem
situations by seminar participants as the
seminar progresses. Records of such
changes are already being kept while maintaining seminar participants’ analytical
confidentiality, which is a hallmark of the
seminars in workshop and web site modes.
3. Developing guidelines for facilitators. As
more people use this method of self-study
for ethics training, they too may want to
become facilitators and learn more by
helping others to start self-study in scientific
ethics.
Bibliography
1. Marcus RJ. Government funding of research. In:
Encyclopedia of Applied Ethics. San Diego: Academic
Press; 1998. Vol. 2, p. 471-80.
2. Gunzburger LK. U. S. Medical Schools’ Valuing of
Curriculum Time: Self-Directed Learning versus
Lectures. Academic Medicine 1993; 68: 700-702.
3. Marcus RJ. Ethical considerations. The New Yorker
1996 March 11; p. 14.
4. Jung CG. Letter to William Kinney (1956 May 26). In:
Letters. Princeton NJ: Princeton University Press;
1973. Vol. 2, p. 300-1. .
5. Clark K. Toward More Ethical Government: An
Inspector General for the White House. Mercer Law
Review 1998; 49: 553-64.
6. Marcus RJ. Ethics in research and government funding.
Chemtech 1997 Nov, p. 8-13.
7. Marcus RJ. Ethics workshops when collegiality fails.
In: Abstracts of Annual Meeting, Association for
Practical and Professional Ethics. 1998; Paper V H.
249
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
8. Marcus RJ. Sequenced self-study for constellating a
moral advisor. In: Abstracts ofJoint Services
Conference on Professional Ethics (JSCOPE 2000).
9. Marcus RJ. Mythological stories as case studies for
teaching scientific ethics. In: Abstracts of 1996 Annual
Meeting, Association for Practical and Professional
Ethics. 1996; Paper II H,
The Web Site
Introduction
Welcome to this web site, the home of “Stories
and Questions,” a personal journey in self
enrichment. Here is a method of exploring who
one is by reading short stories and responding in
the moment to simple questions. These stories
allow one to stop and feel, and the question
permits feelings about one’s life and its direction.
Stories and Questions is a series of individual
seminars facilitated by Rudy Marcus. Rudy has
done stories and questions for 16 years (and if
one counts his research experience in the
sciences, for 50 years) and has experienced for
himself their ability to effect personal change.
On this web site, you can start your own
journey of exploration. Please read the
Introduction and follow its suggestion for “How
Do I Start?” Feel free to contact me at:
[email protected]
This is an introduction to seminars designed
for self-study. Each session or envelope contains
a story or direction for an exercise with the
material of the story. You, the participant in this
seminar, encounter the story or do the exercise,
and then respond to one or more question(s) on
this web site. The response can be in writing or
any other form of expression, and can be
telephoned or sent by e-mail or post to:
[email protected] Rudy will
then send you the next session (e-mail) or
envelope (paper) of the self-study seminar. The
method is adapted for self-study from group
workshops using different stories and questions.
Method
If you were using this material at a group
workshop or seminar, you would be sitting in a
circle. Each person in the circle would hear from
a facilitator what is on the web site as Session I,
or what is in envelope I in the paper version — a
short story, and then a question to which each
member of the circle responds. It is not a
discussion group, and there is not a consensus to
be reached. Rather, each response is respected as
that person’s truth at that particular time and
place. In such a workshop, there would be a long
break after the discussion of the material in
Session I. That break might even take the form of
lunch, a nap, a walk in the woods, and/or a swim.
More thoughts about the story, and additional
responses to the questions occur, and those might
be written in a journal or one’s workshop notes.
Stories
In a group seminar using this material, the
facilitator would have warned participants NOT
to identify with any of the characters in the story.
That is important and it applies as well to the
self-study.
The seminar participant encounters the story
as if the participant were seeing it on a stage. The
participant is not on the stage with the story
characters. The participant is in the audience
watching the actions of all the characters, being
privy to the knowledge, habits, and actions of all
the characters at that point in the story.
In the language of psychology, the
participant brings one’s ego to the story, one’s
own awareness, rather than identifying with,
taking the part of, one or the other character in
the story The more cross-cultural the story is—
for example, all cultures are likely to have
creation stories, and stories about the origin of
science—the more universally valid or typical do
those characters seem, and the easier it is for the
hearer of the story to say, “Hey, that character IS
me, and that is MY story.” Try NOT to do that.
The comparison of story with stage is quite
apt because as action on stage involves feelings
and emotions of onlookers, so encounter with
story can activate an individual participant’s
inner knowledge and experience analogous to the
story character(s)’ knowledge and experience.
That can happen whether or not the participant
had previously been aware of any feeling or
actions corresponding to those of one or more
characters in the story. A shorthand phrase for
such activation is that one or another of those
story characters is constellated in a participant by
the participant’s work with story and questions.
Another way of saying this is that no one
character in the story is or describes the whole of
me, but it often describes a part of me. I may
not have been aware of that part of me prior to
my work with that story.
250
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Marcus, An Interactive Web Site for Ethics Training
Questions
If you encountered this material in a group
workshop, the facilitator would ask one or more
questions after each story. Responses to the
question are addressed not to the facilitator, and
not to other members of the group but to the
center of the circle. The story is “myth,” a
technical term that has been defined by Joseph
Campbell as something that happened and
continues to happen to many people at many
times and places, in contrast to history which
happened at one time and place to one person or
a group of persons (and, a scientist would add, in
contrast to science which is something that
happens and is replicable under specified
conditions). As a myth, the story is considered to
bubble out of the center of the group circle rather
than from the facilitator. Similarly, the question
asked is considered to be coming from the center
of the circle, and therefore responses are
addressed to the center of the circle rather than to
a person at the perimeter of the circle or to the
facilitator. That should minimize any onus on, or
defensiveness by, a respondent about one’s
response. Similarly, others in the circle are
expected to respect any response since it is
addressed to the center and not to them.
Responses are not to be discussed or argued with,
particularly not by the facilitator.
The discussion circle consists of center and
perimeter. Without participation from the
perimeter, in the form of responses to questions,
the circle and therefore the seminar would not
exist. It is important for each participant to hear
oneself verbalize a response, even if that
response sounds similar to one that has already
been given.
The questions are designed to evoke choice
and feelings rather than to test knowledge or
recall. Reasons for, or explanations of, choices
may be asked for. Respondents will be asked to
stick to the subject matter of a question because
one of the easiest ways of escaping a choice or a
feeling is to talk about something else. Similarly,
the question asked at a session is about the story
told in that particular session, not about the end
of the story or about another story. Each
question is a universe of discourse, embedded in
the universe of discourse of the story of that
particular session. (By definition, a universe of
discourse is a separate world surrounded by an
impermeable barrier.) A participant is free to
state different choices, feelings, and opinions at
any time. Such changes are seen as fresh insights
rather than as error or shame for earlier choices.
“I’ve heard that story before”
Of course you have heard that story before. It is
part of our cultural heritage. It is a myth. It is
universal truth. You and I have that story in our
bones.
But in the group workshop I am not listening
to, and in the self-study I am not reading, that
story to find out how it ends, “who done it,” or
what the facts are. I am called on to take a stand
vis-a-vis that story at this particular moment. I
am listening to it, or reading it, to see whether I
still have the same answers to questions evoked
by the story or to see whether the story evokes
new questions.
Bruno Bettelheim has noted that a fairy tale
asked for most frequently by a child most likely
describes what that child feels to be its most vital
problem at that stage of its life. (For example,
the Grimm Brothers’ fairy tale that gripped my
emotions for many decades was “The Goose
Girl.” Like that protagonist, I also lost home,
country, and native tongue as a pre-pubescent.)
Ritual has been described as a process of
activating a myth. In that sense repeated
exposure to a story—by rereading it, by noting
whether responses to story and questions change,
and by asking new questions about it—
reconstellates powerful, often numinous,
characters within myself.
Instructions
You are not at a group workshop now, so you
have an opportunity to create your own pace and
place. Find a private space and time for 45
minutes, turn off the telephone, put out the cat,
and open one of the stories. Read the story,
consider it, and respond to the question(s) in
writing or other art form. Send your response to:
[email protected] Stay with
the story for a day or more—preferably a week.
Look at the story, questions, and your responses
occasionally, and write down any additional
thoughts. Note any additional insights.
Your response will be acknowledged and the
next Session will be sent to you. Repeat the
process with the next Session. Continue in that
manner until the final Session.
Because this kind of work is an ongoing
process and new insights keep popping up, it is
well to keep the Session materials, your
251
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
responses, and the facilitator’s comments in a
notebook. You will find that collection a growing
resource as new insights arise. You will also find
that it becomes a valued friend and adviser in
dark times.
Which Story?
Four stories are available as Sequenced Self
Studies at this time. They are:
• Cracked Pot (11 sessions)
• Moses Mendelssohn’s Dream (5 sessions)
• Rainmaker (5 sessions)
• Becket (8 sessions)
Any of those Sequenced Self-Studies is worth
doing in its own right in the same way that one
goes to a movie or takes a trip for adventure,
enjoyment, or enrichment.
Just as movies or trips also may be taken
with specific purposes in mind, such as
information or education, these stories can be
used for specific purposes as well as in their own
right. For example, Cracked Pot and Moses
Mendelssohn’s Dream have been used for
working with self-worth problems. Rainmaker
has been useful for conflict resolution in highly
polarized situations. Becket is a good practicum
for finding moral advisers in hierarchically
structured organizations. Both Rainmaker and
Becket are excellent self-studies for ethics
training.
How Do I Start?
On the following pages [of the web site] you
will find the first Session of each of the available
self-studies. Choose one, follow the instructions,
and send your response to:
[email protected]
Rudy will then comment on your response
and activate the next session of your self-study.
