The Harvard Trauma Questionnaire: Reliability and Validity

Document technical information

Format pdf
Size 1.6 MB
First found May 22, 2018

Document content analysis

Category Also themed
Language
English
Type
not defined
Concepts
no text concepts found

Persons

James Phillips Wilson
James Phillips Wilson

wikipedia, lookup

Organizations

Places

Transcript

The Harvard Trauma Questionnaire:
Reliability and Validity Generalization Studies of the Symptom Scales
Chantal Darzi
Dissertation submitted to the
School of Psychology
in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
In
Clinical Psychology
Faculty of Social Sciences
University of Ottawa
Ottawa, Ontario
© Chantal Darzi, Ottawa, Canada, 2017
ii
For Raja
et nos deux princesses,
Sophia et Mina
iii
Abstract
The cross-cultural applicability of the PTSD diagnosis has been widely disputed in recent years.
Consequently, an examination of the psychometric properties of instruments that are used to
assess traumatized individuals of various cultures is of utmost importance. To respond to this
need, the overall goal of this dissertation was to evaluate the psychometric properties of the
Harvard Trauma Questionnaire (HTQ; Mollica et al., 1992), a measure that was developed to
assess trauma symptoms across cultures. In the first study, I conducted a search of all
publications and dissertations that used the symptoms scales of the HTQ. This search revealed
that the HTQ is commonly used by trauma researchers, however only a minority of them
reported using established translation and cultural adaptation procedures to adapt the instrument
for their specific sample. In addition, of the 384 studies considered for inclusion, only 44% of
them reported internal consistency estimates of their sample. I then performed reliability
generalization analyses on Cronbach’s alpha coefficients to assess the reliability properties of the
HTQ symptom scales. Overall, 103 samples were included in the analyses, representing various
cultures, languages and countries of study. The findings of this study indicated that both the
HTQ-16 and 30 symptom scales are likely to provide reliable scores across diverse populations.
However, the evidence supporting the reliability of scores produced for the re-experiencing,
avoidance/numbing and arousal subscales is less strong. Significant moderating effects were
found for various sample and methodological variables, such as the gender composition of the
sample, cultural group, cultural orientation of the country of origin and trauma type. Building
upon the findings of study 1, I performed validity generalization (VG) analyses to assess the
overall construct validity of the HTQ symptom scales in Study 2. Seventy-five independent
samples were included in the VG that evaluated the convergent and discriminant validity
iv
properties of both the HTQ-16 and HTQ-30. The findings revealed that the convergent validity
properties of the HTQ-16 are supported to some extent, but the discriminant validity properties
are not. Furthermore, there was limited support for either the convergent or discriminant validity
of the HTQ-30. Several significant moderating effects were also found for both scales (i.e. age,
gender, cultural group, recruitment site, trauma type, being an original sample). Although these
studies shed some light into the overall psychometric strength of the HTQ symptom scales, the
decision whether to use this instrument for the assessment of PTSD should also be guided by
evidence-based assessment guidelines.
v
Acknowledgements
I would like to express sincere gratitude to Dr. John Hunsley, my thesis supervisor, who
has guided me throughout my graduate studies and provided consistent and invaluable support.
Merci infiniment John. I am also indebted to my thesis committee members, Drs. Andrea
Ashbaugh, Catherine Lee and Dave Miranda, and my external examiner, Dr. Paul Frewen. Your
expertise and thoughtful feedback have made this thesis stronger and helped me develop as a
researcher. Thank you to my friends and colleagues who shared their experiences with me and
allowed me to share mine. Enfin, j’exprime ma profonde reconnaissance à ma famille qui a
toujours été là pour moi. Je vous aime beaucoup.
vi
Statement of Co-Authorship
The two manuscripts included in this dissertation were prepared in collaboration with my
dissertation supervisor. I was primary author and Dr. John Hunsley was the secondary author for
the first manuscript, entitled “The Harvard Trauma Questionnaire: A Reliability Generalization
Study”, and the second manuscript entitled “A Validity Generalization Study of the Harvard
Trauma Questionnaire.” As the primary author on all manuscripts, I was responsible for the
conceptualization of the research questions and methods, coding, planning and execution of
statistical analyses, and preparation of manuscripts. Dr. Hunsley was involved in coding and
providing guidance and assistance in all aspects of the project, especially in the refinement of the
research questions and methods, and editing of the manuscripts.
vii
Table of contents
Abstract…………………………………………………………………………………………...iii
Acknowledgements………………………………………………………………………………..v
Statement of Co-Authorship……………………………...………………………………………vi
List of Tables and Figures………………………………………………………………………xii
General Introduction………………………………………………………………………………1
Overview of the Dissertation…………………………………………………………………..3
Psychometric Properties of Clinical Assessment Instruments…………………………………4
Reliability…………………………………………………………………………………...4
Internal Consistency……………………………………………………………………..7
Validity……………………………………………………………………………………...8
Evaluating the Psychometric Properties of Clinical Instruments……………………………..10
Reliability Generalization………………………………………………………………….12
Validity Generalization………………..…………………………………………………..14
Cross-Cultural Assessment……………………………………………..…………………….15
Posttraumatic Stress Disorder (PTSD): An Overview…………………………………..……18
Prevalence of Trauma and PTSD………………………………………………………….21
Cultural Considerations of Trauma………………………………………………………..23
The Assessment of PTSD…………………………...……………………………………..27
Assessment of PTSD Across Cultures…………………………………………………….29
The Harvard Trauma Questionnaire (HTQ)…………………………………………………..31
Potential Moderators of Reliability and Validity……………………………………………..37
The Current Studies…………………………………………………………………………..42
viii
Study One - The Harvard Trauma Questionnaire: A Reliability Generalization Study…………43
Abstract……………………………………………………………………………………….44
Introduction…………………………………………………………………………………..45
The Harvard Trauma Questionnaire………………………………………………………46
Reliability and Reliability Generalization (RG)…………………………………………..49
The Current Study…………………………………………………………………………50
Method………………………………………………………………………………………..53
Literature Search and Data Collection…………………………………………………….53
Inclusion Criteria……………………………………………………………………….53
Data Reduction…………………………………………………………………………54
Coding of Descriptive Data and Moderators…………………………………………...56
Data Analysis……………………………………………………………………………...58
Calculating Mean Effect Sizes…………………………………………………………58
Random-Effects Model…………………………………………………………………59
Heterogeneity of Effects………………………………………………………………..59
Spearman-Brown Prophecy Formula…………………………………………………..60
Possibility of Publication Bias…………………………………………………………60
Analysis of Moderators…………………………………………………………………60
Results………………………………………………………………………………………...61
The HTQ Symptom Scales in the Empirical Literature…………………………………...61
Reporting Practices of Reliability Estimates………………………………………………62
Overview of Studies Included in the Meta-Analyses and Descriptive Statistics of Potential
Moderators…………………………………………………………………………………62
ix
Reliability Generalization Analyses……………………………………………………….63
Spearman-Brown Formula……………………………………………………………...64
Publication Bias………………………………………………………………………...64
Moderator Analyses…………………………………………………………………….64
Discussion…………………………………………………………………………………….65
The HTQ in the Empirical Literature……………………………………………………...65
RG Analyses……………………………………………………………………………….66
Moderator Analyses………………………………………………………………………..68
Limitations of the Study…………………………………………………………………...71
Implications and Recommendations……………………………………………………….72
References…………………………………………………………………………………….74
Study Two: A Validity Generalization Study of the Harvard Trauma Questionnaire…………...90
Abstract……………………………………………………………………………………….91
Introduction…………………………………………………………………………………...92
The Harvard Trauma Questionnaire……………………………………………………….92
Validity and Validity Generalization (VG)………………………………………………..94
The Current Study…………………………………………………………………………96
Method………………………………………………………………………………………..98
Literature Search and Data Collection…………………………………………………….98
Coding of Descriptive Data and Moderators…………………………………...………99
Data Analysis…………………………………..………………………………………...101
Calculating Mean Effect Sizes……………………………………..…………………101
Random-Effects Model…………………………………………….…………………102
x
Heterogeneity of Effects…………………….…………………………..…………….102
Possibility of Publication Bias………………………………………..……………….102
Analysis of Moderators………………………………………………………………..103
Results……………………………………………………………………………………….103
Overview of Studies Included in the Meta-Analyses and Descriptive Statistics of Potential
Moderators………………………………………………………………………………..103
Validity Generalization Analyses………………………………………………………...105
Convergent Validity……………………………………………………………….......105
Discriminant Validity…………………………………………………………………106
Publication Bias……………………………………………………………………….106
Moderator Analyses…………………………………………………………………...107
Discussion…………………………………………………………………………………...108
Moderators ………………………………………………………………………………112
Limitations of the Study………………………………………………………………….115
Implications and Recommendations……………………………………………………...116
References…………………………………………………………………………………...118
General Discussion……………………………………………………………………………..132
The Psychometric Properties of the HTQ Symptom scales…………………………………132
The Role of Moderators……………………………………………………………………..137
Limitations of the Studies…………………………………………………………………...141
Implications………………………………………………………………………………….143
Recommendations for Researchers………………………………………………………….146
Recommendations for Clinicians……………………………………………………………148
xi
Conclusion…………………………………………………………………………………..149
References (General Introduction and Discussion)………… …………………………………152
Appendix A: Reliability and Validity Generalizations Coding Manual………………………..181
Appendix B: Reliability and Validity Generalizations Coding Sheet………………………….187
Appendix C: Studies Included in the Reliability Generalization of the HTQ (Study 1)……….192
Appendix D: Studies Included in the Validity Generalization of the HTQ (Study 2)………….207
xii
List of Tables and Figures
General Introduction
Table 1: Criteria for Rating Psychometric Properties of Instruments……………………………12
Table 2: Summary of Recommendations for Conducting RG…………………………………...13
Study 1
Figure 1: Literature Search Flow Chart………………………………………………………….85
Table 1: Descriptive Statistics of the Adaptation and Administration Procedures Reported in
Samples Using the HTQ Symptom Scales …………………………………………………..86
Table 2: Descriptive Statistics of Studies Included in the RG…………………………………..87
Table 3: Descriptive Statistics for the Reliability Values of the HTQ Symptom Scales and
Subscales..………………………………………….……………………………………………88
Table 4: Summary of Moderating Effects for the Internal Consistency of the HTQ Symptom
Scales…………………………………………………………………………………………….89
Study 2
Figure 1: Literature Search Flow Chart………………………………………………………...128
Table 1: Descriptive Statistics for the Convergent Validity of the HTQ Symptom scales……..129
Table 2: Descriptive Statistics for the Discriminant Validity of the HTQ Symptom scales…...130
Table 3: Summary of Moderating Effects for the Construct Validity of the HTQ Symptom
scales…………………………………………………………………………………………...131
1
General Introduction
Evidence-based practice (EBP) has become increasingly prominent in health care,
including mental health services (Hunsley & Mash, 2007). Although the focus of EBP in mental
health has been mainly on treatment, evidence-based assessment (EBA) has been highlighted in
various EBP guidelines (e.g., American Psychological Association Presidential Task Force on
Evidence-Based Practice, 2006; Canadian Psychological Association Task Force on EvidenceBased Practice of Psychological Treatments, 2012). A key principle underlying EBA is that
research and theory should be used to guide decision-making regarding assessment targets,
methods and measures, and the assessment process itself (Hunsley & Mash, 2007). This can lead
to more consistency regarding diagnosis across clinicians, researchers, and settings, and also to
the provision of higher quality treatment (Speroff et al., 2012). An important element of EBA is
using psychometrically strong measures, that is, ensuring that the reliability and validity
properties of an instrument are sound when used with a specific sample for a specific purpose
(Hunsley & Mash, 2007). For example, if a clinician is planning to use an instrument in the
process of determining whether a Vietnamese refugee meets criteria for a diagnosis of PTSD, the
clinician must ensure that the instrument he or she decides to use has supporting psychometric
evidence for diagnostic purposes within a Vietnamese refugee population.
Psychodiagnostic assessments rely heavily on the conceptualization of the construct they
are intended to measure. If the conceptual underpinnings of the construct were not solid, then the
psychometric properties of the instrument would be compromised (Haynes, Smith, & Hunsley,
2011). One psychological construct that has been heavily debated in regards to its
conceptualization is the posttraumatic stress disorder (PTSD) diagnosis based on the various
editions of the American Psychiatric Association (APA)’s Diagnostic and Statistical Manual of
2
Mental Disorders (DSM). A central argument is that the PTSD construct was developed
according to Western cultures’ conceptualizations of trauma and its sequelae, and may not be
applicable to other cultures (e.g., Jones et al., 2003). Consequently, it is not surprising that there
has also been much questioning regarding the cross-cultural applicability of screening tools used
to assess PTSD. Although research in cross-cultural assessment in general has received increased
attention and several authors have proposed guidelines for validating instruments for different
cultures (e.g., International Test Commission, 2005), the assessment of PTSD across cultures is
especially challenging due to variations in trauma symptom expression (e.g., Pole, Gone, &
Kulkarni, 2008).