252
III. Research Theory and Methods
8. Detection Methods
The Misuse of Statistics: Concepts, Tools, and a Research Agenda
John S. Gardenier, National Center for Health Statistics, Centers for Disease Control and
Prevention, USA
David B. Resnik, Brody School of Medicine, East Carolina University, USA
Key Words. Competence, Ethics, Journals, Research misconduct, Statistics
While it is widely recognized that the proper use of statistics is a key element of research integrity,
there has been considerable debate about how to understand or respond to the misuse of statistics in
research. To understand what is meant by “misusing statistics,” it is important to describe the role of
statistics in the scientific method and relate the concept of “misuse” to other ethical concepts, such as
“misconduct” or “incompetence” or “negligence.” We believe that some misuses of statistics can be
considered misconduct, although most misuses should be viewed as negligence or deficits of
competence.
Statistical methods, theory, techniques, and models play an important role in several stages of the
scientific method, but we will focus here on just two stages (See Figure 1). First, statistics is essential
to good experimental design as in randomized clinical trials, for example. In order to obtain a
rigorous test of a hypothesis, it important to obtain data that can provide evidence for or against the
hypothesis. If the hypothesis is a comparative or quantitative statement, such as “drug x is more
effective than drug y” or “less than five percent of patients suffer serious side effects from drug x,”
then the conclusions must be based on statistically significant results. For example, an experiment
that compares the effects of two drugs on only ten patients is very unlikely to produce statistically
significant results. If some or all of those patients are subjected to health risks in the experiment, this
creates two additional ethical problems. First, it is unethical to expose a human subject to an
unnecessary experimental risk, unless the potential benefits (to the individual or to society) of
exposure to the risk outweigh the potential harms. If the experiment is not well designed such that no
meaningful conclusions can be drawn, then the potential benefits will not outweigh the potential
harms. Second, when patients give informed consent to participate in research, they usually believe
that the research is valuable and may advance science. Encouraging or even allowing subjects to
participate in an experiment that is highly unlikely to yield valid results is implicitly deceptive. It is
important to address the statistical issues before conducting experiments or tests, because once one
has gathered and recorded data, it may be too late to correct statistical (or ethical) flaws in the design
of the experiment (1). The expression “garbage in, garbage out” applies here.
Second, statistics is important in an analyzing and interpreting data. There are many different
statistical tools that one may use to analyze data, ranging from simple procedures, such as t-tests and
Corresponding author: John S. Gardenier, CDC/NCHS, 6525 Belcrest Road, Hyattsville, MD 20782, 301-458-4229
(voice), 301-458-4021 (fax), [email protected]
* Disclaimer: This paper represents the personal views of the authors. It does not necessarily represent policies or positions
of their institutions. The research effort reported here was not supported by grant funds.
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Figure 1: The Role of Statistics in the Scientific Method
Define problems,
questions, and
research aims
Review the literature
Develop a hypothesis
Statistical methods,
measurement tools,
and models
Design experiments
or other tests
Collect and
record data
Revise or modify
protocol or hypothesis
Analyze and
interpret data
Disseminate results
Peer review
Public understanding
of research
linear regression, to more complex ones, such as
analysis of covariance and statistical modeling.
It is not our aim to discuss these methods here,
but we would like to point out that it is relatively
easy to misuse these methods. To apply any
statistical method correctly, one must have
information about the variables used (continuous
or discrete, gaussian or bimodal, etc.),
information about the sampling process used
(sample size, independence, randomness,
representativeness, etc.), and a sound
understanding of the theory and assumptions
underlying that method. If a researcher does not
use a method correctly, then conclusions may
overestimate or underestimate an important
relationship or effect. If we think of statistics as
a tool for distinguishing between random “noise”
Replication of results
Scientific impact
of research
in the data and the real signal, then someone who
incorrectly uses statistics may produce a result
that is distorted or even artificial. A person who
correctly uses statistics will amplify and clarify
the signal without distorting it (2).
With this understanding of the role of
statistics in research in mind, we can clarify what
we mean by “misuse” of statistics. Not all
misuses have equivalent ethical implications, as
we discuss later. A “misuse,” for our purposes, is
an incorrect use, i.e., a use of statistics that is not
appropriate, given the research question, the
experimental design, and the methods being
used. For example, it may be appropriate to
exclude outliers if there is credible evidence that
such points are not part of the statistical
population represented by the sample. It may
256
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gardenier & Resnik, The Misuse of Statistics
also be appropriate to use statistical methods to
fill in (or impute) missing data for the purposes
of statistical analysis. What’s the difference
between appropriate and inappropriate exclusion
of outliers or appropriate and inappropriate
imputation of data? Many books on statistical
methods discuss these topics, but from an ethical
viewpoint they boil down to the following: an
appropriate exclusion (or imputation) is one that
dampens the noise without altering the signal that
describes the relationship or effect.
Misuses of statistics can also occur in the
absence of erroneous or distorted results. Misuse
can also arise from a failure to provide the
research community with important information
about the methods used or the experimental
design. Researchers need to address such
statistical issues as excluding outliers, imputing
data, editing data, “cleaning” data, or “mining
data.”2 These practices are often practical, or
even necessary, but it is important to discuss
them honestly and openly when reporting
research results (3).
Thus, there are two types of misuses in
statistics: (1) using statistical methods,
techniques, or models in ways that produce
distorted or artificial results; (2) failing to
disclose important information about statistical
methodology to researchers. Misuses of statistics
may (or may not) violate several ethical
obligations, such as the duty to be honest, the
duty to be objective, the duty to avoid error, and
possibly the duty to be open (4). There has been
considerable debate about whether “misuse of
statistical methods” should be classified as
misconduct (5). The federal government and the
scientific community have moved toward a
narrow definition of misconduct that focuses on
fabrication, falsification, and plagiarism (6, 7).
The new federal policy implies that the misuse of
statistics could be classified as a form of
misconduct when it involves intentional
deception. Some misuses could be classified as
“fabrication” if they involve making up data or
results, or “falsification” if they involve
manipulating, changing, or omitting data or
results. Misuses of statistics that do not involve
intentional deception could be viewed as honest
error, incompetence, bias, or “serious deviations”
from acceptable practice (8). A person who
makes excessive errors due to haste, ignorance,
or sloppiness may be considered to be negligent
or lacking the needed degree of competence,
statistical or otherwise (9). Professionalism
requires adequate application of both statistical
and subject matter expertise to analyses. There
might be varying degrees of culpability in a
failure to meet this criterion. Clearly, honest
error is never misconduct. Neither is it
misconduct when two or more well qualified
statisticians or other researchers disagree about
technical issues in a given research protocol.
Still, some misuses of statistics in research do fit
the definition of misconduct used by the federal
government. That may be hard to establish by a
preponderance of the evidence. When a person
changes or fabricates data, one at least has some
kind of record that one can use to imply intent.
When a person manipulates analyses of data,
there may be no record to prove the manipulation
was deliberate or even culpably negligent. Thus,
as a purely practical matter, it may be very
difficult investigate or prosecute such cases (10).
The Importance of Correcting Statistical
Misuse
Statistics play vital roles in most aspects of
modern post-industrial societies. Although
statistics are sometimes dismissed as trivia or
fuzzy math, distrusted as biased, or directly
equated with lying, the truth is that they are
inescapably important (11). As noted in the
Preamble to the Ethical Guidelines for Statistical
Practice:
257
The professional performance of statistical
analyses is essential to many aspects of society.
The use of statistics in medical diagnoses and
biomedical research may affect whether
individuals live or die, whether their health is
protected or jeopardized, and whether medical
science advances or gets sidetracked. Life,
death, and health, as well as efficiency, may be
at stake in statistical analyses of occupational,
environmental, or transportation safety. Early
detection and control of new or recurrent
infectious diseases depend on sound
epidemiological statistics. Mental and social
health may be at stake in psychological and
sociological applications of statistical analysis.
Effective functioning of the economy
depends on the availability of reliable, timely,
and properly interpreted economic data. The
profitability of individual firms depends in part
on their quality control and their market
research, both of which should rely on
statistical methods. Agricultural productivity
benefits greatly from statistically sound
applications to research and output reporting.
Governmental policy decisions regarding
public health, criminal justice, social equity,
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
education, the environment, the siting of critical
facilities, and other matters depend in part on
sound statistics.
Scientific and engineering research in all
disciplines requires the careful design and
analysis of experiments and observations. To
the extent that uncertainty and measurement
error are involved – as they are in most research
– research design, data quality management,
analysis, and interpretation are all crucially
dependent on statistical concepts and methods.
Even in theory, much of science and
engineering involves natural variability.
Variability, whether great or small, must be
carefully examined both for random error and
for possible researcher bias or wishful thinking.
...
Because society depends on sound
statistical practice, all practitioners of statistics,
whatever their training and occupation, have
social obligations to perform their work in a
professional, competent, and ethical manner.
(12)
If researchers are careless or deceptive in their
use of statistics, harms and costs to society will
result. Poor statistics in science leads to poor
science. The research record can be corrupted or
polluted, wasting the time and energy of other
researchers. At the very least, research funds lost
in bad research represent an opportunity cost in
that those funds could have been allocated to
more deserving projects.
For all of these reasons, it is important
that scientists and science administrators pay
careful attention to the quality of statistics in
science as funded, performed, and reported in
their areas of jurisdiction and of competence.
Good statistical work should be defended when it
is attacked inappropriately. Bad statistical work
should be detected and corrected as appropriate.
What are the Contributing Factors to
Misuse?
There is not a great deal of evidence that has a
direct bearing on the misuse of statistics in
research. However, if one assumes that many of
the factors that contribute to other ethical
problems in research, such as misconduct,
probably also play a role in the misuse of
statistics, then one could cite the following
factors, i.e., the “usual suspects” (13, 14).
• Pressures to publish, produce results, or
obtain grants
• Career ambitions or aspirations
• Conflicts of interest and economic motives
• Inadequate supervision, education, or training
We believe that all of these factors probably play
a role in misuses of statistics, but our conclusions
are merely speculative. More research is needed
on this topic. However, we would like to discuss
two other possible factors in the misuse of
statistics that are not on the above list of “usual
suspects.”
First, there are now many computer
programs that analyze data. These programs are
very user-friendly; all you need to do is load your
data set and choose your statistical test in order to
get results. One may even run several different
tests in an attempt to increase the significance
level (or p-value), although this can invalidate
the testing. While these programs save a great
deal of time and effort, they may contribute to
statistical misuse in that it is possible to plug
some numbers into one of these programs
without knowing how the analysis works, or why
a certain test would (or would not) be an
appropriate test. We think this problem has a
fairly obvious solution: teach more statistics in
research. If students and researchers understand
how to use statistics properly, then they should
have fewer problems using statistical computer
programs. Indeed, we believe that education is
the key to improving statistical practice.
Second, it has become standard practice in
some areas of research to only publish results
that have a p-value of 0.05 or less. The best
journals use more comprehensive criteria
enforced by competent statistical peer review.
We here address only those journals that place
excessive reliance on the p-value. The value of
0.05 is an arbitrarily chosen number; there is no
sound statistical or philosophical reason why a pvalue of 0.06 is fundamentally different from a pvalue of 0.05. However, under pressure to
publish, researchers may decide to massage or
manipulate data in order to obtain “significant”
results. Furthermore, there is now a growing
body of literature on publication bias in research
(15-17). Publication bias occurs when there are
discrepancies between the published research
record and the complete research record. The
discrepancies occur because journals tend to
publish only “significant” results. There are
some good potential solutions to the p-value
problem. First, researchers should realize that pvalues are merely conventional, not sacrosanct.
258
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gardenier & Resnik, The Misuse of Statistics
Second, they are also often sensitive to various
theoretical assumptions and may give erroneous
results due to mere artifacts of a data sample.
Third, not all statistical computer packages
compute all tests correctly. Fourth, journals
should be willing to publish results that are
substantial contributions to the literature of the
field, not just those that appear to have met a
conventional statistical test. The test result
reported may not be correct, and even a correct
conclusion that a certain hypothesis was not
statistically supported by data from a welldesigned study may be useful in limiting future
fruitless research by others. Finally, researchers
and research organizations should create
databases for unpublished data of archival value
and make those data publicly available (18).
Statistical Ethics, a Powerful Tool for
Research Integrity
Statistical ethics is a relatively recent
development. The seminal work, by W. Edwards
Deming, was first published in 1965 (19). The
American Statistical Association developed a
series of statistical ethics codes or guidelines
starting in 1979. Their current official Ethical
Guidelines for Statistical Practice was
promulgated in 1999 (12). The International
Statistical Institute instituted its Declaration on
Professional Ethics in 1985 (20). The United
Nations has published Fundamental Principles of
Official Statistics in the early 1990s, the current
official version being dated 1994 (21).
The pattern that emerges from this brief
history is that initial efforts to approach the issue
tend to be optimistically simple. Corrections
over time add to the scope and complexity of the
documents. The most recent document breaks
out areas of ethical responsibility for all people
using statistical methods professionally (12). It
covers separately, for example, responsibilities in
publications and testimony, responsibilities to
funders or employers, to research subjects, to
research team colleagues, and responsibilities
regarding allegations of misconduct. Beyond
addressing responsibilities of the individuals,
moreover; it also addresses the responsibility of
those employing practitioners of statistical
methods to provide a suitable moral climate for
that work.
Such statistical ethics documents become
tools for research integrity when they are integral
to actual practice. For example, if a federal
research funding agency were to adopt a policy
of stating in grant announcements that all grant
proposals received for projects employing
statistical methods would be expected to be
performed in accordance with the Ethical
Guidelines for Statistical Practice, that would put
real moral pressure on both proposers and
grantees to avoid misuse of statistics. If journal
editors were to state in notices to authors that any
papers containing statistical methods submitted
to that journal would be implicitly subject to
those same guidelines, some of the authors would
be more highly motivated to avoid misuse of
statistics.
If all scientists and engineers who are
competent in statistical methods would note
published examples of misuse of statistics and
report those to the funding agencies or journal
editors involved, then the recipients would
become more motivated to enforce sound
statistical practice. In short, we should not let
ethics documents sit unused on shelves or in
unvisited cyberspace. Ethical considerations
have practical consequences for good or evil.
The failure of good people to use them
effectively contributes to the likelihood that other
people may perpetuate statistical misuse either
through intent to deceive or simply through
deficits of statistical competence.
A Proposed Research Agenda
While we believe that there are still many
important conceptual and theoretical issues
relating to the use/misuse of statistics in research,
it should be clear from this brief discussion that
more empirical research is required on the
incidence of statistical misuse, its causes and
effects¸ and on the efficacy of using ethics
education and ethics documents as tools for
improvement. The following are some of the
empirical research questions we think are
important to study:
1. How many (or what percentage of) published studies make statistical mistakes?
2. How many allegations of research misconduct involve misuses of statistics?
3. How many researchers believe that the
misuse of statistics is an important ethical
issue in research?
4. Do different fields have different statistical
practices or take different approaches to the
misuse of statistics?
259
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
5. What is the incidence of publication bias in
various fields?
6. What do researchers and students know
about statistics?
7. Where, when, and how do students learn
about misuses of statistics in research or
other ethical issues in statistics?
8. How often do researchers use statisticians
or other statistical consultants?
9. Are editors and reviewers able to catch
statistical misuses?
10. Can data audits detect misuses of statistics?
11. Do research ethics codes or policies
address misuses of statistics?
12. When ethics education or ethics documents
are used as tools to improve research
integrity, how effective are they at promoting the proper use of statistics?
13. How often do institutional review boards
(IRBs) discuss statistical issues in human
subjects research? Do IRBs use statisticians?
14. How do misuses of statistics affect the
public? Do such misuses ever cause harm
to the public or threaten public health or
safety?
15. How often do statistical issues arise in
public policy debates?
16. What does the public know (or not know)
about statistics?
17. How do lay people interpret important
statistical concepts, such as “probability,”
and “risk”?
7. Office of Science and Technology Policy. Federal policy
on research misconduct. http://www.ostp.gov/html/
001207_3.html , 2000.
8. Panel on Scientific Responsibility and the Conduct of
Research. Responsible Science: Ensuring the Integrity
of the Research Process. Washington: National
Academy Press, 1992.
9. Grinnell F. The Scientific Attitude. New York: The
Guilford Press, 1992.
10. Dahlberg, J. Personal communication. Office of
Research Integrity, Rockville, MD. 20 November 2000.
11. Huff D. How to Lie with Statistics. New York: WW
Norton, 1954.
12. American Statistical Association. Ethical Guidelines for
Statistical Practice. Alexandria, VA: American
Statistical Association, 1999. Available at http://
amstat.org/profession/ethicalstatistics.html.
13. Macrina, F. Scientific Integrity. Washington: American
Society for Microbiology Press, 2000.
14. Commission on Research Integrity. Integrity and
Misconduct in Research. Washington: Public Health
Service, 1995.
15. Ioannidis J. Effect of the statistical significance of
results on the time to completion and publication of
randomized efficacy trials. JAMA 1998; 79: 281-86.
16. Stern J and Simes R. 1997. Publication bias: evidence
of delayed publication in a cohort study of clinical
research projects. BMJ 1997; 315: 640-45.
17. Easterbrook P, et al. Publication bias in clinical
research. Lancet 1991; 337: 867-72
18. Rennie D. Fair conduct and fair reporting of clinical
trials. JAMA 1999; 282: 1766-68.
19. W. Edwards Deming. Principles of Professional
Statistical Practice. The Annals of Mathematical
Statistics 1965; 36: 1883-1900.
20. International Statistical Institute. Declaration on
Professional Ethics. 1985. Available at http://
www.cbs.nl/isi/ethics.htm.
21. United Nations. Fundamental Principles of Official
Statistics. 1994. Available at http://www.cbs.nl/isi/
fundamental.htm .
Bibliography
1. Johnson R. and Bhattacharyya G. Statistics: Principles
and Methods. New York: John Wiley and Sons, 1985.
2. DeMets D. Statistics and ethics in medical research.
Science and Engineering Ethics 1999; 5: 97-117.
3. Resnik D. Statistics, ethics, and research: an agenda for
education and reform. Accountability in Research 2000;
8: 163-88.
4. Bailar J. Science, statistics, deception. Annals of
Internal Medicine 1986; 104: 259-60.
5. Guston D. Changing explanatory frameworks in the
U.S. government’s Attempt to define research
misconduct. Science and Engineering Ethics 1999; 5:
137-54.
6. Office of Research Integrity. Federal policy on research
misconduct. Federal Register: October 14, 1999; 64,
198: 55722-55725.
260
Images as Evidence: Forensic Examination of Scientific Images1
John W. Krueger, Division of Investigative Oversight, Office of Research Integrity, DHHS,
USA
Keywords: Autoradiograms, Blots, Image processing, Manipulation and falsification, Scientific misconduct
A “questioned” scientific image, i.e., suspicions of falsification (or plagiarism) of image data, such as
photographs of PAGE gels, autoradiograms, and blots (Western, Northern, and Southern) can give rise
to an allegation of misconduct in science. Pursuing oversight review of institutional investigations
and reviewing allegations that ORI receives directly, ORI commonly examines the evidence through
image processing. Typically, the examination can extend beyond merely asking “what is the evidence
the image is/isn’t authentic?” and/or “are two contested images really the same?” Examples from
these cases illustrate the general principles in forensic image processing and several methods that ORI
has found useful in resolving the questions at hand. They provide an opportunity for further
instruction as to what constitutes data falsification in an image.
Design/Methods
Source of Material: The material for this presentation was taken from a survey of 19 ORI cases
that involved allegations of falsification or plagiarism of the images of gels, blots, auto-radiograms,
and micrographs. The cases span a period from 1990 to 2000. The number of such questioned image
allegations has generally increased, as has their incidence relative to other ORI cases. (Figure 9) A
compilation from this review is discussed below.
Software: Most of ORI’s image analysis was done on a Macintosh® computer. The reason is
both historical and practical; files transfer easily from the Windows® platform to the Macintosh®;
but the opposite is not always true.
ORI has found several different image processing programs that are readily available and well
documented so that the results can be easily shared with all parties in a potentially adversarial dispute.
(1, 2) Each separately —or in combination with the others— offers distinct advantages. The image
processing was conducted using either NIH Image (3) and/or Adobe Photoshop® (4), both of which
were equipped with the Image Processing Tool Kit® (IPTK) plugins. (5) NIH Image, developed at
the National Institutes of Health, is in the public domain and is ideal for analytical treatment of 8 bit
(256 shades) monochromatic images. Photoshop is better suited for conducting overlay comparisons
of two images and for working with color, but it requires the IPTK’s plugins for analytical work.