A trauma assessment instrument specifically developed to be used and adapted across
cultures is the Harvard Trauma Questionnaire (HTQ; Mollica et al., 1992). Systematic reviews of
tools used to assess the health of refugees have found that the HTQ has been quite extensively
used by researchers to assess trauma and its sequelae (Gagnon, Tuck, & Barkun, 2004; Hollifield
et al., 2002). Among the various PTSD instruments evaluated in these reviews, the HTQ was
described as being a scientifically strong measure because of the procedures used in its
development and its psychometric properties. In addition, the HTQ has been recommended by
other experts in the field of trauma assessment such as Keane, Silberbogen, and Weierich (2008)
and Nakeyar and Frewen (2016). Although this instrument has been recommended above others
to assess PTSD in different cultures, a more detailed investigation of individual studies that have
used the HTQ suggests that its psychometric properties may not be so robust across populations
(e.g., Jakobsen, Thoresen, & Johansen, 2011; Rasmussen, Verkuilen, Ho, & Fan, 2015; Silove et
al., 2007). Given the popularity of this instrument and the inconsistent findings among the few
3
validation studies available on the HTQ, a further investigation into the overall psychometric
strength of the HTQ is needed.
Overview of the Dissertation
This dissertation contributes to the EBA literature by examining the empirical evidence
of the psychometric properties of the HTQ symptom scales. The goal of this research is to
provide information regarding the cross-cultural applicability of the HTQ symptom scales,
thereby helping clinicians and researchers make evidence-based decisions regarding their use of
this instrument with individuals from different cultures. The data for two meta-analytic studies
were extracted from empirical publications and dissertations that used the HTQ symptom scales
in their own data collection. These studies were coded and the information obtained was used to
(a) provide descriptive statistics of how the HTQ symptom scales were used in the empirical
literature and (b) evaluate its psychometric properties, specifically reliability and validity. The
first study included a reliability generalization (RG) study in which I calculated overall mean
reliability estimates (i.e., Cronbach’s alpha) across studies. The second study is a validity
generalization (VG) study in which I examined the evidence for convergent and discriminant
validity properties of the HTQ symptom scales across the identified primary studies. I conducted
moderator analyses in both studies to assess whether sample and/or methodological
characteristics had a significant impact on the mean estimates of these psychometric properties.
The general introduction of this dissertation is comprised of six sections. The first is an
overview of the psychometric properties of clinical assessment instruments, which is then
followed by a discussion on the psychometric evaluation of clinical instruments. In the third
section, I discuss important features of cross-cultural assessment. This section is followed by an
overview of PTSD, which includes the prevalence of the disorder and of traumatic events,
4
possible cultural elements of PTSD, and a description of the assessment of PTSD, both in
general and across cultures. Next, I provide a description of the development and psychometric
properties of the Harvard Trauma Questionnaire. In the final section, I present the variables to be
examined as potential moderators of the reliability and validity of the HTQ.
Psychometric Properties of Clinical Assessment Instruments
Measurement is an inherent part of psychological research and clinical work. Researchers
and clinicians regularly make important decisions in the selection of appropriate assessment
strategies as they approach a new case or develop a new research protocol. This entails selecting
the most appropriate measurement instruments, or more precisely, instruments that are
scientifically supported for both the purpose and population at hand. Thus, understanding the
properties that affect the scientific quality of psychological measures is crucial throughout this
decision-making process. Two fundamental concepts that compose the scientific foundations of
measurement are reliability and validity. These concepts are described in the following sections.
Reliability. Reliability, or the precision of a measure, represents “the consistency of the
scores across instances of the testing procedures” (American Educational Research Association
[AERA], American Psychological Association [APA], & National Council on Measurement in
Education [NCME], 2014, p. 33). Furthermore, reliability estimates represent the proportion of
variance that is explained by the true score itself (characteristics measured in the test), as
opposed to the error score, the part of the variance that is explained by sampling or measurement
error (Graham, Yenling, & Jeziorski, 2006). As the reliability of test scores is also a prerequisite
for the validity of an instrument, examining the degree of consistency of scores is imperative in
determining if the instrument is indeed scientifically sound.
5
There are four main types of reliability: (a) test-retest, (b) interrater, (c) parallel forms,
and (d) internal consistency (Salkind, 2012). Test-retest reliability refers to temporal stability,
and can be operationalized as a correlation of scores of a test that has been administered at
different time-points (Haynes et al., 2011). Interrater reliability can be defined as the assessment
of the agreement between two or multiple raters. There are different types of interrater reliability,
including percentage agreement and interrater correlations (Haynes et al., 2011). Parallel forms
reliability refers to the examination of different forms of the same instrument (Salkind, 2012).
Finally, internal consistency refers to the degree of consistency among the scores on items of an
instrument (Haynes et al., 2011) and can be evaluated by comparing two halves of an instrument
(split-half reliability), or by computing an index such as Cronbach’s alpha (Cronbach, 1951).
Cronbach’s alpha is the most commonly used and reported index of reliability (Tavakol &
Dennick, 2011) and it is often used in meta-analytic studies of the reliability properties of an
instrument (Vacha-Hasse & Thompson, 2011). Details regarding internal consistency and
Cronbach’s alpha are described further.
When describing reliability, it is also important to note that it refers to the properties of
test scores of a specific sample and not to the properties of an instrument itself. Saying “this test
is reliable” rather than “this test produces reliable scores for this sample” is one of the most
common errors made in psychology with regard to measurement (Thompson, 1994). Indeed,
factors such as sample characteristics can dramatically influence the degree of consistency of test
scores. For instance, the same instrument can produce scores with differing reliability when
administered to more or less homogenous samples (Thompson, 1994). In sum, reliability is not
only dependent on the quality of a measure, but also the characteristics of the sample, sample
size, and administration and scoring procedures used (Barnes, Harp, & Jung, 2002). Therefore,
6
gathering reliability evidence is especially important for measures that are commonly used with
various types of samples, such as the HTQ.
Researchers often report reliability estimates of prior studies but do not report the data
from their own studies (Vacha-Hasse, Kogan, & Thompson, 2000). Ignoring the sample-specific
nature of reliability may then lead researchers to interpret unreliable data (Henson, Kogan, &
Vacha-Haase, 2001). Although the American Psychological Association Task Force on
Statistical Inference specified that all studies should include the reliability coefficients of the
scores of their sample even if the study is not psychometric in nature (Wilkinson & American
Psychological Association Task Force, 1999), researchers continue to omit reliability coefficients
from reports of their studies. To examine this widespread tendency, Vacha-Haase and Thompson
(2011) conducted a review of 47 previous meta-analytic studies of reliability. Their results
showed that 54% of the 12,994 primary studies included in their review made no mention of
reliability in their studies, whereas 15.7% mentioned reliability but only provided the coefficients
reported in previous studies.
This tendency to generalize reliability properties across studies has been described by
Vacha-Haase et al. (2000) as “reliability induction.” They argued that reliability induction is only
plausible if the previous samples have sample composition and score variability that are similar
to the current set of participants. Even when samples may seem to be comparable, examining
score reliability is still important as there may be subtle differences between samples. One way
of examining the generalizability of the reliability of test scores is by conducting a reliability
generalization (RG) study, which is a meta-analytic method described in further detail in a
subsequent section.
7
Internal consistency. Internal consistency refers to the degree to which all the items in a
test measure the same construct. Cronbach’s alpha is a measure of the inter-relatedness of test
items and has a value between 0 and 1. If the items of a test are highly correlated to each other,
Cronbach’s alpha will be higher (Tavakol & Dennick, 2011). Calculating Cronbach’s alpha is
particularly useful when examining the reliability properties of an instrument that does not have
right or wrong answers (Salkind, 2012), such as the HTQ-symptom scale. It is important to note,
however, that Cronbach’s alpha is affected by the number of items on a test. Therefore, there can
be high internal consistency for other reasons, such as redundant items or using a large number
of items (Haynes et al., 2011). Other difficulties and criticisms associated with coefficient alpha
include: (a) it is often based on unmet assumptions (e.g. assuming that the true score variance is
the same across all items); (b) these assumptions can inflate alpha, thus providing imprecise
estimations of internal consistency; (c) alpha cannot be generalizable if an item is deleted, and
(d) variability is not accounted for in a point estimate of alpha (e.g. Dunn, Baguley, & Brunsden,
2014). Some strategies have been developed to address these concerns about alpha. For instance,
researchers can use bootstrapping methods to produce confidence intervals around the alpha
point estimate to provide a range of probable values (Dunn et al., 2014). Preferably, researchers
are recommended to calculate omega (McDonald, 1999), a different measure of internal
consistency, which has shown to be a more accurate index as compared to alpha (e.g. Zinbarg,
Revelle, Yovel, & Li, 2005). However, until this alternative becomes more accessible and used
in the literature, Cronbach’s alpha currently remains the best available measure of internal
consistency that can be examined by meta-analysis.
Establishing internal consistency is important as it can affect the accuracy of
interpretation of scores and is relevant in both clinical and applied research settings. Calculating
8
typical reliability estimates help determine the effect of measurement error on the observed score
of a sample of participants in research settings (Haynes et al., 2011). In clinical settings, a typical
internal consistency value can also be used to reveal the effect of measurement error on the
observed score of an individual through the calculation of the standard error of measurement
(Haynes et al., 2011; Tavakol & Dennick, 2011). The standard error of measurement is then used
to calculate confidence intervals, providing a range of values that is likely to contain an
individual’s true score. In addition, producing mean internal consistency estimates for samples
allows clinicians and researchers to calculate a reliable change index, which measures the extent
to which an individual makes gains in treatment (Haynes et al., 2011).
As described earlier, reliability is an important step in establishing validity. In other
words, an instrument needs to produce reliable scores in order to produce valid scores. The
internal consistency of an instrument can be used to determine the maximum value of the
validity coefficient that can be obtained. The index of reliability, or the square root of the
reliability estimate, represents the highest validity estimate of the scores produced by the same
sample (Haynes et al., 2011). Once the degree of consistency of the scores of an instrument is
assessed, the next step in the evaluation of a measure is to examine the evidence of validity. The
index of reliability can thus serve as an evaluation benchmark of the validity coefficients.
Validity. Validity refers to the extent to which an assessment instrument measures what
it is supposed to measure. In other words, an instrument that has been evaluated as having good
evidence of validity has variation in scores that reflects the variation in the construct that is to be
measured (Haynes et al., 2011). According to the Standards for Educational and Psychological
Testing (AERA, APA, NCME, 2014, p. 11), validity can also be defined as “the degree to which
9
evidence and theory support the interpretations of test scores entailed by proposed uses of tests”
and is considered the most fundamental element of test development and evaluation.
Similar to reliability, there are various kinds of evidence of validity. The four main
categories are as follows: (a) content validity, (b) criterion validity, (c) incremental validity, and
(c) construct validity. Content validity refers to the extent to which items of an instrument are
representative of the construct that is intended to be measured (Haynes, Richard, & Kubany,
1995), or that the test items cover the range of elements needed to assess a certain construct.
Criterion validity relates to whether an instrument is associated with other current or future
outcomes that are measured by different instruments, and is often described as either predictive
or concurrent: predictive validity is when a test score predicts a certain outcome, and concurrent
validity is when a test score is related to an outcome that is measured simultaneously (Haynes et
al., 2011). Incremental validity entails the question of whether the information retrieved from
one or more measures increases the validity or utility of a clinical decision more than other
already established sources of information (Hunsley & Meyer, 2003). Finally, construct validity
refers to the degree to which a test score reflects the construct of interest, and includes both
convergent and discriminant validity. Convergent validity refers to the extent to which the scores
of an instrument are related to scores of a different measure that is intended to assess the same
construct, or other variables that have theoretically established associations with the construct
(Messick, 1995). Discriminant validity refers to the degree to which the scores of an instrument
are unrelated to measures of unrelated constructs (Haynes et al., 2011). Thus, low correlations
between the two instruments can be indicative of good discriminant validity.
Validity is not a static characteristic of an instrument but can vary according to context
and different sample characteristics. This was described as “situational specificity” by Schmidt
10
and Hunter (1998) and was used to support their hypothesis that there are statistical artifacts
(e.g., sampling error, measurement error) that can cause instruments to demonstrate adequate
validity in one situation and inadequate validity in another. This echoes the concept that
reliability estimates are sample-dependent. Those using an instrument are encouraged to examine
validity evidence for their specific population and purpose. This evidence for validity can be
collected through various means (AERA, APA, & NCME, 2014) including: (a) test content (e.g.,
content items reflect the construct of interest), (b) response processes of test-takers (e.g., looking
at how respondents interpret items), (c) internal structure of the instrument (e.g., examining the
relations among items through factor analysis), (d) consequences of completing the instrument,
and (e) relations to other variables. This last method of examining validity properties includes
convergent and discriminant evidence, test-criterion relations, and validity generalization, which
is described in more detail in a subsequent section.