Finally, ImageJ (6) is an update of the NIH public domain software that is compatible across
computer platforms and will process images at 8, 16, and 32 bit depth; thus, it can detect vastly fainter
features that might be hidden in the image.
Other Resources: Articles that can serve as guidance to issues involved in the forensic
examination of contested documents can be obtained on the Internet. (1, 2) Those sites can serve as
Corresponding author: John Krueger, Office of Research Integrity, 5515 Security Lane, Suite 700, Rockville, MD, 20852,
301-443-2263 (voice), 301-594-0039 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Figure 1. Original Western blot data. The results of an
electrophoretic mobility shift assay to show bands reflecting
the gene expression of separate mutant proteins. However,
the shape of the bands and the pattern of the background in
the 1st, 4th, and 5th lanes look alike.
links to find other material.
Reasons for Examination and Some
Principles of the Image Analysis Methods
The usual motivation for image analysis is to
examine the authenticity of a particular document
or to determine whether two purportedly
different images really were derived from the
same experiment.2 In fact, image analysis
provides information that addresses other issues.
For example, features can be detected that reveal
the source of the image, whether it is compatible
with laboratory records such as autoradiograms
or prior blots (see note 2), and whether the
questioned image existed on a computer as a file,
or on a website as a component of someone
else’s homepage. Second, the analysis of the
latter sources can provide dates of creation,
which can be corroborated with laboratory
records, etc. Third, image enhancement may
reveal evidence for the mechanics of the figure’s
construction, such as edges of photographic
prints and presence of “white-out” and may
uncover “hidden” details, such as erasures of
labels. Fourth, an analysis of the new facts
produced, such as the sources, dates, and
incidence of re-use, may establish whether a
pattern of misrepresentation existed that rules out
honest error. Examples from ORI’s cases
illustrate these points.
Figure 1 represents a photographic mock-up
of Western blot data, consisting of five
photographic strips, in which the 2nd to 4th lanes
were on one strip. Although purportedly
showing differentdeterminations of protein
created by separate mutant gene constructs, the
1st, 4th, and 5th lanes look unexpectedly similar,
but it is difficult to say for certain that they are
the same.
One generic principle in making comparisons
to determine the authenticity of data is to look at
the features that would otherwise be unnoteworthy, such as fine features hidden in the
background.3 There may be random features that
are hidden from our perception. The human eye,
which responds to contrast, can distinguish only
~50 shades of gray (7) or less (8), but it can
detect 100 shades of color (8).4 However, the
computer's response is not dependant on contrast;
it can selectively amplify very slight differences
in shade. The ability to detect such differences
can be affected by the “depth” used to digitize
the image, which in this case is 256 shades of
gray.5 The amplified differences in gray shades
can next be shadowed and assigned false-colors
to make faint differences even more visible, as
shown in Figure 2.
These steps reveal faint artifactual features
that were “hidden” in the background which are
common to three of the lanes. Thus the
respondent's claim, that at least two of the three
lanes (1, 4, or 5 in Figure 1) represented evidence
for gene expression of different mutant proteins,
was a clear falsification of data.
Enhancement of the small difference in
shades can also expose minute structural details
in the morphology of bands, which otherwise
would look smooth and featureless. Figure 3
illustrates a photo-montage from the above case;
the bands appear similar in the 1st and 5th lanes.
Contrast enhancement and false-coloring of
the above image as shown in Figure 4
demonstrate that the respective bands share
similar miniature features. Thus, the image
analysis showed that the first and the last lanes
were from the same experiment.
In both examples above, the falsification was
associated with false labeling of data that had
been “re-used” from another experiment. The
262
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Krueger, Images as Evidence
Figure 2 (left). Image enhancement of the questioned
Western blot data. This ORI image analysis figure shows
only the 1st, 4th, and 5th lanes from Figure 1. Contrast
enhancement of the monochromatic gray-scale image,
followed by shadowing and false-coloring (using NIH
Image), revealed small features in the background artifact
that are common to all three lanes (arrows) which the
respondent had falsely represented as different. Note that
in this case some differences can also be found, such as an
horizontal artifact under the top band in the 4th lane, but
they are in the background and represent artifacts that were
introduced at some later point.
Figure 3. Western blot data. The results purportedly found
a good agreement between the observed and the predicted
size of five mutant proteins. However, the 1st and the 5th
lanes’ bands look similar.
Figure 4. Image enhancement of the 67 kDa MW and 32 kDa MW bands from Figure 3. The bold lines denote
miniature features in the bands’ morphology that indicate both were actually the same data, which the respondent had
falsified in re-use.
263
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
second example showed an additional
falsification involving a false claim that the
molecular weights had been determined. In this
case, the intent to falsify the data is prima facie,
because the molecular weight could not have
been measured for the last re-used band. Finally,
because the molecular weights were purported to
approach the predicted values, the evidence also
indicates that the falsifications are significant.
These elements strengthen the findings.
Background detail and miniature features
cannot be examined by image enhancement in all
specimens. Fortunately, numerous other
approaches are available in image processing to
compare two questioned images. In general a
combination of methods is determinative. For
example, the morphology, size, and vertical
arrangement of the bands and the existence of
larger artifacts are the most obvious features to
compare. Moreover, the horizontal spacing
between the bands should not be overlooked;
because substances rarely migrate on gels
absolutely parallel, there may be slight
differences in lateral disposition that are also
significant. Some forms of artifact might reoccur, such as that introduced by a faulty film
dryer and/or the edge of a blot on an
autoradiographic film. The key question in cases
of “replicating” artifacts is whether a relationship
to other features should exist.
How to best visually represent the results of
an image overlay is always a challenge. A
visually effective and efficient method is to
overlap color-coded plots of the “contour” map
of the intensities in two separate blots, where the
areas of overlap generate a third color. If two
gray scale images are overlaid, the interpretation
of the origin of features in the overlay becomes
problematic unless each is first converted to a
suitably chosen monochrome color scheme.
Reconstruction of a Missing Document:
Analysis of an image can also be used to test the
proffered source of a questioned image under
circumstances in which the original raw data are
missing. Figure 5 represents a composite image,
which was created by combining a figure of a
questioned Northern blot in the mentor’s
manuscript with a figure of a different
experiment shown in the student’s thesis.
Unfortunately, the original blot and its
PhosphoImager computer file were missing, but
the mentor provided laboratory data purporting to
be a different representation of the same blot (an
ethidium bromide stain) that showed two groups
Figure 5. Overlay of the mentor’s Northern blot mRNA
data (small rectangle) with a figure from a different
experiment from the student’s thesis (tall rectangle). In this
ORI image analysis, the actual fit was determined
mathematically and showed the missing blot actually had at
least seven lanes, indicating the respondent’s claim was
false.
of six lanes, separate by an empty lane.
However, the overlay, shown in Figure 5, which
was established as the best mathematical fit
between the two sources, demonstrated that the
missing original blot had to have had at least
seven lanes. Thus, the proffered laboratory
records could not be evidence of the mentor’s
“missing” data.
Analysis of Poor Images: The poor quality of
an image is not necessarily a deterrent to the
application of the above tools to its examination.
The left side of Figure 6 shows a poor quality
photocopy of data that was submitted in a
mentor’s National Institutes of Health (NIH)
grant application, which purported to be a
Western blot of an immunologic protein, “P-48,”
using 125I-labeled human lymphocytes. The
figure on the right side of Figure 6 represents the
enhanced image of an autoradiogram from his
student’s experiments, which used 35S methionine labeling of cultured rats cells.
The distracting artifact due to overphotocopying could be minimized by image
processing. This step revealed additional bands
in the former with more certainty, and it more
264
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Krueger, Images as Evidence
Figure 6. Examination of a poor quality photocopy. The mentor submitted the left hand “125I-labeled” figure in an NIH
application. At right is shown the student’s 35S-labeled autoradiogram, in which the band pattern was shadow-enhanced
(green arrows). An artifactual lane distortion is denoted by the red arrows, which is weakly indicated in the photocopy.
clearly exposed a similar artifactual distortion of
the lanes, as shown in Figure 7. The mentor had
falsified the preparation, the experimental
conditions, and the molecular weights in the
photocopy that he had submitted to the NIH.
Recovery of Probative Details:
Examinations of images may even reveal new
evidence that bears upon other elements that are
required for a finding of scientific misconduct.
In another case, the allegation involved six
instances where different sets of autoradiograms
were allegedly falsely labeled and presented as
different experiments. The student claimed these
were honest errors, due, in part, to her
inadvertent use of unlabeled autoradiograms.
However, image enhancement by one of the
institution’s committee found evidence that the
original label on one autoradiogram had been
Figure 7. Computer enhancement of the bad photocopy
shown in Figure 6. In ORI’s image analysis, the distracting
artifact in the photocopy can be removed by filtering, while
false-coloring further enhanced the bands. The lane
distortion artifact, present in the student’s autoradiogram
(Figure 6) was apparent in the same relation to the bands in
the enhanced image, showing the student’s results were
falsified by the mentor to NIH.
265
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Figure 8. An example from one of six sets in which an autoradiogram had been falsely labeled and re-used. The
institution’s image analysis found evidence that the label for the prior experiment had been erased on the corner of the
autoradiogram. The visible ink is blue, while evidence for the enhanced erasures is shown in red. Originally barely visible
only as a faint and diffuse blue smear, the erased label differed from the film’s background by only one gray level out of 256.
The erasures were visualized here by ORI, after the film had been scanned at 42 bit resolution and the erasures had been
selected for enhancement using their hue. The erased “MyoG” and “MyoD” denoted experiments on chickens and not
mice. Thus, the re-use of the autoradiogram could not have been an honest error from mixup of unlabeled films, as the
respondent originally claimed.
erased, but not fully (Figure 8). Thus, image
processing revealed evidence that the
falsification was not an honest error. ORI’s
subsequent analysis of figures in publications
found that there was a pattern as to the six
instances of re-use that was not consistent with
their selection by chance.