Establishing the validity of scores produced by an instrument is important as it also
speaks to the validity of theories on the construct being measured and the operational definition
of the construct in question (Haynes et al., 2011; Kimberlin & Winterstein, 2008). The evidence
of validity of a measure continuously evolves as the theories of the constructs being measured
change according to new research developments (Haynes et al., 2011). Thus, evaluating the
validity of measures should be an ongoing process. This most certainly applies to the HTQ,
given the changes in the PTSD construct in the most recent version of the DSM and the cultural
variation of symptoms.
Evaluating the Psychometric Properties of Clinical Instruments
As highlighted, establishing psychometric evidence is an inherent component of the
development and adaptation of assessment instruments. Once the reliability and validity
11
estimates have been calculated, it is then important to determine whether or not the values
obtained are adequate for using and interpreting the data provided by an instrument.
Consequently, several experts have proposed criteria to evaluate assessment tools (e.g., Bickman
et al., 1999; Hunsley & Mash, 2008; Robinson, Shaver, & Wrightsman, 1991) to help clinicians
and researchers determine the suitability of measures for a specific purpose.
Hunsley and Mash’s (2008) criteria can be applied to measures that assess mental health
issues across populations. They developed a rating system that includes nine psychometric
categories: norms, internal consistency, inter-rater reliability, test-retest reliability, content
validity, construct validity, validity generalization, sensitivity to treatment change, and clinical
utility. For each category, the following ratings can be assigned: less than adequate, adequate,
good, excellent, unavailable, and not applicable. A rating of ‘adequate’ indicates that an
instrument attains a minimal level of scientific support, ‘good’ indicates that there is solid
scientific evidence that supports the instrument, and the rating of ‘excellent’ indicates extensive,
high quality evidence supporting the instrument. The authors recommended, for both research
and clinical purposes, only using instruments that fall in the ‘good’ or ‘excellent’ ranges. See
Table 1 for a description of ratings for select psychometric categories. Because the focus of this
dissertation is on internal consistency, convergent validity, and discriminant validity, only the
categories of internal consistency and construct validity are presented. Hunsley and Mash’s
rating system for these categories also served as guidelines when interpreting the results of the
two dissertation studies.
12
Table 1
Criteria for Rating Psychometric Properties of Instruments
Internal consistency
Adequate: Preponderance of evidence indicates alpha values of .70-.79.
Good: Preponderance of evidence indicates alpha values of .80-.89.
Excellent: Preponderance of evidence indicates alpha values of  .90.
Construct validity
Adequate: Some independently replicated evidence of construct validity (e.g.,
predictive validity, concurrent validity, convergent and discriminant validity).
Good: Preponderance of independently replicated evidence, across multiple types of
validity (e.g., predictive validity, concurrent validity, convergent and discriminant
validity).
Excellent: In addition to the criteria used for a good rating, evidence of incremental
validity with respect to other clinical data.
Note. Adapted from “Developing Criteria for Evidence-Based Assessment: An Introduction
to Assessments That Work,” by J. Hunsley and E. J. Mash, 2008, A Guide to Assessments
That Work, pp. 8-9.
Reliability generalization. As reliability estimates vary across samples and different
administrations of a given instrument, it can be difficult to assess its overall reliability properties.
To obtain comprehensive information on reliability of scores for a given instrument, researchers
conduct meta-analyses that combine data from studies that have used the instrument (e.g.,
Churchill & Peter, 1984; Peter & Churchill, 1986). There has been a marked increase in interest
and the practice of conducting meta-analyses on reliability coefficients since Vacha-Haase
(1998) coined the term “reliability generalization” (RG). In general, RG studies are designed to:
(a) determine a mean reliability of scores generated by a test, (b) examine the variability in score
reliability, and (c) characterize the sources of variance by examining the relations between the
reliability estimates and study and sample characteristics. This method is an extension to the
“validity generalization” procedures developed by Schmidt and Hunter (1977) that examines the
relations between scores from a given measurement instrument and other criteria as a means to
evaluate the overall validity properties of the instrument.
13
RG vary considerably in their methodology (López-López, Botella, Sánchez-Meca, &
Marín-Martínez, 2013). In her dissertation, Henchy (2012) summarized RG best practice
guidelines found in the existing literature (e.g., Howell & Shields, 2008; Thompson, 1999;
Warne, 2008). These guidelines are presented in Table 2. Essential recommendations include
important RG practices that have been commonly found and recommended in previous research,
whereas optimal recommendations are somewhat newer practices that should be considered.
Thus, these guidelines can help researchers conduct RG that are consistent with best practices.
They served as a roadmap for the present RG and, with the exception of the use of a power
analysis (because of the availability of scores of studies), are included in the methodology of this
dissertation.
Table 2
Summary of Recommendations for Conducting RG
Essential recommendations
Conduct thorough searches for studies
Determine criteria for inclusion of primary studies
Address file-drawer problem or publication bias
Code both instrument and sample characteristics
Use multiple raters and report inter-rater reliability
Use box and whisker plots or Confidence Intervals to present variability
Optimal recommendations
Do not combine multiple types of reliability
Do not combine multiple subgroups in one analysis
Conduct separate analyses for multiple subscales or instruments
Examine homogeneity of population correlations
Conduct a power analysis before conducting the RG
Note. Adapted from “Review and Evaluation of Reliability Generalization Research”, by A.
M. Henchy, 2012, Unpublished Doctoral Dissertation, pp. 34-36.
RG analyses for the field of psychology can help those who use a given test make
decisions regarding the applicability of the test to a specific group in research settings or to
individuals in clinical settings (Vacha-Haase et al., 2000). Another potential benefit of
14
conducting RG, especially in multicultural research, is its potential effect on public policy and
the development of treatment programs for ethnic minorities (Chun, Organista, & Marín, 2003).
For instance, if an instrument that is being used to assess posttraumatic symptoms in a given
ethno-cultural group has not yet demonstrated adequate reliability properties for this specific
group, the results may be misinterpreted and could have a negative impact on the decisionmaking process regarding access to services. Values generated by RG analyses may provide an
indication of the range of reliability estimates that are applicable for different groups. Thus the
importance of the evaluation of psychometric properties of instruments cannot be
underestimated, and more awareness of the potential impacts of instruments with inadequate or
unknown psychometric information needs to be considered.
Validity generalization. Following their initial observations regarding situation
specificity in validity, Schmidt and Hunter (1977) developed a meta-analytic method called
“validity generalization” (VG) to examine the correlations between a specific test and a criterion.
Specifically, VG is intended to estimate a mean validity coefficient of a measure based on the
correlations between the measure and validity criteria of different studies. For example, to
evaluate the discriminant validity properties of a PTSD questionnaire, one could gather relevant
studies that provide correlations (or other measures of the relations between two variables)
between this questionnaire and a measure of impression management, and then calculate a mean
correlation. This method can also provide indications of the effects that might account for
variability in the values found in a given sample, and is similar to the RG meta-analytic
procedures proposed by Vacha-Haase (1998).
An important issue that needs to be mentioned is the assessment of the strength of the
relation between two variables. In other words, how can one determine if a correlation is strong
15
or weak? Although r-values range from -1 to +1 (-1 = perfect inverse correlation, 0 = no
correlation, +1 = perfect positive correlation), it is difficult to establish the strength of values
across the full range of possible values. However, Cohen (1992) proposed the following r-value
effect-size benchmarks to help with interpreting correlations: .10 = small effect, .30 = medium
effect, .50 =large effect. Although these benchmarks have been widely used in various fields of
research, Hemphill (2003) suggested other benchmarks based on 380 meta-analytic studies in
applied psychological research that are more appropriate for psychology (i.e. <.20 = small effect,
.20-.30 = medium effect, >.30 = large effect).
The VG method has been primarily used in the field of employment selection and testing
(e.g., Dye, Reck, & McDaniel, 1993; Ones, Viswesvaran, & Schmidt, 2003) and it has been
recommended in the Standards for Educational and Psychological Testing as a general approach
to evaluate validity (AERA et al., 2014). According to these standards, VG should preferably be
used when the following conditions are met: a large meta-analytic database of data representing
more or less the type of situation to which the use of scores on the instrument will be
generalized, and when statistical artefacts are corrected and this correction produces validity
evidence that is consistent (AERA et al., 2014).
Cross-Cultural Assessment
A central aspect of cross-cultural assessment is the question of whether test scores can be
interpreted the same way across various cultural groups. Thus, in order for a psychological
assessment instrument to be used among different cultures and sub-groups, many elements need
to be considered. Two concepts that are critical to this issue are bias and equivalence (e.g. He &
van de Vijver, 2012).
16
Bias refers to various elements that can impact the accuracy and validity of instruments
across cultures and it occurs when score differences between cultural groups do not reflect the
actual cultural differences of the underlying construct being measured (van de Vijver & Tanzer,
2004). As He and van de Vijver (2012) summarized in their review, the several types of bias can
be classified as being either construct bias (i.e. a construct being measured is different across
cultural groups), method bias (i.e. bias deriving from the sampling procedures, structure of an
instrument, different response styles or the administration process), or item bias (i.e. an item that
has a different psychological meaning across cultures).
Measurement equivalence implies that a same instrument measures the same construct
across diverse cultural groups. There are several types of equivalence, including: (a) content
equivalence; (b) linguistic and semantic equivalence; (c) conceptual equivalence; (d) scale and
technical equivalence; (e) normative equivalence (Keane et al., 1996). Content equivalence refers
to the notion that the items of a measure reflect the ways that individuals of a culture experience
the phenomenon that is being measured. Using the psychological effects of a traumatic event as
an example, if several items measure the presence of numbing symptoms, but numbing
symptoms are not a typical posttraumatic reaction within the culture in question, the instrument
would be lacking in content equivalence. Linguistic and semantic equivalence ensures that the
grammar, syntax, and meanings of a translated instrument are comparable to the original version.
The most commonly recommended method to ensure this type of equivalence is to use both a
blind-back translation and a consensus approach to determining the phrasing of items (Brislin,
1970, 1986). Conceptual equivalence refers to the idea that a concept is equivalent across
cultures. More specifically, it addresses the question of whether a translated expression assesses
the same aspect or construct in different cultures. Scale and technical equivalence ensures that
17
the methods used to assess a construct are comparable across cultures and yield scores on an
instrument that mean the same thing across groups. Lastly, normative equivalence pertains to the
establishment of normative standards specific to the culture in question, instead of relying on
pre-established norms developed in a different culture.
The assessment of the quality of adapted instruments must take into account the methods
that were used to ensure these types of equivalencies, as well as verify the psychometric
properties of the instrument. A methodological review of 47 instrument translation studies
revealed that there was great variability in the methods used to translate and validate instruments
used for cross-cultural research (Maneesriwongul & Dixon, 2004). In addition, Sperber (2004)
noted that the process of translating and cross-culturally validating instruments is usually not
considered important in research protocols and that the most common translation procedure used
is the simple forward translation.
Many different guidelines and recommendations for the cross-cultural adaptation of
instruments exist, however very few of them have been empirically investigated. A crossdisciplinary review of cross-cultural adaptation guidelines identified 31 of them and found no
consensus in their methodology (Epstein, Santo, & Guillemin, 2015). However, this review
showed that most guidelines included the recommendation of using committees, focus groups,
and back translations. Although the empirical evidence regarding these guidelines is lacking, it
appears that the use of back translations may not have much added value if a consensus approach
is already used (e.g. da Mota Falcão, Ciconelli, & Ferraz, 2003; Epstein, Osborne, Elsworth,
Beaton, & Guillemin, 2013). Although currently there is limited empirical support for these
guidelines, using strategies to help minimize bias remains unquestionably essential in cross-
18
cultural measurement. If the adaptation process is not carefully implemented, the results may be
invalid (Epstein et al., 2015).
Posttraumatic Stress Disorder: An Overview
The psychological responses after experiencing traumatic life events have been of interest
for many centuries. For instance, authors such as Homer and Shakespeare wrote about combat
stress that appears to closely resemble current PTSD symptomatology (Trimble, 1981).
Following these early depictions, such post-trauma reactions became more recognized as
syndromes in the 19th and 20th centuries, with a diverse nomenclature being used, including
spinal concussion, soldier’s heart, traumatic neurosis, and shell shock, to name a few (ParryJones & Parry-Jones, 1994).
Although early editions of the Diagnostic and Statistical Manual of Mental Disorders
(DSM) included disorders that occurred as a consequence of a stressful event such as “gross
stress reactions” (DSM-I: APA, 1952), and “anxiety neurosis/transient situational disturbance”
(DSM-II: APA, 1968), PTSD was first recognized as a distinct diagnosis in the third edition of
the DSM (DSM-III: APA, 1980). Since then, the PTSD diagnosis has undergone many changes
throughout the subsequent DSM editions, with most modifications involving the stressor
criterion, specific symptoms and specifiers. In the DSM-5 (APA, 2013), the PTSD diagnosis was
moved from the “Anxiety Disorders” chapter to a newly created one entitled “Trauma and
Stress-Related Disorders.” The stressor criterion was significantly changed (e.g., Criterion A2
that required that the individual must have reacted to the traumatic event with “fear, helplessness,
or horror” (APA, 2000, p. 467) was removed, and dissociative and preschool subtypes were
added. Because the HTQ symptom scales are based on the 17 symptoms of Criteria B, C and D
19
of the DSM-IV (APA, 1994) PTSD diagnosis, I will specifically focus on these symptoms and
describe them in light of the changes in the current DSM-5 (APA, 2013).
One major change to the PTSD symptoms in the DSM-5 was a new conceptualization of
the factor structure of the disorder. The DSM-IV edition described a three-factor model, or three
symptom clusters: (a) Criterion B: re-experiencing, (b) Criterion C: avoidance/numbing, and (c)
Criterion D: hyperarousal (APA, 1994). However, numerous confirmatory factor analyses
revealed that this three-factor model was not the best to capture the underlying symptoms
structure of PTSD (e.g., Baschnagel, O’Connor, Colder, & Hawk, 2005; Elklit & Shevlin, 2007;
Krause, Kaltman, Goodman, & Dutton, 2007). Although a range of one to seven-factor models
have been proposed, the model that has received the most support in the literature is a four-factor
conceptualization, and the APA sub-group included this one in the DSM-5 (Friedman, 2013).
This model is primarily based on King, Leskin, King, and Weathers’ (1998) conceptualization,
which separates avoidance and numbing into two clusters. The DSM-5 model includes the
following symptom clusters: (a) Criterion B: intrusion (e.g., “recurrent, involuntary, and
intrusive distressing memories of the traumatic event,” APA, 2013, p. 271), (b) Criterion C:
avoidance (e.g., “avoidance of or efforts to avoid distressing memories, thoughts, or feelings
about or closely associated with the traumatic event(s),” APA, 2013, p. 271, (c) Criterion D:
negative alterations in cognitions and mood (e.g., “feelings of detachment or estrangement from
others,” APA, 2013, p. 272) and (d) Criterion E: alterations in arousal and reactivity (e.g.,
“exaggerated startle response,” APA, 2013, p. 272). More recently, a seven-factor hybrid model
that includes re-experiencing, avoidance, negative affect, anhedonia, externalizing behaviours,
and anxious and dysphoric arousal symptom clusters has preliminary evidence of being a
superior fit as compared to all other models (Armour, Müllerová, & Elhai, 2016).
20
The DSM-5 edition contains all the 17 symptoms of DSM-IV, however some
descriptions have been revised and reworded to provide further clarification. Three new
symptoms were also added. One example of an added item is “reckless or self-destructive
behaviour” (APA, 2013, p. 272) in the Criterion E section. This is intended to reflect the fact that
externalizing behaviours such as risk-taking and suicidal behaviours are common in individuals
with PTSD (e.g., Miller, Kaloupek, Dillon, & Keane, 2004; Nock, Hwang, Sampson, & Kessler,
2010). This increase in the number of symptoms to 20 expanded the possible criteria to meet the
diagnosis and, consequently, increased the number of possible symptom profiles. In addition,
although the diagnostic thresholds were not changed (i.e., a minimum of six symptoms is still
needed to meet criteria), they have been redistributed among the new symptom clusters. In other
words, instead of requiring one re-experiencing symptom, three avoidance/numbing symptoms,
and two arousal symptoms to meet full PTSD criteria, an individual must experience one
intrusion symptom, one avoidance symptom, two negative alterations in cognitions and mood
symptoms, and two alterations in arousal and reactivity symptoms.
Studies that have compared the DSM-IV and DSM-5 PTSD diagnostic criteria have
found differences in prevalence rates between the two (e.g., Calhoun et al., 2012; Gentes et al.,
2014; Kilpatrick et al., 2013). A study conducted of a community sample in Wales found that the
DSM-IV prevalence was 14.3%, but only 8% when DSM-5 criteria were applied (White et al.,
2015). The change in prevalence was attributed to the exclusion of DSM-IV stressor criterion
qualifying events such as life-threatening illnesses for DSM-5. Changes in diagnostic rates in
other studies were also attributed to the splitting of avoidance and numbing symptoms into two
distinct clusters (e.g., Gentes et al., 2014).
Studies on the rate of PTSD across different cultures using DSM-5 criteria are limited
21
due to the recency of its publication. Therefore, it is too early to establish the cross-cultural
applicability of the new criteria. However, considerable data regarding the prevalence rates
across world have been gathered with the past versions of DSM. These studies are summarized
in the following section.
Prevalence of trauma and PTSD. It is important to note that the methods used to collect
data, the diagnostic criteria used to evaluate trauma and PTSD (e.g., DSM-IV vs. DSM-5), and
the composition of research samples (e.g., war veterans vs. civilians; refugee vs. non-refugee)
can all have a tremendous impact on the rates of trauma and PTSD reported in the literature.
Thus, it is critical that these factors are kept in mind when interpreting results of epidemiological
studies. There are also three different concepts to be aware of when studying the epidemiology of
trauma: (a) the prevalence of being exposed to traumatic events, (b) the conditional prevalence,
or the risk of having PTSD after being exposed to a traumatic event, and (c) the total prevalence
of PTSD in the population (Norris & Slone, 2014).
With regards to the lifetime prevalence of trauma exposure, studies have shown that
experiencing traumatic events is quite common among general populations across the world. For
example, Kilpatrick and colleagues (2013) found in a national probability study of approximately
3,000 U.S. adults that 89% of participants reported being exposed to at least one DSM-5
qualifying event. One study on the exposure to trauma in four post-conflict, low income
countries/regions (i.e., Algeria, Cambodia, Ethiopia, and Gaza) found that the rates of trauma
according to various trauma categories (e.g., torture, youth domestic stress, and conflict-related
events) varied across countries, but that overall exposure to severe trauma was consistently high
(de Jong et al., 2001). The most recent data collected from the World Mental Health Surveys that
22
were conducted in 24 countries between 2001 and 2012 show rates of lifetime trauma exposure
ranging from 28.6% in Bulgaria to 84.6% in Ukraine (Koenen et al., 2017).
Even though traumatic events are common, the majority of individuals who experience
them do not develop PTSD. For instance, in one study based on DSM-IV criteria, 9.5% of
individuals exposed to trauma developed PTSD (Breslau et al., 1998). In addition, in Kilpatrick
and colleagues’ (2013) study, the conditional probability of lifetime PTSD was found to be only
11.7%. It is interesting to note that although, in general, men are more likely to experience
trauma than are women, women have a higher tendency to develop PTSD than do men (e.g.,
Breslau et al., 1998; Kessler et al., 1995). Another factor that influences the conditional
probability of PTSD is type of trauma experienced. For instance, survivors of violent and
intentional traumas are more likely to develop PTSD than are those who experience
unintentional traumas such as natural disaster or those who are witnesses to a traumatic event
(e.g., Kessler et al., 1995).
Although traumatic events and negative psychological impacts due to trauma are
prevalent in all cultures (Keane et al., 2008), epidemiologists have found varying rates of the
overall lifetime prevalence of PTSD across countries. Studies in the United States have shown
that PTSD is present in 6.4% to 9.5% of the general population (Breslau, Davis, Andreski, &
Peterson, 1991; Kessler et al., 1995; Pietrzak, Goldstein, Southwick, & Grant, 2011). The World
Mental Health Surveys from 24 countries found estimates for lifetime prevalence for PTSD that
ranged from 0.3% in China to 8.8% in Northern Ireland (Koenen et al., 2017). In addition, a
systematic review on the prevalence of mental disorders in refugees who were resettled in
Western countries showed that 9% of them suffered from PTSD (Fazel, Wheeler, & Danesh,
2005). Furthermore, research in areas of conflict or where other mass traumatic events took
23
place, such as natural disasters or terrorist attacks, reveals higher rates. For instance,
approximately 20 months after the December 2004 Indian Ocean tsunami, the prevalence of
PTSD among victims in Sri Lanka was 21% (Hollifield et al., 2008). Similarly, 19.5% of a
sample of rescue and recovery workers in the aftermath of 9/11 continued to screen positive for
PTSD five years after the attacks (Brackbill et al., 2009). Although there are differences in rates
of PTSD among studies, it is apparent that the negative psychological effects of traumatic events
are considerable among people across cultures.
Cultural considerations of trauma. The relation between culture and mental health is
typically seen as one that is highly intertwined (e.g., Sam & Moreira, 2002). Cultural factors that
can influence how mental disorders are experienced and expressed include ethnicity, race,
acculturation, individualism-collectivism, and social acceptance of expressed distress, among
others (Eshun & Gurung, 2009; Yeomans & Foreman, 2009). The degree to which culture plays
a role in psychodiagnostic classification is central to the debate between emic and etic
approaches to psychopathology. The emic approach considers psychopathology to be specific to
individual cultures, as opposed to the etic approach which views psychopathology as universal
with some sociocultural factors influencing symptom expression (Murphy, 1982).
Although most trauma research has been conducted in Western countries, more and more
research has been done with people from non-Western cultures and, as the cross-cultural research
on PTSD develops, more and more questions arise regarding the cross-cultural applicability of
the diagnosis. Specifically, questions about face validity and clinical utility have provoked much
debate (Hinton & Lewis-Fernández, 2011; Summerfield, 1999). Some researchers have argued
that some PTSD criteria, such as flashbacks, are a Western phenomenon (Jones et al., 2003).
Other researchers have gone further, suggesting that the psychological impact of trauma varies so
24
much, and depends on so many factors, that it cannot be standardized into a single disorder or
diagnosis such as PTSD (e.g., Bracken, Giller, & Summerfield, 1995; Summerfield, 2004) and
that local idioms of distress or cultural syndromes are better to describe the specifics of trauma
response. Rather than adopt an either-or stance, a number of researchers have argued that the use
of both the PTSD construct and local idioms of distress are relevant for understanding the
sequelae of trauma (e.g., Miller et al., 2009).
As indicated, there has been evidence of PTSD in numerous cultures. To further
illustrate, studies that have used the same instrument (i.e., Composite International Diagnostic
Interview for DSM-IV-TR [CIDI]; Robins et al., 1988) to assess the presence of PTSD, have
found evidence of the disorder in various countries. For example, the 12-month prevalence rate
of PTSD using the CIDI was 0.4% in Japan (Kawakami et al., 2005), 0.6% in Mexico (MedinaMora et al., 2005), 0.6% in South Africa (Williams et al., 2008) and 3.5% in the U.S. (Kessler et
al., 1995). Although these studies show that the DSM PTSD construct can be found across
cultures, there are significant differences in the prevalence rates reported. Numerous factors may
contribute to these variations, such as the prevalence of trauma exposure and methodological
variables (e.g., lack of equivalency among the various adaptations of the instrument), to name a
few. However, some studies have attempted to control for these factors and have found true
differences in PTSD prevalence rates. For example, Alcántara and colleagues (2013) found
differences after controlling for prevalence of exposure, and de Jong and colleagues (2001)
found differences after ensuring measurement equivalency.
Another factor that may contribute to the variation in the epidemiology of PTSD is the
variation across cultures in the manifestation of trauma symptoms. For instance, it appears that
the salience of avoidance/numbing symptoms may vary across cultures. A study of Kalahari
25
Bushmen found that several avoidance/numbing symptoms were not endorsed among the group
of traumatized individuals, but that the re-experiencing and arousal symptom were (McCall &
Resick, 2003). Other studies among Vietnamese refugees (Norris, Van Landingham, & Vu,
2009), Cambodians (Mollica, Mcinness, Poole, & Tor, 1998), Middle Eastern women (Norris &
Aroian, 2008), and Senegalese refugees (Tang & Fox, 2001) found that avoidance and numbing
symptoms were less salient than symptoms of the other clusters. Symptom variations have also
been found with the re-experiencing cluster, with the presence of distressing dreams or
nightmares being more salient in certain cultural groups, such as American Indians (Shore,
Orton, & Manson, 2009), Alaskan Natives (Palinkas, Petterson, Russell, & Downs, 2004), and
Cambodian refugees (Hinton et al., 2009).
Another example of culturally bound symptom variation is the higher prevalence rate in
the United States of PTSD among Latinos, African Americans, and Native Americans, compared
to Whites. Apart from some variations in prevalence of traumatic exposure (except for Latinos),
other factors that have been hypothesized to contribute to these differences are an overendorsement of PTSD symptoms related to cultural differences in response styles, the effects of
racism and discrimination, possible overlapping symptoms with other cultural idioms of distress,
and a tendency to experience peri-traumatic dissociation (Alcántara et al., 2013; Pole et al.,
2008).