A scientific image is
simply a picture constituting
evidence that a test was
carried out and/or that the test
produced a certain outcome.
In this context, the image is
construed as qualitative
“data.” It could also be the
basis for quantitative
measurements, i.e., by
measuring the “size” of a
substance, or as the raw data
for determine the amount of a
substance. Thus, one
consequence of discovering
the falsification in an image is
that there may be false claims
elsewhere in a paper.
Figure 9. Incidence of 19 ORI cases involving contested scientific images. The
data reflect when ORI’s case file was opened; this formal step can occur at any
phase in a case’s history (i.e., at the allegation assessment, inquiry, or investigation
stages). Thus the act of misconduct differs slightly in its date of occurrence. The
percentages indicate the fraction of all ORI cases opened in those years.
“Tamper” refers to allegations where the intensity of bands was selectively
altered. “High-Tech” indicates manipulation by a computer to modify the image.
266
Compilation of
Information from 19 ORI
image analysis cases
In all of the cases above, the
questioned image qualified as
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Krueger, Images as Evidence
Type of Misconduct Number of ORI Cases
Alleged
Falsely label as a
different experiment
(re-use)
Falsify molecular weight
marker positions
Cut, Graft, and Reuse,
alter lane positions to
fabricate data
Tampering: selective
manipulation of
image content,
enhance/subtract
bands, move position
Plagiarism of images
(from Internet or
journals), with false
claims of element(s)
from above
Image Source
Thesis (student)
13
Others (plagiarized)
Prior publication (self)
>13
Respondent
8
(5 students)
(3 mentors)
3
2
Status:
5
Senior Faculty
Junior Faculty
Fellows
Students
4
7
4
3
5
Allegation Source:
Student/Mentor/Co-Invest.
Reviewers
Inquiry Committee
Audiovisual Technician
Audience
3
Table 2. Characteristic of allegations of falsification of
images in 19 ORI cases.
Table 1. Falsification of data with images–compilation from
review of 19 ORI cases. This compilation indicates the
incidence as number of cases, which under- represents the
instances of misconduct, i.e., the actual number of figures or
publications involved. The impact of the acts in each case
was, in some cases, dramatically higher; one case involved
40 separate figures and multiple publications.
data, the intentional fabrication or falsification of
which is the key element of research misconduct.
On three occasions, a component of the
allegation also involved plagiarism. The
allegations reviewed by ORI generally involved
use of fairly low-tech mechanisms for
misrepresenting the data (Figure 9), such as reuse with false labels; in one case there were
multiple instances of use of a felt-tip pen to add a
band. Use of a computer to alter the content of
the image has been seen less frequently.6
Table 1 compiles the nature of the
misrepresentations involving questioned images
in 19 ORI cases. The most common allegation
was falsification of data by misrepresenting the
results as being a different experiment, which
also includes the attendant falsification of
molecular weights. Only five examples occurred
in which the lane markers were cut, re-grafted,
and shifted so as to fabricate a test that was never
run. Purposeful tampering with the image to
selectively enhance or remove bands has
occurred, but it was not very common. The
allegations of plagiarism involved falsification of
9
5
2
1
1
figures copied from published journal figures or
by use (and falsification) of images obtained
from the Internet homepages of other scientists.
Other aspects of these image-related
allegations are described in Table 2. Thesis
research appears to provide a relatively frequent
source of questioned images, falsified by both
students and mentors. In three cases, the images
were allegedly obtained from others, and in two
other cases they involved falsification of images
that had been published earlier by the same
laboratory. The source of most of these
allegations was co-workers, although in five
cases it was a reviewer who recognized the
image as being from another source, or saw
intrinsic evidence that the image could not be
authentic. Most allegations did not arise because
the images looked inauthentic, but simply
because they were either recognized as false or
represented claims that a reviewer frankly
disbelieved. The questioned image was often
the one concern in a case that could not be easily
dismissed.
Discussion
The facts uncovered by the forensic examination
of questioned images can often be corroborated
by other evidence, such as absence of records/
experiments on the requisite date(s), the
existence of dated computer files or other
versions, parallel images in publications, etc. In
addition to the basic image processing, a clear
267
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
follow-up analysis is important.
The most useful analysis of questioned
scientific images is done with a clear
understanding of the experiment in question.
This often requires collaboration between the
image analysis and individuals who have a direct
familiarity with the conduct of the scientific
experiments at issue. (9) To date, only two
institutions have reported to ORI using a
computer-based image analysis. Only one
institution documented those results; in that
instance, image processing by a committee
member uncovered details that were
determinative (see Figure 8). The information
from ORI’s cases indicates that most allegations
involved “reuse” of the image to represent data
from a purportedly different experiment.
Occasionally, photographs of gels or blots were
“cut and pasted” together in different
combinations. Manipulations by computer were
less common.
An image by itself creates a mantle of
authenticity, if only because we give unusual
weight to what we see. Yet in those cases where
scientific misconduct was found, discovery of
one falsified image often led to the discovery of
another, and in all the “original’ laboratory
records were “missing.” Thus good laboratory
practice may help to deter or to minimize the
impact of falsification.
Notes
1
2
3
4
5
6
Any views expressed in this article are my own and do not
necessarily reflect those of the Office of Research
Integrity. The citation of items in this article does not
connote a product endorsement.
The questions are not limited to examining items that
look alike. For example, immunoblots from the same
gel can be stripped and re-probed with entirely new
labeled antibody to reveal different protein bands.
The forensic value of the background information is
completely analogous to the basis for numerical forensic
analyses developed by Dr. James Mosimann in another
presentation at this meeting.
A simple “thought” experiment makes the point more
elegantly than examining the underlying physiology of
visual perception: any two gray levels, so alike that they
could be fairly represented as one shade, could still be
assigned two separate colors, say red and blue, of the
same intensity. (8)
Notice that digitizing at greater bit depth, such as 12 bit,
would in principle detect fainter differences in shading
to 1/4096 parts, rather than the 1/256 parts shown here.
It is debatable as to whether it would be more or less
difficult to detect computer alterations. What can be
said is that an allegation rarely arose because an image
on its face appeared inauthentic.
Bibliography
1. Blitzer, Herbert, “Digital Imaging,” Institute of Forensic
Image Processing, Purdue School of Engineering and
Technology, IUPUI. PDF document available on the
Internet at http://www.engr.iupui.edu/ifi/articles.
Published in Law Enforcement Technology, March
2000.
2. Scientific Working Group for Imaging Technologies,
“Definitions and guidelines for the use of Imaging
Technologies in the Criminal Justice System” (vs. 2.1,
June 1999). Department of Justice PDF document
available on the Internet at http://www.for-swg.org/
swgitin.htm.
3. NIH Image is public domain software that was
developed at the National Institutes of Health. A large
number of analytical macros for NIH Image are
available via an active electronic bulletin board. NIH
Image is particularly useful for any quantitative and
analytical measurements, false-coloring to better
visualize or code details of processed images, but it is
limited to 8 bit gray scale images on the Macintosh.
NIH Image is available on the Internet at http://
rsb.info.nih.gov/nih-image/.
4. Photoshop is a popular commercial program for both
computer platforms (Windows and Macintosh).
Photoshop records the history of the computer’s actions,
to document the steps used in the image processing, but
its analytical capabilities are limited. Photoshop can be
obtained from Adobe Systems, Inc., San Jose, CA
95110-2704.
5. Photoshop-compatible “plugins” are available
commercially as the Image Processing Tool Kit®
(IPTK), Reindeer Games, Inc., 20 Battery Park Ave.,
Suite 502, Asheville, NC 28801. IPTK is also
compatible with NIH Image; it greatly extends the
analytical capabilities of either program for analysis of 8
bit gray or 24 bit color images. Information can be
found on the Internet at http://members.aol.com/
ImagProcTK.
6. The base in support for analytical macros for ImageJ is
not as large as the forerunner, NIH Image, but it is
growing. ImageJ and its plugins can be obtained on the
Internet at http://rsb.info.nih.gov/ij/plugins.
7. Inoué, Shinya, “Video Microscopy,” Plenum Press, New
York, N.Y. 1986. pp 80-83.
8. Russ, John C. “The Image Processing Handbook,” CRC
Press, Inc., Boca Raton, 2nd Edition. 1994. p. 213.
9. Indeed, the examples presented in this paper represent
the collaboration and assistance from my ORI
colleagues, Drs. John Dahlberg, Kay Fields, and Nancy
Davidian.
268
Terminal Digits and the Examination of Questioned Data
James E. Mosimann, Office of Research Integrity, DHHS, USA
John E. Dahlberg, Office of Research Integrity, DHHS, USA
Nancy M. Davidian, Office of Research Integrity, DHHS, USA
John W. Krueger, Office of Research Integrity, DHHS, USA
Keywords. Digits, Research misconduct, Statistical forensics, Terminal Digits, Uniform distribution
Our objective is to illustrate the use of statistical methods to examine the authenticity of data in the
investigation of research misconduct. We present examples of statistical analyses of questioned data
from several cases that illustrate the experience of the Office of Research Integrity. We show that the
statistical examination of numbers that are normally unrepeatable when experiments are repeated, or
otherwise are of inconsequential meaning, may reveal substantial clues as to the authenticity of
questioned data when compared with numbers in data that are unquestioned. We illustrate the
occurrence of the uniform distribution of non-leading (insignificant rightmost) digits in unquestioned
numbers, along with examples of deviation from such uniformity for fabricated or falsified numbers.
(Most people are unable to choose digits randomly.) We describe several cases in which a variety of
anomalies in data sets provided the impetus for the examination of rightmost digits. The anomalous
behavior of rightmost digits, when added to testimony and other physical evidence, can greatly
enhance or decrease the credibility of witnesses. The cases discussed involve: 1 and 2, Anomalous
behavior of terminal digits in published or recorded numbers; 3, Terminal odd digits in event times
that should have exhibited only even digits (and why); and 4, Data that were falsified by calculations
from computer spreadsheets (detected by the inclusion of an additional digit of accuracy).
Introduction
Allegations of research misconduct1 often are of the form that a particular experiment was not done
as described, or not done at all. In considering such allegations it is often necessary to examine
“questioned” data. Such data can establish that the experiment was performed as described.