In addition, somatic symptoms have been found to be common reactions to trauma in
certain cultures. For instance, Salvadoran and Senegalese refugees have reported sensations of
bodily heat (Jenkins & Valiente, 1994; Tang & Fox, 2001), Bhutanese refugees who were
tortured reported experiencing bodily heat, and several studies have also shown various somatic
symptoms among Cambodian refugees such as neck pain and orthostatic dizziness (Hinton et al.,
26
2006; Hinton et al., 2008). Therefore, such missing symptoms from the PTSD construct may also
play a role in the variation of prevalence rates.
As reported by Hinton and Lewis-Fernández (2011), studies demonstrating cross-cultural
variations of posttraumatic symptomatology were considered in the development of the new
DSM-5, and the recommended changes to the PTSD diagnosis were based on the results of these
studies. As they noted, the decision to have a threshold of only one symptom for the distinct
avoidance cluster was, in part, based on the lower rate of avoidance reported in certain cultures.
The additions of a new section on culture-related diagnostic issues illustrate other efforts made to
ensure that the DSM-5 classification of PTSD is both culturally sensitive and culturally
appropriate (Lewis-Fernández, Hinton, & Marques, 2014).
In addition to studies examining the PTSD construct across different cultures, there have
been studies that define posttraumatic responses according to more emic approaches. For
instance, some researchers have found distinctive posttraumatic symptom expressions in
different cultures that are quite dissimilar to the symptoms of PTSD. One of these cultural
trauma syndromes is baksbat or “broken courage” in Cambodia (Chhim, 2012), which can
include symptoms such as an inability to trust others, becoming more submissive, feelings of
cowardliness, and becoming deaf and mute. Another example of a cultural syndrome is
ihahamuka among Rwandan genocide survivors which is mainly characterized by a shortness of
breath (Hagengimana & Hinton, 2009).
In summary, researchers are still trying to understand whether, when, or how
posttraumatic psychological symptoms are more indicative of a culturally bound disorder, a
universal construct, or something in between. Although further studies are warranted to clarify
this situation, it is important to remember that these symptoms affect thousands every year and
27
that there is evidence to support the cross-cultural applicability of the PTSD construct (e.g., de
Jong et al., 2001). Accordingly, the development of both appropriate instruments to assess PTSD
and evidence-based treatments to treat the condition should be a high priority.
The assessment of PTSD. Accurate assessment of PTSD can be challenging, as
researchers and clinicians rely mainly on subjective reports of symptoms, and the overlap of
PTSD symptoms with other psychopathologies can be difficult to tease apart (e.g., Brewin,
Andews, & Rose, 2003). Similar to other disorders that have high rates of comorbidity, using
multiple methods to assess trauma symptoms is highly recommended (Keane et al., 1985). These
methods can include not only a thorough clinical history from the affected individual and/or
collateral informants, but also structured diagnostic interviews, self-report measures, and
psychophysiological measures.
Structured diagnostic interviews are often recommended for the assessment of PTSD due
to higher diagnostic accuracy than self-report measures (Keane et al., 1985), and to their
conformity with EBA standards. Although they are commonly used in research settings, they
appear to be less commonly used in clinical settings (except for forensic practices) (Aboraya,
2009; Keane, 1995; Keane, Buckley, & Miller, 2003). Structured diagnostic interviews or
clinician-administered instruments can be sections of a more comprehensive assessment
instrument, such as the PTSD module of the Structured Clinical Interview for DSM-IV (SCIDIV; First, Spitzer, Williams, & Gibbon, 1996), or stand-alone tools such as the ClinicianAdministered PTSD Scale (CAPS; Blake et al., 1990). A 2005 survey of trauma specialists found
that these tools were the two most commonly used clinician administered instruments in clinical
settings and were also among the most frequently used for research purposes (Elhai, Gray,
Kashdan, & Franklin, 2005). In addition, in their review of instruments used to assess PTSD
28
using EBA guidelines, Keane and colleagues (2008) rated the CAPS and the SCID-IV as the
most strongly recommended clinician-administered instruments for diagnosis. The CAPS was
especially recommended after receiving excellent ratings across all psychometric categories
(e.g., norms, internal consistency, content validity, clinical utility), and was also suggested for
use for purposes of treatment monitoring and treatment outcome evaluation (Keane et al., 2008).
Both of these structured interviews have been updated to reflect the changes in the DSM-5 (First,
Williams, Karg, & Spitzer, 2016; Weathers et al., 2013).
Self-report measures can provide information about the trauma experienced by the
individual and about the presence, frequency, and duration of PTSD and other trauma-related
symptoms. Self-report instruments continue to be widely used, as they tend to take less time to
complete and are less costly than structured interview schedules. Based on their survey results,
Elhai et al. (2005) found that the most commonly used self-report measures in clinical settings
were the Trauma Symptom Inventory (TSI; Briere, 1995), the PTSD Checklist (PCL; Weathers,
Litz, Herman, Huska, & Keane, 1993), and the Posttraumatic Diagnostic Scale (PDS; Foa,
Cashman, Jaycox, & Perry, 1997). Two of these self-report measures (i.e., PCL and PDS) were
also Keane and colleagues’ (2008) most highly recommended self-report instruments for
diagnosis, treatment monitoring, and treatment outcome, based on Hunsley and Mash’s (2008)
EBA guidelines.
Another category of PTSD assessments includes psychophysiological measures which do
not involve subjective reports but, rather, examine psychophysiological reactivity to traumarelevant stimuli by measuring different physiological reactions, such as heart rate, skin
conductance, or even cerebral activity through electroencephalography. Although there is
evidence of their clinical utility (Keane et al., 1998), they are not easily accessible for most
29
clinicians as they are often expensive and require special training for their use (Orr, Metzger,
Miller, & Kaloupek, 2004). Psychophysiological measures have also been criticized for their
diagnostic accuracy, as studies have shown than approximately 40% of individuals who are
diagnosed with PTSD do not manifest any physiologic reactivity (Orr et al., 2004).
Although using a multi-method approach to the assessment of PTSD is recommended, it
is not always feasible. For instance, there may be time or financial constraints, especially in
research contexts such as epidemiological surveys where large numbers of participants need to
be assessed. Whether using multiple methods or not, a critical component in choosing an
instrument is examining its psychometric properties for the group that is to be assessed. As
discussed in the next sections, variations in the expression of posttraumatic symptoms across
individuals with differing cultural backgrounds and the cultural utility of an instrument also need
to be considered in selecting and using an instrument.
Assessment of PTSD across cultures. Both the diagnostic criteria of PTSD and
psychological measures of PTSD have mainly been developed in Western and industrialized
countries (Keane et al., 2008). Assuming that this Western-based symptomatology is universal,
that is using a pseudoetic approach to trauma (Renner, Salem, & Ottomeyer, 2007), may fail to
take into account the culture-specific idioms of distress and, therefore, might not be considered a
truly evidence-based approach to assessment. In other words, a pseudoetic approach to the
assessment of PTSD may increase the risk of measurement bias.
When it comes to using evidence-based trauma assessment instruments among different
cultures, there are three options available (van de Vijver & Tanzer, 2004): (a) conducting a literal
translation into a target language; (b) culturally adapting an existing instrument; or (c)
developing a new one that is tailored for a specific culture. Although some researchers have
30
chosen this last option (e.g., Miller et al., 2006), developing an entirely new measure can be very
costly and time-consuming, and would limit cross-cultural comparisons. It is also not feasible to
develop standardized measures for every existing culture and sub-culture.
Most PTSD instruments found in the literature that have been used in diverse cultural
groups were literal translations. For instance, the CAPS has been translated into several
languages such as Bosnian (Charney & Keane, 2007), Swedish (Paunović & Öst, 2005), and
German (Schnyder & Moergeli, 2002). With regards to self-report instruments, the PCL has been
translated into various languages including, but not limited to, Spanish (Miles, Marshall, &
Schell, 2008), Chinese (Wang, Su, Bi, Wei, & Mo, 2012), and Malay (Bahari, Alwi, Ahmad, &
Saiboon, 2015). Although these translated instruments have evidence of rather good
psychometric properties for their specific sample (e.g. Bahari et al., 2015), they may not fully
capture all the symptoms related to posttraumatic stress within the target culture by not including
culturally relevant idioms of distress.
The only PTSD instrument that was specifically designed to be culturally adapted and
used across diverse cultural samples is the Harvard Trauma Questionnaire (Mollica et al., 1992).
This option, that is incorporating an emic approach by including culturally relevant constructs to
an existing instrument, appears to be more efficient and has long been proposed as yielding the
most accurate assessment possible (e.g., Berry, 1989). However, the overall psychometric
properties of the HTQ have not been evaluated to support this claim. Indeed, adapting crosscultural instruments for the assessment of posttraumatic stress symptoms comes with many
challenges and many factors need to be considered to optimize their measurement accuracy. For
adapted instruments that have already been used in research, a meta-analytic examination of the
psychometric properties of these instruments can provide essential information regarding their
31
characteristics and likely value when used in subsequent research. In addition, because the
assessment of traumatized individuals can have major clinical and societal implications (e.g.,
access to treatment, obtaining disability compensation, obtaining refugee status), an in-depth
analysis of instruments designed for use in these evaluations is of the utmost importance.
The Harvard Trauma Questionnaire
The Harvard Trauma Questionnaire (HTQ) was originally developed by the Harvard
Program in Refugee Trauma (HPRT) and the Indochinese Psychiatry Clinic in Massachusetts
after years of extensive research and clinical experience with refugee populations (Mollica,
McDonald, Massagli, & Silove, 2004). After unsuccessful attempts at assessing trauma with an
Indochinese population using another standard instrument, staff at the clinic decided to create the
HTQ by following the format of the Indochinese versions of the Hopkins Symptom Checklist-25
(HSCL-25; Mollica, Wyshak, de Marneff, & Lavelle, 1987), an instrument that assesses
symptoms of anxiety and depression. The HTQ was developed as a cross-cultural, clinicianadministered instrument to assess trauma and torture related to mass violence and their
psychological impacts. It was intended to be used with clinical and community refugee
populations, in both research and clinical settings (Mollica et al., 2004). Although the developers
initially recommended its use for refugee populations, they have also used the HTQ among nonrefugees (e.g., Silove et al., 2007).
The HTQ was the first cross-cultural trauma assessment instrument to be created and
validated across various cultural groups. The first versions of the HTQ were validated among
three Indochinese refugee populations: Cambodian, Vietnamese, and Lao (Mollica et al., 1992).
Mollica and colleagues (1994) specified that the HTQ should not simply be translated into
another language and then administered to traumatized individuals. Instead, they recommended a
32
rigorous adaptation and revision process supported by detailed knowledge of the culture, relevant
life events of those who would complete the measure, and culture-specific symptoms for each
new form of the HTQ and for each distinct group of trauma and torture survivors. In the HTQ
manual, Mollica and collegues (2004) described evidence-based approaches to the cross-cultural
adaptation of instruments that can help researchers and clinicians develop their own versions of
the questionnaire. These adaptation procedures included essential components of cross-cultural
adaptation such as guidance on establishing a group of experts to consult on the adaptation
process, gathering qualitative information on cultural expressions of distress, translating items,
and evaluating cross-cultural equivalency. They particularly noted the importance of crosscultural equivalency by describing Flaherty et al.’s (1988) five equivalence dimensions (i.e.,
content, semantic, technical, criterion, and conceptual equivalence), which are similar to the ones
described in the previous section. They also emphasized the use of focus groups to gain a better
understanding of the cultural meaning and symptomatology of traumatic experiences of a
specific group.
Systematic reviews of the literature have shown that the HTQ is commonly used in
studies among refugee populations (Gagnon et al., 2004; Hollifield et al., 2002), and can be
useful to exclude PTSD non-cases in specific forcibly displaced populations (Nakeyar & Frewen,
2016). The HTQ has also been recommended as a tool to be used across cultures by trauma
experts (e.g. Keane et al., 2008). However these assertions were not supported by sufficient
empirical evidence that would justify qualifying it as a culturally sensitive measure of PTSD
symptoms. Nor have evidence-based assessment guidelines such as Hunsley and Mash’s ever
been used to evaluate the HTQ and support these claims. Because of these factors, it is
imperative that a thorough review of its psychometric properties be conducted. The results of a
33
review could provide both evidence for the scientific status of the HTQ and also an indication of
what is possible for the cultural adaptation of PTSD measures in general.
In the following paragraphs I will describe the structure of the HTQ and the range of
populations with which it has been used. Although the different sections of the instrument vary
among the various versions of the HTQ, four general sections can be found in most versions.
Because of the need for the instrument to be culturally appropriate, the number and content of
items of each section can also vary across versions. Part I includes items that describe a range of
traumatic experiences, such as “combat situation” and “forced separation from family members.”
There are four possible responses for each item: (a) experienced, (b) witnessed, (c) heard about,
and (d) no. The respondent is asked to indicate which of these responses (one or more) best
represent his or her level of exposure to the trauma. Part II includes two open-ended questions
related to the subjective experience of the respondent’s most traumatizing event. Part III inquires
about direct head injury and other events that could possibly lead to brain damage (e.g.,
starvation, suffocation). Part IV assesses posttraumatic symptoms, and it is this part that is the
focus of the dissertation studies.
In Part IV, the first 16 items are derived from the DSM-III-R/IV PTSD criteria (APA,
1987, 1994) and are the same in every version of the HTQ. I refer to these 16 items as the HTQ16 throughout the dissertation. The items represent the criteria for the intrusion/re-experiencing,
avoidance/numbing, and hypervigilance/arousal symptom clusters. Sample items include:
“recurrent thoughts/memories,” “hard to concentrate,” and “can’t feel emotions.” The
respondent is asked to answer each item according to the following scale: 1 = “not at all,” 2 = “a
little,” 3 = “quite a bit,” 4 = “extremely.” They are also asked to evaluate how much each
symptom bothered them within the past week. The next set of PTSD symptom items are culture-
34
specific questions and, therefore, are tailored for each different version. For instance, the Arabic
version that was validated among Iraqi refugees in the USA (Shoeb, Weinstein, & Mollica,
2007) consists of a total of 45 trauma symptom items and includes unique cultural items such as
“Nafseetak ta’bana” or tired soul. As a further example, the Japanese version developed for
earthquake survivors includes the cultural item “Disappointment upon awakening that life is not
better,” which is not found in other versions. The English version of the questionnaire includes
14 cultural symptoms that are based on clinical observations among Indochinese refugees
resettled in the USA (i.e., literal translation of the Indochinese versions). This version has been
commonly used in the empirical literature.
Scoring the trauma symptoms part of the HTQ includes calculating the mean item score
for both the HTQ-16 and the full symptom scale (i.e. 16 DSM symptom items and the cultural
symptom items). In both cases the mean item score is obtained by summing the individual item
scores and then dividing by the number of items, yielding a mean item score for each scale. The
range of possible scores for each scale is 1 to 4. To establish clinical caseness, Mollica and
colleagues (2004) proposed two methods: (a) cut-off scores and (b) an algorithm approach. They
provided instructions as to how to calculate a cut-off score for a specific group of respondents for
three purposes: (a) screening, (b) clinical care, and (c) research. Although the community sample
cut-off scores of 2 (Silove et al., 2007) and clinical cut-off score of 2.5 (Mollica et al., 1992)
were established in some studies, the developers advised users to establish caseness benchmarks
for each new population. In situations where the usual cut-offs are not applicable, and the
establishment of new benchmarks is not possible, an algorithm approach is preferred. This
approach is based on the DSM-III-R/IV criteria and it conforms to the well-established DSM
guidelines for caseness. Specifically, to meet criteria for the diagnosis of PTSD, an individual
35
needs to score a minimum of 3 on one or more items of the intrusion/re-experiencing subscale,
on three or more items on the avoidance/numbing subscale, and on two or more items of the
hypervigilance/arousal subscale.
The developers of the original English and Indochinese versions of the HTQ have
developed and validated other versions including: revised Cambodian (HPRT, 2000), revised
English (HPRT, 1998), Bosnian (HPRT, 1998), Croatian (HPRT, 1998), Japanese (HPRT, 1996),
Peruvian (HPRT, 2000), Kurdish (HPRT, n.d.), Arabic (HPRT, 2006), Spanish (HPRT, 2000),
and Russian (HPRT, n.d.). The HTQ has also been translated and adapted for use in many other
languages and for different cultural groups by other researchers (e.g., de Fouchier et al., 2012;
Kleijn, Hovens, & Rodenburg, 2001), has been commonly used in research with refugees
(Gagnon et al., 2004; Hollifield et al., 2002), and has been shown to be useful to exclude PTSD
non-cases in certain forcibly displaced populations (Nakeyar & Frewen, 2016).
Although the measure is considered a gold standard for the cross-cultural assessment of
posttraumatic stress symptoms, the psychometric properties of the HTQ symptom scales have
only been examined in detail in a few studies beyond what has been reported in the instrument
manuals. For example, Mollica and colleagues (1992) conducted a study on a sample of 91
Southeast Asian refugee patients (from Cambodia, Laos, and Vietnam) attending a psychiatric
outpatient clinic. Good reliability properties were demonstrated for the 30 item symptoms scale
(referred as the HTQ-30): interrater reliability for the trauma-related symptoms (k = .98); scale
test-retest reliability with a 1-week interval (r = .89); and internal consistency (Cronbach α =
.90). The sensitivity was found to be 78% and the specificity 65%, which indicate that the
measure was better at identifying individuals within these cultural groups who have PTSD than
at correctly identifying those who do not. Although this study showed evidence that the HTQ has
36
reasonably good psychometric properties for this population, the sample size was relatively small
and the authors did not report the psychometric properties for the three cultural groups
separately. Other investigators (Kleijn et al., 2001) examined the evidence of the psychometric
properties of translations of the HTQ-16 in five other languages: Arabic, Farsi, Serbo-Croatian,
Russian, and English. They reported that the internal consistencies of the different versions
ranged from .74 to .89. This study also found that some of the items on the PTSD scale had low
item-total correlations, which could potentially compromise the score reliability of the measure
when used in other samples.
Some authors have suggested that the validity properties of the HTQ symptom scales are
limited and that it may not be generalizable due to having been developed from a psychiatric
outpatient population (Hollifield et al., 2002). In addition, it appears that the Cambodian version
of the HTQ-16 may not be efficient at identifying true cases of PTSD among a non-refugee
Cambodian community sample (Silove et al., 2007), although Mollica and colleagues (1992)
reported more encouraging findings on this point. In a more recent study, Rasmussen and
colleagues (2015) evaluated aspects of the construct validity of the HTQ-16 by examining its
measurement invariance among 878 survivors of torture or other human rights violations who
belonged to various cultural groups. Their results showed that the basic content validity, or
overall structure of the PTSD construct as measured by the HTQ, was acceptable. However,
there were significant response-style differences across cultures, as well as variations in the
clinical thresholds of PTSD. Another study on the factor structure of PTSD using the HTQ-16
compared three models across three groups living in non-Western low- and middle-income
countries (i.e. torture survivors in Northern Iraq, sexual violence survivors in the Democratic
Republic of Congo [DRC], and Burmese refugees in Thailand) (Michalopoulos et al., 2015). The
37
factorial models examined were: (a) 3-factor DSM-IV model, (b) 4-factor “numbing” model, and
(c) 4-factor “dysphoria” model. The results showed that all models had an adequate fit for the
DRC and Burma samples. The 4-factor “numbing” model was the best fit for the DRC sample
and the 4-factor “dysphoria” model for the Burma sample. None of the models had an adequate
fit for the Northern Iraq sample. These important findings speak to the potential limitations of
construct validity of the HTQ-16. Further research is also needed to evaluate other dimensions of
its construct validity, such as convergent and discriminant validity.
Potential Moderators of Reliability and Validity
As mentioned previously, the data provided by an assessment instrument can be more
reliable and valid for some populations than others. To obtain a clearer picture of the
heterogeneous quality of this variance, it is important to examine its potential causes. In the
event that the reliability and validity coefficients produced by the current RG and VG analyses
show signs of heterogeneity (i.e. significant Q statistic), moderator analyses were conducted to
assess the potential impact of sample and methodological characteristics on scores produced by
the HTQ symptom scales, thus further examining the potential sources of variance. Commonly
reported variables that have shown mean differences between groups with and without PTSD, or
that have often been found to moderate reliability or validity estimates, were examined.
The sample characteristics examined as moderators included: (a) culture/ethnicity, (b) cultural
context of the country of origin (individualist vs. collectivist), (c) type of trauma experienced, (d)
being forcibly displaced, (e) age, and (f) gender. These moderator analyses were exploratory in
nature due to the limited information available regarding their impact on the reliability and
validity of HTQ scores.
As described in detail throughout the introduction of this
dissertation, there is evidence of the variation of PTSD symptomatology across cultures and
38
ethnicities (e.g., Kessler et al., 1999; Pole et al., 2008). Because of these findings and the fact
that the studies using the HTQ were conducted across many cultures and ethnicities, this factor is
considered a potential source of variance in the reliability and validity of HTQ scores.
As an extension of this potential moderator, the cultural background of the sample in
terms of the construct of collectivism versus individualism was examined for its possible
moderating effects. Individuals in collectivist cultures tend to be interdependent within their
reference group (e.g. family, community, tribe), give more importance to the goals of this group,
and behave according to the norms established within the group (Mills & Clark, 1982). On the
other hand, people from individualist cultures are mostly independent from their reference group,
give high priority to their own personal goals and behave according to their personal attitudes as
opposed to the norms of their reference group (Triandis, 2001). Some studies have shown that a
cultural group’s orientation (i.e. collectivist or individualist) can have an impact on individual
response styles on Likert-type questionnaires (e.g. Harzing, 2006), which may impact symptom
endorsement. Specifically regarding posttraumatic stress, research has shown that individuals
that come from more independent cultures and who have developed PTSD are more likely to
have a change in self-definition and personal identity following a trauma that individuals from
more interdependent cultures (Jobson & O’Kearney, 2008). In addition, Elsass (2001) found that,
among South American cultures, individualist cultures tended to view PTSD avoidance
symptoms as maladaptive coping, whereas collectivist cultures tended to consider avoidance
symptoms to be adaptive. Such differences in perspective may have an impact on symptom
reporting, and consequently, on score reliability. Because this qualitative study included a small
sample, it only offers preliminary findings. Larger quantitative studies are needed to further
examine the role of collectivism and individualism on the psychological effects of trauma. To
39
categorize cultures as collectivist or individualist in the current study, Hofstede’s dimensions of
national culture was used (Hofstede, Hofstede, & Minkov, 2010). As a result, only the countries
in Hofstede’s research studies were categorized and included in the moderator analyses using
this cultural distinction.
Another variable that may contribute to the variance of the psychometric evidence of the
HTQ is the type of trauma experienced by research participants. This variable was considered
because the prevalence rate of PTSD varies according to the type of trauma experienced by
individuals. For instance, the National Comorbidity Survey revealed that rape, childhood
physical abuse, and childhood neglect were the traumas the most related to PTSD (Kessler et al.,
1999). In a more recent study, the probability of developing PTSD after experiencing a sexual
assault was 40.2%, as compared to 9.6% for severe accident, 9% for unexpected death of
someone close, and 5.1% for natural disaster (Breslau, Troost, Bohnert, & Luo, 2013). The types
of traumas in the RG and VG data were merged into two broader categories (i.e., intentional
trauma and non-intentional trauma). Studies have shown that individuals who are exposed to
intentional traumas had worse health outcomes than did individuals who experienced traumas
that were not deliberate, such as natural disasters or accidents (Lange et al., 2003; Matthieu &
Ivanoff, 2006; Van der Velden et al., 2006).
Being a forcibly displaced person (FDP), which includes refugees, asylum seekers and
the internally displaced was also considered as a potential moderator. These individuals flee their
home to escape traumatic events such as war, civil conflict, or persecution. The impact of these
pre-migration stressors can also be compounded by resettlement difficulties or other postmigration problems such as detention (Ichikawa, Nakahara, & Wakai, 2006; Porter & Haslam,
2005). Researchers tend to find higher rates of psychological difficulties among individuals who
40
have survived war and have been displaced as compared to non-refugee groups (Fazel et al.,
2005; Porter & Haslam, 2005; Steel et al., 2009). However, research has also found that
estimates of PTSD prevalence among refugee populations can vary greatly, from 4% (Hauff &
Vaglum, 1994) to 86% (Carlson & Rosser-Hogan, 1991).
Age was considered a relevant potential source of variance as some studies have found
that rates of PTSD can differ according to age. For instance, Kessler and colleagues (1999) found
in the National Comorbidity Survey in the USA that there were age differences in the prevalence
of PTSD according to gender. Among women, there was no significant association between age
and PTSD. Among men, there was a significant association between age and PTSD, however,
there was no association after controlling for trauma exposure or risk of PTSD after trauma
among the different age cohorts. Furthermore, analyses conducted on data from this survey
indicated that the conditional risk for PTSD declined as age increased (Kessler et al., 1995;
Norris, 1992).
A final sample characteristic that was considered as a potential moderator is gender, as it
has also been frequently studied as a risk factor for the development of PTSD and may play a
role in any heterogeneity of variance. Overall, the main findings of epidemiological research
show that, although women tend to experience fewer traumatic events, they are more likely to
develop PTSD than are men (Breslau et al., 1998; Perkonigg, Kessler, Storz, & Wittchen, 2000).
More specifically, women had a significantly higher prevalence of PTSD than did men in large
national community surveys in the US (Kessler et al., 1999; Pietrzak et al., 2011). These findings
have also been replicated in other parts of the world (e.g., Creamer & Parslow, 2008; Ditlevsen
& Elklit, 2010).