However, if the allegation is true, then these questioned data are necessarily falsified or fabricated.
A useful way to assess questioned data is to examine inconsequential components of data sets that
are not directly related to the scientific conclusions of the purported experiment. Thus if the
allegation is true and the data are falsified, the falsifier typically devotes attention to numbers that
establish the desired scientific outcome. Properties of the numbers that are not directly related to the
desired outcome are less likely to receive consideration by the falsifier.
The same principle of examining details inconsequential to the scientific outcome appears valid
Corresponding author: J.E. Mosimann (present address), ABL Associates Inc., 5 Balmoral Court, Rockville, MD 20850,
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
whether the data are expressed in non-numeric
form (images, written descriptions) or as
numbers. Here we consider several cases where
the data are numeric and lend themselves to
immediate statistical description.
In all these cases we stress the importance of
comparing “questioned” data with similar
unquestioned data from the same laboratory or
individuals.
Rightmost digits
Consider counts of radioactivity for a biological
preparation; for example, 5071. In a recount of
the sample, or in a replication of the assay, it is
highly unlikely that the rightmost digits will be
the same. Thus with two repetitions of the
experimental procedure, instead of 5071, one
might obtain respectively, 5109 and 4966. The
rightmost, non-leading digits of these three
numbers are not the same. Thus _071 differs
from _109, and in turn both differ from _966.
Digits are often recorded well beyond the
repeatability of the experimental procedure. For
such rightmost digits, theoretically2 there is a
tendency to be uniformly distributed as expected
in a lottery. For example, a uniform distribution
of digits is expected in the Maryland Lottery.
Figure 1 shows the frequencies of the digits 0 to
9 found in 5,106 winning “Pick-3” numbers (of 3
digits each) for the past ten years.3 This
distribution is not significantly different from
uniform. All digits have occurred with nearly the
same frequency, as they should in a lottery.
Figure 1. Ten years of Maryland Lottery Pick Three Digits,
January 2, 1990 to December 31; 15,318 digits.
Case 1: Uniformly distributed rightmost
digits in scintillation counts
In the first case, experimental measurements
were known not to have been done because
radioactive spots on the experimental sheets had
not been excised and hence could not have been
counted in the scintillation counter. Yet the
respondent’s notebook contained (falsified)
handwritten counts for that experiment. In this
case, faced with the evidence, the respondent
admitted to the falsification of the numbers in the
notebook.
In addition to the questioned counts, the
notebook contained handwritten counts that were
supported by counter output, and thus not
falsified. Both questioned and unquestioned
numbers occur in pairs (a numerator and
denominator) and have large numbers of digits
(Table 1).
The following procedure was used to find
digits. The rightmost digit of a number was
designated as occupying “Place 1,” then the digit
to its left occupied “Place 2,” etc. Digits were
examined in four places for each number, except
that the leftmost digit was never included in the
analysis. Thus by way of example, the
underlined digits would be included in the
analysis: 1078, 251183, 735, 62034. It is clear
that a three-digit number contributes two digits
for analysis and a four-digit number, three digits.
Numbers of five or more digits contribute four
digits.
Chi-Square tests for uniformity of digit
distributions from 252 falsified counts from
notebook pages 141-152 are presented in Table 2.
The distributions are not uniform. Three of the
four Chi-Square values have probabilities less
than .05, and when digits from all four places are
grouped together, the total distribution is far from
uniform (Chi-Square = 30.94, df = 9, p=.0003).
Chi-Square tests for uniformity of the digit
distributions from 222 unquestioned counts also
are presented in Table 2. The distributions are
not significantly different from uniform. All of
the four Chi-Square values have probabilities
greater than .05, and when digits from all four
places are grouped together, the total distribution
is not significantly different from uniform (ChiSquare = 11.09, df = 9, p=.27).
The unquestioned counts have uniform or
nearly uniform rightmost digits, whereas the
falsified counts do not.5
270
–––––––––––––––––––––––––––––––––––––––––– Mosimann, et al., Terminal Digits and the Examination of Questioned Data
Falsified counts
(Notebook page 145)
Unquestioned counts
supported by counter
printouts
(Notebook page 135)
Numerator
1078
1770
Denominator
251183
217763
Numerator
82267
105584
Denominator
170679
190994
1091
1434
1247
1131
54350
225853
238995
241139
260074
220938
87592
83341
88426
105068
90707
181133
197822
172062
194570
150614
Table 1. Illustrative falsified and unquestioned counts from the respondent’s laboratory notebook.
Numerator (summation of reaction produced counts) and denominator (residual substrate count)
are associated with a given clone, and activity is expressed by the ratio, numerator divided by
denominator. Note that the 28 counts illustrated each contain from four to six digits.4
Chi-Square Results
For Falsified and Unquestioned Counts
Digits from 252 Falsified Counts
Place 4
Place 3
Number
185
250
Chi-Square
34.8
29.3
D. Freedom
9
9
Probability
.00006
.00058
Place 2
252
13.2
9
.1521
Place 1
252
27.1
9
.0013
Total
939
30.94
9
.0003
Digits from 222 Unquestioned Counts
Place 4
Place 3
Number
195
218
Chi-Square
14.3
9.89
D. Freedom
9
9
Probability
.11
.36
Place 2
222
8.72
9
.46
Place 1
222
11.33
9
.25
Total
857
11.09
9
.270
Table 2. Chi-square results for tests of uniformity of digit frequencies for falsified and unquestioned
counts. The rightmost place is “Place 1”; the next place to left is “Place 2” etc. (Leftmost digits of
numbers were excluded, so there are fewer “Place 4” digits than “Place 3,” etc.)
Case 2: Unlikely Patterns in Rightmost
Digits.
In this case, we again demonstrate the ability of
uniformly distributed digits to distinguish
questioned from unquestioned data. However,
the digit analyses lead further to the identification
of unlikely patterns in numbers that should not be
related, given the purported experiment.
Table 3 (next page) reproduces the means
and standard deviations from a published Table
that was challenged by a coworker.
Lipopolysaccharide extracts (LPS) were purified
from endotoxin from various bacteria. The five
rows in each half represent, respectively,
different bacterial sources for the endotoxin.
LPS was added at various concentrations to the
cell cultures. Thus the five columns of the Table
represent different levels of LPS (left to right,
respectively: 5000, 500, 50, 5, and .5 ng/ml).
The upper half of Table 3 represents cultures to
which endotoxin and stimulator cells were added
at the same time. The lower half represents
cultures to which endotoxin was added 24 hours
prior to the addition of stimulator cells.
However, while supporting notebook data could
be found for the first four columns, no supporting
271
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Column 1
Column 2
Column 3
Column 4
Column 5
Mean
Std Dev
Mean
Std Dev
Mean
Std Dev
Mean
Std Dev
Mean
Std Dev
17697
1739
17399
1680
15085
1342
18262
2934
27191
1404
20164
3540
16746
1171
19397
1133
17889
3919
26999
7107
23323
3861
24154
722
19094
1340
28763
3373
28611
967
24474
4042
18918
4459
14224
828
24596
6327
29152
1407
29711
1519
21855
8458
23840
1695
29669
3222
28765
7104
24752
1455
22498
4591
21639
1347
32825
3063
70714
2106
32683
8535
26321
2753
20015
2020
34030
3917
68177
7155
43411
4682
41980
1705
34026
3906
47703
1894
66004
3924
26535
2349
41592
5699
31262
2796
54588
5065
74316
2192
33216
3762
37036
2071
27513
5062
32033
8307
71117
6817
Table 3. Published Table (Column 5 has questioned data).
Number
Chi-Square
D. Freedom
Probability
Column 1 Column 2
70
69
8.57
5.93
9
9
0.478
0.747
Column 3
69
8.54
9
0.481
Column 4
70
9.14
9
0.42 4
Column 5
69
26.22
9
0.0019
Columns 1-4
278
4.45
9
0.880
Table 4. Tests of Uniformity of Digits for the Columns of the Published Table. Chi-Square tests of rightmost digits
for, respectively, Columns 1 to 5 of the Published Table, and for Columns 1-4, together.
notebook data could be found for the questioned
numbers in column 5.
Of statistical importance is the fact that
means and standard deviations in this Table are
reported to several places. Thus numbers are
recorded with greater precision than the
repeatability that the biological experiment
allows, permitting a digit analysis.
The treatment of rightmost digits is the same
as that for the previous case. Digits are analyzed
in four places with no leftmost digit included in
the analysis.
Only the digits of the questioned Column 5
are significantly different from uniform (p =
.0019). Columns 1 to 4 separately are not
different from uniform (the probability ranges
from .424 to .747). In the aggregate, columns 1
to 4 are again not different from uniform (p =
.88).
Based on the contrast between the digit
distributions for the questioned Column 5 and the
unquestioned columns, the complainant’s
assertion that the experiment for Column 5 was
Place 4
Place 3
Place 2
Place 1
1
7
4
1
9
4
1
0
0
6
0
0
4
7
7
7
4
1
7
Table 5. Vertical Pattern of Digits
not done is strengthened.
Furthermore, examination of the standard
deviations in the upper half of Column 5 of Table
3 reveals a remarkable “vertical” pattern. These
numbers should be statistically independent from
row to row. However moving vertically
downward at each digit place reveals a
symmetrical recurrence of digits: 1,7, blank, 1, 7;
4, 1, 9, 4,1; then 0, 0, 6, 0, 0; and finally, 4, 7, 7,
7, 4 (Table 5).
The vertical pattern does not appear
consistent with the five presumably statistically
independent experiments depicted by the separate
272
–––––––––––––––––––––––––––––––––––––––––– Mosimann, et al., Terminal Digits and the Examination of Questioned Data
determined to occur whenever a
peak on the recording of current
equaled or exceeded 10 picoAmps.
26428
406
6428
406
116428
3406
Since the spontaneous “firings”
were infrequent, the continuous
7824
376
7824
376
17824
3761
record of the electrical signal was
24840
1107
24840
1107
124840
7107
not retained. Instead, an “event
detection” device sent the time and
the amplitude of the firing to Excel
26660
345
6501
355
116660
34511
spreadsheets as a permanent
7791
407
7906
348
17791
407
record of the experiment.