The methodological variables included as potential moderators were: (a) type of sampling
41
used (general community sample; sample seeking mental health services; sample seeking
medical treatment), (b) the study sample was one for which the HTQ was validated, and (c) type
of adaptation procedure used. The moderator analyses of the type of sampling variable were
exploratory, and the analyses of the other two variables were hypotheses driven.
The sites from which the samples were recruited were examined. Psychometric properties
of instruments can be dependent on the severity of the disorder measured (Haynes et al., 2011)
and, as symptoms are generally more severe in individuals who are seeking mental health
services or medical treatment than they are in individuals from non-clinical samples, the effect of
recruitment site (i.e., clinical, community, or mixed clinical and community) were considered.
In previous sections of this introduction I discussed the important notions of equivalency
when adapting instruments for different populations or purposes. Although Mollica and
colleagues (2004) also explicitly highlighted this notion and provided recommendations with
regards to adaptation procedures for other researchers in their manual, early in the article search,
I noticed that many authors used original versions of the HTQ with populations that had not
undergone a previous validation process. To illustrate, Gorst-Unsworth and Goldenberg (1998)
used the HTQ symptom scale with a sample of male Iraqi refugees resettled in the United
Kingdom. When they described their use of the HTQ, they stated only that it was administered in
Arabic and Kurdish with the help of an interpreter. There was thus no indication that the
researchers followed any adaptation guidelines. Internal consistency and convergent validity
estimates should be highest in studies involving samples for which the HTQ was originally
intended to be used, and discriminant validity estimates lower in such samples. Accordingly,
moderator analyses were conducted to determine whether having a study sample that was
culturally similar to the samples used in the development of the HTQ affected reliability and
42
validity estimates.
The Current Studies
The assessment of PTSD is a rather challenging endeavour due to the continuously
evolving conceptualization of the construct and the absence of clear differentiation from other
disorders. The cross-cultural variations in posttraumatic symptom expression further add to this
complexity, thus making the assessment of PTSD across diverse cultural groups an even more
challenging task. The various sources of cultural bias and measures to address them further
complicate the development and validation of instruments, making the evaluation of the
psychometric properties of these instruments highly imperative.
The available evidence regarding the psychometric properties of the HTQ symptom
scales is inconsistent, and its overall psychometric characteristics across populations remain
unknown. Because this instrument is intended to be used across cultures, a comprehensive
analysis of the reliability and validity properties of this instrument across different populations is
warranted. The current studies are the first attempt at aggregating empirical findings across
studies and cultures to assess the cross-cultural applicability of the HTQ in light of evidencebased assessment guidelines. The results from these RG and VG meta-analyses will provide
pertinent information that will help clinicians and researchers decide whether or not it is an
appropriate assessment tool for their clients/participants of different cultural backgrounds.
43
Study 1
The Harvard Trauma Questionnaire: A Reliability Generalization Study
44
Abstract
We conducted reliability generalization (RG) meta-analyses of The Harvard Trauma
Questionnaire (HTQ; Mollica et al., 1992), a widely used cross-cultural instrument used to assess
trauma and its psychological sequelae. A search of all publications and dissertations that used the
symptom scales of the HTQ revealed that of the 384 studies considered for inclusion, only 44%
of published studies and dissertations reported internal consistency estimates of their sample. In
addition, only 16% of samples that would have required a cultural verification or adaptation of
the scales were indeed administered a culturally adapted version. The random-effects RG
included 183 reliability coefficients from 95 studies, representing data from 33,376 individuals.
We characterized the variance of score reliabilities by considering sample and methodological
variables that are predictive of the reliability coefficients. The mean reliability coefficients
ranged from α = 0.74, p < 0.001, 95% CI [0.72, 0.77] for the re-experiencing subscale to α =
0.93, p < 0.001, 95% CI [0.92, 0.94] for the HTQ-30. Overall, the HTQ-16 and HTQ-30 are
likely to provide reliable scores across a large diversity of populations. However, the reexperiencing, avoidance and arousal subscales should be used with caution due to lower overall
mean reliability estimates and inadequate performance in certain samples. We discuss the
importance of following journal article reporting standards and carefully implementing
adaptation strategies in cross-cultural assessment.
Keywords: posttraumatic stress disorder, assessment, cross-cultural assessment, reliability, metaanalysis
45
The Harvard Trauma Questionnaire: A Reliability Generalization Study
Although traumatic events and posttraumatic stress disorder (PTSD) are considered
prevalent in all cultures (Keane, Silberbogen, & Weierich, 2008), epidemiological research has
found varying rates of the overall lifetime prevalence of PTSD across countries. Studies in the
United States have shown that PTSD is present in 6.4% to 9.5% of the general population
(Breslau, Davis, Andreski, & Peterson, 1991; Kessler, Sonnega, Bromet, Hughes, & Nelson,
1995; Pietrzak, Goldstein, Southwick & Grant, 2011). Based on data from the World Mental
Health Surveys in 24 countries, the World Health Organization reported estimates for lifetime
prevalence for PTSD ranging from 0.3% in China to 8.8% in Northern Ireland (Koenen et al.,
2017). In addition, a systematic review on the prevalence of mental disorders in refugees who
were resettled in Western countries showed that 9% of them suffered from PTSD (Fazel,
Wheeler, & Danesh, 2005). Even though there are differences in rates of PTSD across studies,
there is no question that the negative psychological effects of traumatic events are considerable.
Although the diagnostic criteria have changed across editions of the Diagnostic and
Statistical Manual of Mental Disorders (DSM) and the International Classification of Diseases
(ICD), PTSD is commonly associated with symptoms such as re-experiencing, avoidance,
numbing and hyperarousal. Despite the widespread occurrence of PTSD, there has been some
criticism surrounding the cross-cultural applicability of the diagnosis. Some investigators have
regarded some diagnostic criteria as reflecting Western cultural constructions and not being
generalizable across cultures (e.g., Jones et al., 2003). For example, avoidance/numbing
symptoms may be less pronounced in some cultures (Dyregrov, Gupta, Gjestad, & Mukanoheli,
2000; Marsella, Friedman, Gerrity, & Scurfield, 1996; McCall & Resick, 2003).
As psychodiagnostic assessments rely heavily on the conceptualization of the construct they are
46
intended to measure, it is not surprising that there has also been much disagreement about the
cross-cultural applicability of screening tools used to assess PTSD. Psychological measures of
PTSD have mainly been developed in Western and industrialized countries (Keane et al., 2008).
Assuming that the symptomatology is universal (i.e., using an etic approach to trauma
assessment) fails to take into account the culture-specific idioms of distress. Clearly, it is not
feasible to develop standardized measures for every existing culture and sub-culture, but
including culturally relevant constructs in the assessment is ideal for more accurate measurement
(i.e., incorporating an emic approach) (e.g. Renner, Salem, & Ottomeyer, 2007). Researchers
who choose to use an instrument with different cultural groups must be cognizant of the
increased risk of various types of measurement bias, such as construct, method (sample,
instrument, response styles, administration) and item bias (He & van de Vijver, 2012) and adapt
the instrument accordingly. Establishing the various types of equivalence such as content,
linguistic, semantic, conceptual, scale, technical, normative, and cross-cultural equivalencies can
help minimize bias in the cross-cultural adaptation process (Keane, Kaloupek, & Weathers,
1996). Because of these complexities in cross-cultural measurement that can impact accurate
assessment, an examination of the psychometric properties of instruments that are intended to be
used across diverse samples is of utmost importance.
The Harvard Trauma Questionnaire
The Harvard Trauma Questionnaire (HTQ) was developed by Mollica and colleagues
(1992) in response to the need for culturally sensitive trauma measures. The HTQ is a clinicianadministered questionnaire that is intended to be adapted and used across cultures, and has been
validated by its developers for various cultural and linguistic groups (Shoeb, Weinstein, &
Mollica, 2007). Most versions of the HTQ consist of four parts. Part 1 measures different
47
traumatic events, part 2 is an open-ended description of the most traumatic event experienced by
the respondent, part 3 evaluates the circumstances surrounding possible head injury and, finally,
part 4 is a list of trauma symptoms. The first 16 items (HTQ-16) of this last part are derived from
the DSM-III-R/IV (American Psychiatric Association [APA], 1987, 1994) PTSD criteria, and
they are the same in every version of the HTQ. A second set of trauma symptom items in part 4
are culture-specific questions and are tailored for each version of the HTQ. The HTQ has also
been translated and adapted in many other languages by other researchers (e.g., Kleijn, Hovens,
& Rodenburg, 2001). Systematic reviews of tools used to assess the health of refugees have
found that the HTQ has been extensively used by researchers to assess trauma and its sequelae
(Gagnon, Tuck, & Barkun, 2004; Hollifield et al., 2002), and can be useful to exclude PTSD
non-cases in certain forcibly displaced populations (Nakeyar & Frewen, 2016). In addition, in
their review of PTSD measures, Keane et al. (2008) recommended the use of the HTQ for the
assessment of PTSD across cultures. They described the HTQ as having “linguistic equivalence
across the many cultures and languages with which it has been used thus far” (p. 297). However,
they did not provide sufficient empirical evidence to support the cross-cultural applicability of
the instrument or evaluate it according to established evidence-based assessment guidelines, such
as Hunsley and Mash’s (2008).
Although the measure is frequently used and appears to have a reputation of being a good
example in terms of cross-cultural trauma assessment, the psychometric properties of the HTQ
have only been examined in a small number of studies. For example, Mollica and colleagues
(1992) conducted a study on a sample of 91 Southeast Asian refugee patients (from Cambodia,
Laos, and Vietnam) attending a psychiatric outpatient clinic. Evidence of good reliability of the
HTQ-30 (16 DSM PTSD items and 14 Indochinese cultural items) was demonstrated in this
48
sample, including interrater reliability (k = .98), test-retest reliability (r = .92), and internal
consistency (Cronbach α = .96). With respect to the DSM-III-R PTSD diagnosis based on semistructured clinical interviews, sensitivity was found to be 78% and the specificity was 65% for
the HTQ-16. When the 14 cultural symptom items were added, the sensitivity remained the
same, whereas the specificity increased to 72%. The inclusion of the cultural symptoms thus
improved the ability of the scale in accurately detecting PTSD. Although this study showed that
the HTQ symptom scales had reasonable psychometric properties with this sample, the authors
did not report the reliability coefficients for the three cultural groups separately and the total
sample size was relatively small. Kleijn and colleagues (2001) examined the psychometric
properties of translations of the HTQ in five languages: Arabic, Farsi, Serbo-Croatian, Russian,
and English. They reported that the internal consistencies of the different versions of the HTQ-16
ranged from .74 to .89. This study also showed that some of the items on this scale had low itemtotal correlations, which could potentially compromise the score reliability of instrument when
used with other samples. Further to this point, the measurement invariance of scores on the
HTQ-16 across various populations has been found to be questionable (Rasmussen, Verkuilen,
Ho, & Fan, 2015).
Although these studies showed that the scores obtained on the HTQ were, at minimum,
adequately reliable according to Hunsley and Mash’s (2008) benchmarks, its overall
psychometric properties with different cultural groups remains unknown. Because this
instrument is intended to be used across cultures, a further analysis of studies that have used the
HTQ in different populations is warranted. Aggregating empirical findings across studies and
cultures will provide pertinent information regarding the potential cross-cultural utility of the
49
HTQ, thereby helping clinicians and researchers decide whether it is an appropriate assessment
tool for use with individuals of different cultural backgrounds.
Reliability and Reliability Generalization
Reliability, or the precision of a measure, represents “the consistency of the scores across
instances of the testing procedures” (American Educational Research Association, American
Psychological Association, & National Council on Measurement in Education, 2014, p. 33). In
other words, reliability estimates represent the proportion of variance that is explained by the
true score itself (i.e., characteristics measured in the test), as opposed to the part of the variance
that is explained by sampling error or measurement error (Graham, Yenling, & Jeziorski, 2006).
As the reliability of test scores also sets a limit on the possible validity estimates of an
instrument, examining the degree of consistency of scores is imperative in determining if the
instrument is indeed scientifically sound.
It is also important to note that reliability refers to the properties of test scores of a
specific sample, and not to the properties of an instrument itself. Numerous factors, including
sample characteristics and the context in which the assessment occurs, can influence the degree
of consistency of test scores. For instance, the same instrument can produce scores with differing
reliability when administered to more or less homogenous samples (Thompson, 1994). In sum,
reliability is not only dependent on the quality of a measure, but also the characteristics of the
sample, sample size, and administration and scoring procedures (Barnes, Harp, & Jung, 2002).
Researchers often assume that reliability estimates of prior studies are directly
generalizable to data collected in their own studies (Vacha-Hasse, Kogan, & Thompson, 2000).