9276
1498
12016
1476
9276
1498
To graph the activity of
muscles from different genetic
Table 6.
crosses, the firings of various
rows of Table 3. Such a pattern is consistent with amplitudes were accumulated into bins of 5picoAmp width (10-15, 15-20, 20-25, etc), with
the formation of the numbers after the outline of
accumulation continuing until some bin
the published Table had been established.
Finally, to check for the possible existence of contained 100 firings.6 The resulting frequency
a pattern, three publications by the respondent
distribution represented the pattern of firings (for
(two journal articles and a book chapter) were
Experiment 1, see Figure 2, below, in which there
examined. Examination of these publications
are just over 100 events in the 20-25 bin).
reveals patterns of digits that are inconsistent
Prior to publication, the respondent’s
with biological observations. Consider Table 6
coworkers thought that firings should only be
(above), which contains numbers from tables in
defined as those peaks 20 picoAmps or greater.
three different publications by the author, all for a Thus they asked the respondent to prepare a new
similar experimental procedure.
graph like that of Figure 2, but sampling only
In these three publications, rightmost digits
peaks 20 picoAmps or greater (i.e. resampling
that should not be reproducible are the same in
the Excel spreadsheet until some bin contained
the first and third rows, and they would be the
100 such firings.)
same in the second row except for the arbitrary
The respondent submitted a new frequency
addition of a “1” after the “376” in the last
graph that appeared like the first, but truncated at
column. Further, in the fifth row two of the
20 rather than 10. Since one would expect the
standard deviations are “407” while the
shape of the new graph (above 20 picoAmps) to
corresponding means are “7791” and “17791.”
differ, the coworkers questioned the result.
Note that the standard deviation 7107 occurs in
the book chapter and also in
Column 5 of the published
table already discussed. The
respondent in this case
agreed that the data were
“flawed” and retracted the
relevant articles.
Journal 1
Trauma patients
Journal 2
Cancer patients
Case 3: Banker’s
rounding and “odd”
terminal digits
For the purposes of a genetic
study, electro-physiological
measurements of
spontaneous “firings” (action
potential spikes) of isolated
muscle fibers were made.
Initially, a firing was
Book
Trauma patients
Figure 2. Binning of amplitudes into bins of 5-picoAmp width (initial 321
records of Experiment 1).
273
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
The respondent asserted that the new graph
was not simply a truncated version of the first,
but represented a fresh sampling of peaks greater
than 20 picoAmps. He asserted that he had
properly sampled the peaks in an Excel
spreadsheet by counting beyond the initial 321
records on which the first graph (Figure 2) was
based. The respondent furnished an Excel
worksheet, “Experiment 1,” of 551 records in
support of the new graph. This worksheet
contained the initial 321 records along with 230
additional records.
In addition to the Excel worksheet for
Experiment 1, the respondent also provided a
worksheet of unquestioned data “Experiment 2”
with 1026 records. For Experiment 1 and the 10
picoAmp peaks, the initial 321 records of
Experiment 1 are largely determined since the
initial Figure 2 is known. Thus the last 230
records of Experiment 1 are more questionable.
Since all 551 records were provided after the
allegation, the opportunity existed to falsify or
fabricate time points, but if falsifications occur,
most would be expected in the last 230 records.
Table 7, below, presents the first 12 records of
Experiment 1.
It is interesting to note that all of the time
values in Table 7 terminate in an even digit. The
occurrence of only even time values can be
explained by a long-used7 practice sometimes
known as “Banker’s Rounding.8 ”
A simple explanation of the even terminal
digits for time values is that two successive timevalues are used in determining a peak, and the
Figure 3. Experiment 1: first 321 time points; 321
terminal digits from 321 numbers. (Note presence of six
odd digits.)
mid-point of the two is recorded. Thus when
successive time values are added and divided by
2, the resulting terminal digit is 5 and would be
rounded to an even digit, for example: (1000 +
1001)/2 = 1000.5 rounds to 1000, and (108.7 +
108.8)/2 = 108.75 round to 108.8. Therefore if
numbers ending in 5 are rounded, only even
numbers occur. The rounding of terminal 5’s to
the nearest even digit is the ANSI/IEEE standard9
for rounding terminal 5’s in computers.
Examination of the terminal digits of the 1026
time values of the unquestioned data in
Experiment 2 reveals no times ending in an odd
digit. (The distribution of the 1026 penultimate
digits of the times for Experiment 2 is not
different from uniform (Chi-Square = 14.6, df =
9, p = .10).) In contrast, the questioned
Experiment 1 contains time values that end in
odd digits, reflecting insertions and alterations.
In the initial 321 time points, six terminate in an
odd digit (Figure 3). (The distribution of the 315
penultimate digits from the potentially unaltered
even times is not different from uniform (ChiSquare = 8.14, df = 9, p = .52).)
Examination of the graph (Figure 4) of the final
230 records of Experiment 1 reveals many more
(58) time values with odd terminal digits10 than
Figure 3. (The distribution of the 172 penultimate
digits from the even, potentially unaltered, times is
not different from uniform (Chi-Square = 12.3, df =
9, p = .20), whereas the distribution of 58
penultimate digits from falsified times ending in an
odd digit deviates significantly from uniform (ChiSquare = 33.0, p = .00013).
Figure 4. Experiment 1: last 230 time points; 230 terminal
digits from 230 numbers. (Note presence of 58 odd digits.)
274
–––––––––––––––––––––––––––––––––––––––––– Mosimann, et al., Terminal Digits and the Examination of Questioned Data
Time in Minutes
0 .0648
0 .4904
0 .4952
0 .5398
0 .9454
1 .7182
2 .6950
3 .3626
3 .7294
3 .8586
4 .3494
4 .3712
Many more time
Experiment 1 - First 12 Records
values
terminate in odd
Amplitude in picoAmps Terminal Digit of Time
digits
in
the final portion of
16.1
8
Experiment
1, as expected
22.7
4
if
falsification
occurred.
33.2
2
The
occurrence
of time
19.8
8
values
ending
in
odd digits,
36.1
4
mostly
in
the
latter
part of
44.4
2
Experiment
1
(and
the
lack
20.5
0
of
uniformity
of
their
19.3
6
penultimate digits) indicates
17.6
4
data
falsification. The
14.9
6
timing
of the occurrence of
12.9
4
the
minutes
ending in odd
45.4
2
digits is illustrated in
Figures 5 and 6.
From Figure 6 it can
be seen that not only do
most of the odd time
values occur in the last part of
Experiment 1 (after minute
137.3006); it also appears from
the denseness of the plot in the
latter that the values
immediately after this time
point are quite close together.
Further statistical tests of the
intervals between events
confirms the increased density
in the latter part of Experiment
1, indicating the insertion of
fabricated firing events.
Table 7. The first 12 records of Experiment 1. Note that amplitudes include values less
than 20, as expected. Also note that the terminal digit of the time is an “even” number
for all 12 records.
Figure 5 (above). Experiment
2, unquestioned, 699 amplitudes
(abs>20). (No amplitude is
associated with an odd
minutes.)
Figure 6 (right). Experiment 1,
questioned; 371 amplitudes with
abs>20, 52 with odd minutes.
(Negative values, even minutes;
positive values, odd minutes.)
275
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Rats
M-1
M-2
M-3
M-4
M-5
M-6
M-7
Weights-1
2.495
1.695
0.738
0.780
0.276
4.128
1.131
Weights-2
3.008
2.272
1.495
0.231
0.122
3.413
1.224
Weights-3
2.7515
1.9835
1.1165
0.5055
0.199
3.7705
1.1775
Weights-4
4.631
3.019
1.768
0.394
0.155
2.261
2.805
Weights-5
2.250
0.702
0.843
0.085
0.205
1.187
0.726
Weights-6
3.4405
1.8605
1.3055
0.2395
0.180
1.724
1.7655
Table 8. Portion of Excel spreadsheet with weights of muscles of rats 1-6. Note that some entries for columns Weights-3
and Weights-6 have four decimal digits and end in 5, whereas other entries have at most three decimal digits.
followed timely procedures.
The respondent presented to university
officials blood flow and weight data for six rats
on an Excel spreadsheet as well as purportedly
original data sheets with handwritten entries for
the muscle weights for six rats. Weights of 28
muscles and three other body parts for six rats
extracted from the Excel printout are presented in
Table 8 .12 Further weights as found in
handwritten entries on separate data recording
sheets for six rats are presented here in Table 9.
In Table 8, columns Weights-1, Weights-2,
Weights-3 and Weights-6 correspond,
respectively to columns 314-1, 314-2, 315-1 and
316-2 in Table 9. Thus the handwritten
“original” data on the four data recording sheets
(314-1, 314-2, 315-1 and 316-2) correspond to
the columns labeled, respectively, Weights-1,
Weights-2, Weights-3, and Weights-6 on the
Excel spreadsheet. The columns Weights-4 and
Weights-5 do not correspond to two additional
data recording sheets labeled 315-2 and 316-1.
When values within a spreadsheet are
calculated, rather than transcribed, the numbers
may display more digits of accuracy than the
original numbers that
are the source of the
Rats
calculated values.
314-1
314-2
315-1
315-2
316-1
316-2
Therefore, looking for
M-1
2.495
3.008
2 .725
3.859
3.479
3.440
enhanced precision in
M-2
1.695
2.272
1 .984
2.087
1.881
1.861
spreadsheet numbers
can indicate that
M-3
0.738
1.495
1 .117
1.464
1.320
1.306
certain numbers have
M-4
0.780
0.231
0 .506
0.269
0.242
0.240
been calculated or
M-5
0.276
0.122
0 .199
0.202
0.182
0.180
randomly generated by
M-6
4.128
3.413
3 .771
1.933
1.743
1.724
the spreadsheet
M-7
1.131
1.224
1 .178
1.980
1.785
1.766
software.
Since data are
Table 9. A portion of rat muscles weights from handwritten entries on six data recording
presented
for six rats,
sheets. Note that all numbers have a precision of three decimal places.
Case 4: One terminal digit too many
An investigator conducted studies on the effect of
rhythmic contractions of skeletal muscle on
blood flow using the hind limbs of rats. Blood
flow was measured at rest and during nerve
stimulation. In addition to measurements of
blood flow, weights of individual skeletal
muscles were recorded on data sheets. The
experimental results for six rats were presented in
a laboratory seminar. Sometime later a coworker discovered that two of six data sheets
were blank, and became suspicious that the
measurements (blood flow/weights) had not been
made for those rats. Suspicions were confirmed
when frozen rat carcasses were checked.
Although four had the hind limb muscles
dissected, two were still intact and un-dissected.
When confronted, the investigator (now
respondent) admitted to falsifying data for two
experimental animals. However, he subsequently
withdrew the admission and denied the charges.
The respondent stated that there was no evidence
to support the claims that the research was
falsified,11 and that the university had not
276
–––––––––––––––––––––––––––––––––––––––––– Mosimann, et al., Terminal Digits and the Examination of Questioned Data
Rat-3
M-1
M-2
M-3
M-4
M-5
M-6
M-7
Rat-6
Mean 1,2
Weights-3
315-1
Difference
Mean 4,5
Weights-6
316-2
Difference
2 .7515
1 .9835
1 .1165
0 .5055
0.199
3 .7705
1 .1775
2.7515
1.9835
1.1165
0.5055
0.199
3.7705
1.1775
2.725
1.984
1.117
0.506
0.199
3.771
1.178
0.0265
-0.0005
-0.0005
-0.0005
0
-0.0005
-0.0005
3.4405
1.8605
1.3055
0.2395
0.18
1.724
1.7655
3.4405
1.8605
1.3055
0.2395
0.18
1.724
1.7655
3.440
1.861
1.306
0.240
0.180
1.724
1.766
0.0005
-0.0005
-0.0005
-0.0005
0
0
-0.0005
Table 10. A Portion of the Weights for Rat 3 and Rat 6. The weights for Rat 3 are precisely the means of the respective
weights for Rats 1 and 2. Additionally, the weights for Rat 3 correspond to three decimals to the handwritten weights for
Rat 315-1. (The only exception is the weight for M-1 (shaded) where the rounded 2.752 is transcribed as 2.725.
Correspondingly, the weights for Rat 6 are precisely the means of the respective weights for Rats 4 and 5. Additionally, the
weights for Rat 6 correspond to three decimals to the handwritten weights for Rat 316-2, without exception.
and at most four allegedly were measured, the
spreadsheet was evaluated for signs that some of
the columns contained calculated values, rather
than valid data entered from experimental
records. The columns Weights-3 and Weights-6
in the Excel spreadsheet (Table 8) contain a
number of entries that are recorded to one more
decimal accuracy than the other columns
(Weights-1, Weights-2, Weights-4, Weights-5).
Additionally, these same entries for Weights-3
and Weights-6 contain one more digit than the
purported original handwritten data as recorded
on the sheets labeled 315-1 and 316-2 (Table 9).
This extra precision could not occur from manual
entry of the weights from the raw data sheets.
Instead, the presence of an extra digit
indicates the possibility that these two columns
represent calculated data. Further, where the
extra digit occurs, it is always a “5.” This
indicates the calculation may have involved
division by “2,” suggesting that those numbers
could be the means of two columns. (When the
sum of the two numbers is even, there is no
increase of the non-zero digits; however, when
the sum is odd, division by 2 produces an
additional “5” digit.)
In fact, the column Weights-3 is precisely the
mean of columns Weights-1 and Weights-2 (see
Table 10, below). Correspondingly, the column
Weights-6 is the mean of columns Weights-4 and
Weights-5 (Table 10).
Since these two columns are calculated on
the spreadsheet, the “original” data on the
recording sheets 315-1 and 316-2 are copied,
respectively, from the spreadsheet-calculated
columns Weights-3 and Weights-6. The only
modification is that the “original” copied data are
only transcribed to three-decimal accuracy as
found on the (presumably) valid sheets labeled
314-1 and 314-2.
Lacking muscle-weight data for two rats, the
respondent generated weights by twice forming
means of measurements of other rats. The
presence of the extra digit in the Excel
spreadsheet provided the needed clue. When the
respondent was shown that the two rats’ weights
were clearly produced as means, not measures,
he accepted the finding of scientific misconduct.
Notes
1.
2.
3.
4.
277
65 Federal Register 76260, December 6, 2000.
A theoretical discussion is found in J. E. Mosimann
and M. V. Ratnaparkhi, “Uniform occurrence of digits
for folded and mixture distributions on finite
intervals,” Communications in Statistics, 1996, 25(2),
pp 481-506. Among other issues, this paper discusses
approximations to continuous distributions by
histogram-distributions for which the uniformity of
terminal digits up to a specified place is known. Such
theoretical issues are important, but our emphasis here
is on direct comparison of questioned data with
unquestioned data.
On May 1, 1995, the Maryland Lottery initiated a
midday pick-3 drawing for weekdays. This is in
addition to the nightly drawing. Thus there are more
than 3,650 winning pick-3 numbers over the ten-year
period. Maryland Lottery results may be found at the
official website, http://www.mdlottery.com.
In all there are 474 counts: 252 admittedly falsified
(notebook pages 141-152) and 222 unquestioned
counts that are supported by counter printouts
(notebook pages 104-106, 130-131, 134-135). Each
count, falsified or unquestioned, contains from three to
six digits. Digits were tested in four places, but no
digit that was itself the leftmost digit was included in
the analysis. Total analyses included 939 digits from
252 falsified numbers and 857 digits from 222
unquestioned numbers.
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
5.
See “Data Fabrication: Can people generate Random
Digits?” J. E. Mosimann, C. V. Wiseman and R. E.
Edelman, Accountability in Research, 4, 31-55, 1995.
This study shows that many people have difficulty
fabricating random digits, even when trying to do so.
6.
“Inverse” sampling until a certain number of a
particular event occurs has a long history, particularly
where rare events are to be studied. (For example, see
J. E. Mosimann, “On the compound negative
multinomial distribution and correlations among
inversely sampled pollen counts,” 1963, Biometrika,
50, 47-54).
7.
“It is conventional to round off to the nearest even
digit when the number to be rounded is exactly half
way between two successive digits.” pp. 13-14, Paul
S. Dwyer, Linear Computations, 1951, John Wiley &
Sons Inc., i-xi , 1 – 344. (See also the next two
footnotes.)
8. “PowerBASIC always rounds towards the closest even
number. For example, both 1.5 and 2.5 would be
rounded to 2. This is called banker’s rounding. …” p.
169, User’s Guide, 1997, PowerBASIC, Inc. 316 Mid
Valley Center, Carmel, California, i-vi, 1-318.
9. ANSI/IEEE Std 854-1987, October 5, 1987, “ANSI”
denotes the American National Standards Institute and
“IEEE” denotes the Institute of Electrical and
Electronic Engineers, Inc. “4.1 Round to Nearest. …if
the two nearest representable values are equally near,
the one with its least significant digit even shall be
delivered.” “5.4 Round Floating Point Number to
Integral Value. …when rounding to nearest, if the
difference between the unrounded operand and the
rounded result is exactly one half, the rounded result is
even.”
10. 46 of these 58 time values that terminate in odd digits
occur with amplitudes greater than 20 picoAmps. In
the initial 321 records of Experiment 1, 6 of 6 odd
time values occur with amplitudes greater than 20
picoAmps.
11. It is only after the respondent denied the charges and
findings of the institution that the ORI demonstrated
which two rats on the spreadsheet represented falsified
data, and the manner of falsification.
12. The spreadsheet also contains columns of numbers
representing blood pressure measurements and
radioactive counts, some of which the university
committee regarded as falsified. These are not
presented here.
278
9. Publication Practices
Guidelines on Plagiarism and Paraphrasing in Writing Manuals Across
Various Disciplines
Miguel Roig, Department of Psychology, St. John’s University, NY, USA
Jaclyn de Jacquant, Department of Psychology, St. John’s University, NY, USA
Keywords: Paraphrasing, Plagiarism, Writing guides
Discussions of plagiarism in conventional writing manuals typically focus on acknowledging the
source of borrowed ideas and text. Such coverage often includes guidelines for proper attribution and
citation practices. A number of manuals also provide specific guidelines for correct paraphrasing. By
correct paraphrasing, we mean the extent to which text from an original source should be modified in
order for it not to be considered a potential case of plagiarism. Those manuals that cover proper
paraphrasing practices (1-3), generally suggest that, in addition to providing a citation, authors should
always paraphrase others’ work using their own words and expressions and avoid the use of the
original author’s language. For example, in a widely used guide, the authors state “When you
paraphrase or summarize, you should use your own words and sentence structure (4). Imitating
syntax, rearranging words and phrases, and borrowing phrases even as brief as two or three words do
not change the original sufficiently to avoid plagiarism” (pg. 66).
Aside from the above guideline on paraphrasing, we are not aware of any other major writing
manual that provides as close an operational definition for correct paraphrasing as the above example
illustrates. However, the examples of proper paraphrasing provided by conventional manuals that
offer such coverage suggest that a correct paraphrase must represent a very substantial modification
of the original text, otherwise the paraphrase may constitute plagiarism. Moreover, some manuals
such as the one quoted above, even suggest that, to avoid plagiarism when paraphrasing, not only
should the original words be changed, but also the sentence structure of the newly paraphrased text
must be different from that of the original (4-7).
As the reader might suspect, the criteria for correct paraphrasing appear to differ from writer to
writer, particularly for inexperienced writers. For example, recent studies by one of the present
authors have reported wide differences in plagiarism/paraphrasing criteria among college students (8,
9). Furthermore, similar differences also appear to exist among professionals, including physicians,
English professors, and journal editors, and between college professors from a variety of disciplines
(10-11). Some authors have even begun to express concern about the writing practices of those who
engage in ‘light’ paraphrasing of others’ works and terms, such as ‘patchwriting’ and
‘paraphragiarism’, have been offered to describe some of these inappropriate paraphrasing practices
(12-14).
Depending on a number of factors, federal agencies, such as the National Science Foundation and
the Office of Research Integrity do not classify inappropriate paraphrasing as instances of research
Correspon