The common error of referring to the reliability of an instrument as an inherent characteristic of
the instrument may be responsible for researchers ignoring examining the score reliability of
50
their sample, which may then lead researchers to unwittingly interpret unreliable data (Graham,
Diebels, & Barnow, 2011; Henson, Kogan, & Vacha-Haase, 2001). This tendency to generalize
the reliability properties across studies has been called “reliability induction” by Vacha-Haase et
al. (2000). These authors argued that reliability induction is only plausible if the previous
samples have similar sample composition and score variability to the current set of participants.
Even when samples may seem to be comparable, examining score reliability is still important as
there may be subtle differences between the samples.
As reliability estimates vary across samples and different administrations of a given
measure, Vacha-Haase (1998) employed a meta-analytic method called “reliability
generalization” (RG) to examine the likely generalizability of reported score reliabilities of a
specific measure across studies. When conducting a RG, all studies that have provided a
reliability estimate of their sample are retrieved and coded. This collection of estimates is used to
produce a mean reliability score and to examine the variability of score reliabilities. When the
variability across estimates is significant, the RG can also be used to identify which sample and
study characteristics account for the variations in score reliability for the measure (e.g., sample
size, country of study, mean age of sample). RG results such as the mean reliability estimate,
confidence intervals around the mean reliability estimate, and the identification of moderator
variables that account for variations in score reliability provide useful information to clinicians
and researchers when selecting an assessment tool appropriate for a specific task.
The Current Study
Considering the wide range of score reliabilities for the HTQ symptom scales reported in
development and validation studies of the measure, the present research was designed to evaluate
the extent to which the original and adapted versions of the HTQ symptom scales (and subscales)
51
produced reliable scores in the research literature. The internal consistencies of these scales,
across samples, were investigated by using a reliability generalization analysis. This metaanalytic method determines the average reliability of scores obtained on the HTQ symptom
scales and subscales, and also accounts for variability in score reliability across samples through
moderator analyses.
Specifically, moderator analyses were conducted to assess the potential impact of sample
and methodological characteristics on scores produced by the HTQ symptom scales and
subscales, thus further examining the potential sources of variance. Commonly reported
variables that have shown mean differences between groups with and without PTSD, or that have
often been found to moderate reliability estimates, were examined. With regards to sample
characteristics, age was considered a relevant potential source of variance as some studies
suggest that the rates of developing PTSD after exposure to a traumatic event may decline as age
increases (Kessler et al., 1995; Norris, 1992). Gender was also included as a moderator, as
epidemiological research shows that, compared to men, women tend to experience fewer
traumatic events but are more likely to develop PTSD (Breslau et al., 1998; Perkonigg, Kessler,
Storz, & Wittchen, 2000).
Because a main purpose of this study was to evaluate the cross-cultural applicability of
the HTQ, and considering that there is evidence of the variation of PTSD symptomatology across
cultures (e.g., Kessler et al., 1999; Pole, Gone, & Kulkarni, 2008), the impact of the cultural
background of study participants of the study was examined. As an extension of this potential
moderator, the cultural background of the sample in terms of the construct of collectivism versus
individualism was examined for its possible moderating effects. For example, Elsass (2001)
found that, among South American cultures, individualist cultures tended to view PTSD
52
avoidance symptoms as maladaptive coping, whereas collectivist cultures tended to consider
avoidance symptoms to be adaptive. Such differences in perspective may have an impact on
symptom reporting, and consequently, on score reliability. To categorize cultures as collectivist
or individualist in the current study, Hofstede’s dimensions of national culture were used
(Hofstede, Hofstede, & Minkov, 2010).
Other sample characteristics that were included as possible moderators were type of
trauma and whether individuals had been forcibly displaced. The types of traumas were grouped
into two categories: intentional trauma (e.g., war, torture, sexual assault) and unintentional
trauma (e.g., natural disasters, accidents). These categories were chosen because studies have
shown that individuals who are exposed to intentional traumas had worse health outcomes than
did individuals who experienced traumas that were not deliberate (Lange et al., 2003; Matthieu
& Ivanoff, 2006; Van der Velden et al., 2006). Whether study participants were forcibly
displaced was also included as a moderator variable, as researchers tend to find higher rates of
psychological difficulties among individuals who have been displaced as the result of war as
compared to non-displaced groups (Fazel et al., 2005; Porter & Haslam, 2005; Steel et al., 2009).
Two methodological variables were included as potential moderators. Psychometric
properties of instruments can be dependent on the severity of the disorder measured (Haynes,
Smith, & Hunsley, 2011) and, as symptoms are generally more severe in individuals who are
seeking psychotherapy or medical treatment than they are in individuals from nonclinical
samples, the effect of recruitment site (i.e., clinical, community, or mixed clinical and
community) was considered.
Finally, it was predicted that internal consistency estimates would be highest in studies
involving samples for which the HTQ was originally intended to be used. Accordingly,
53
moderator analyses were conducted to determine whether having a study sample that was
culturally similar to the samples used in the development of the HTQ affects reliability estimates.
Method
Literature Search and Data Collection
To identify studies for possible inclusion in the RG, a literature search of studies having
used the HTQ was performed in the PsycINFO, PubMed, PILOTS (Published International
Literature on Traumatic Stress), and Web of Science databases. The search strategy was based on
a cited reference search, and enabled the identification of both published studies (i.e. article or
book format) and unpublished dissertations. Most authors who use the HTQ cite Mollica and
colleagues’ first published article describing the development and initial psychometric property
evaluation of the measure (Mollica et al., 1992). Some authors also cite the HTQ manual, which
was published a few years later and includes a more detailed description of the measure (Mollica
et al., 1996). Therefore two separate cited reference searches were conducted. The studies
collected were published within a timeframe ranging from the year 1992 (the year of the first
published study of the HTQ) to September 2016, when the final literature search was performed.
Inclusion criteria. Studies meeting the following inclusion criteria were selected: (a) the
authors used at least one of the HTQ symptom scales; (b) the study was empirical offering
quantitative data on the HTQ symptom scales (c) the sample consisted of participants 7 years of
age or older (to be consistent with DSM-5 [APA, 2013] criteria); (d) the study was written in
English or French (languages spoken by the authors); (e) the study was accessible through our
university’s library network (including inter-library loans); and (f) the study was available in
article, book chapter, or dissertation format.
54
Data reduction. The cited reference search of Mollica et al. (1992) yielded a total of
1,505 entries. The cited reference search of Mollica et al. (1996) yielded 66 entries. There was a
final total of 1,571 entries for all searches combined. More specifically, there were 126 entries in
PsycINFO, 125 in Medline, 724 in PILOTS, and 596 in Web of Science. After removing
duplicate entries, 862 unique studies were retrieved for further review. The full text of the
selected studies was read by the first author and retained if the study fit the inclusion criteria.
This detailed examination reduced the number of potential studies to 384. Of the excluded
studies, 41% were not empirical quantitative studies, 35.8% did not use the HTQ symptom
scales, 0.4% did not use a sample of participants aged 7 years and older, 4.6% were not written
in English or French, 12.1% were not accessible within our university library network, and 6.1%
were not available in article, book chapter, or dissertation format (see Figure 1).
Once all the studies considered for inclusion were gathered, the first author (a clinical
psychology doctoral student with previous coding experience) used a coding manual to code the
journal discipline and the reliability mentioning practices of the studies’ authors. Journal
discipline included four possible categories: (a) medical; (b) psychiatry; (c) general mental
health; and (d) other. Although “psychology” was initially considered as a possible category,
journals of this specific discipline were difficult to distinguish from other general mental health
journals and were therefore included in this latter category. The manner in which authors
commented on the reliability properties of the HTQ symptom scales included four possible
options: (a) provided a reliability coefficient from their specific sample; (b) reported a reliability
coefficient from another study; (c) mentioned that the measure is reliable without providing a
value; and (d) made no reference to score reliability.
55
Authors who did not report Cronbach’s alpha of their samples and who provided their
email addresses were contacted and asked to either provide the coefficient alpha for their sample
or to send their data sets to calculate the coefficient. Out of the 108 authors who were contacted,
35 responded (32.4% response rate), and 18 provided the internal consistency coefficient for
their sample. Independent t-tests revealed no significant differences between the mean published
and unpublished reliability estimates for both the HTQ-16, t(11) = 0.49 , p = .63, and HTQ-30,
t(4) = -2.52, p = .65.
Prior to conducting the RG analyses, multiple publications on the same data set were
identified to ensure the statistical independence of the studies to be included in the analyses.
Most multiple publications on the same data set were easily identifiable but, to ensure that all
were identified, the detection heuristic developed by Wood (2008) was used as a guideline. This
method included verifying if one or more of the “same” authors were included in the suspected
multiple publications and determining whether the study and sample characteristics were similar
in these studies. When deciding which study to eliminate from further use, several factors were
considered, such as the amount of descriptive data reported and the reporting of the coefficient
alpha of the sample. In addition, when studies provided alpha coefficients of different samples
within the same study, each sample was recognized as separate and was retained for inclusion in
the analyses.
In summary, 384 studies matched the inclusion criteria of which 185 studies had sample
reliability estimates, either reported in the publication or provided by the authors. After removing
studies using the same samples, studies that included modified versions of the HTQ symptom
scales (i.e., added or subtracted items), and unusable reliability information (such as reporting
only the range of alpha across subscales), there were a final total of 95 studies and 111 samples
56
to be included in the analyses. Specifically, this included 36 alpha coefficients for the HTQ-30
(i.e., the 16 DSM items and the 14 Indochinese culture-specific items), 70 for the HTQ-16, 22 reexperiencing subscale, 27 for the avoidance/numbing subscale, and 28 for the arousal subscale.
Coding of descriptive data and moderators. Once all the relevant articles were
identified, internal consistency values, sample characteristics, and methodological characteristics
were coded by the first author for the HTQ symptom scales and subscales. A coding manual (see
Appendix A) was developed to help guide the coding process and data were entered on a coding
sheet (see Appendix B). The coding manual and the selection of coding variables were based on
a review of the literature on trauma, cross-cultural measurement and reliability generalization.
The first author initially developed the manual and then revised the document in collaboration
with the second author. Some variables initially considered (e.g., level of education, length of
stay in host country/area) were omitted due to the lack of reporting of such data, or because the
data were presented in different incomparable formats. The study and sample variables were
used in the primary descriptive statistics of the overall pool of studies that used the scales and
subscales. Some of these variables were also examined as potential moderators. Study and
methodological characteristics included the following categorical variables: (a) language of
administration of the HTQ; (b) country of study; (c) type of sample (community; seeking mental
health services; seeking medical treatment sample); (d) version of the HTQ (original or adapted
version), (e) whether an adaptation or cultural verification was needed for the sample (yes or no);
(f) adaptation procedure used (cultural adaptation only, translation only, or both); (g) adaptation
procedure followed the developers’ recommendations (yes or no), (h) adaptation procedure
followed other experts’ recommendations (yes or no); (i) translation procedure used (oral
translation; forward translation; back translation; blind back translation); (j) a consensus
57
approach was used during adaptation (yes or no); (k) the adaptation included a pre-test (yes or
no); and (l) type of administration (clinician administered; non-clinician administered; selfreport).
Sample characteristics included sample size, the mean age of participants, and gender
composition of the sample (percentage of males), which were coded as continuous variables.
Sample characteristics also included the following categorical variables: (a) country of origin of
participants; (b) country of origin’s cultural orientation (individualist or collectivist) based on
Hofstede’s categorization (only countries classified by Hofstede were coded); (c) type of trauma
experienced (intentional [e.g., war, torture, sexual assault] or unintentional [e.g., natural
disasters, accidents]); and (d) sample consisted of forcibly displaced persons (yes or no).
To assess the reliability of the coding procedures, the second author (a clinical
psychologist and professor with extensive experience in meta-analysis) coded 20% of the studies
that were included in the meta-analyses. Reliability analyses were performed using SPSS version
20. This step was important to help eliminate random variation in the analysis and increase the
reliability and power of the results (Dieckmann, Malle, & Bodner, 2009). The inter-coder
reliability was calculated by using the kappa (k) statistic for the categorical variables, which is a
measure of the agreement between two raters (Cohen, 1960). Specifically, the k statistic
calculates the proportion of agreement while controlling for the proportion that would occur by
chance. The results had a rating of “good” according to Hunsley and Mash’s (2008) criteria, with
a mean k = 0.76. The inter-coder reliability of continuous variables was calculated by using
intraclass coefficients (ICC), (Shrout & Fleiss, 1979) and were also considered “good” as
assessed with Hunsley and Mash’s benchmarks for ICC, with a coefficient of 0.85. Several
discrepancies were caused by unclear reporting of data in the studies, vague descriptions of the
58
versions of the HTQ used, and vague descriptions of any adaptation procedures that were used.
Discrepancies between the coders were discussed until a unanimous agreement was reached.
Data Analysis
Calculating mean effect sizes. The first step of the analyses was to enter the reliability
information of each sample (sample identifier, sample size, and reliability coefficient) into the
Comprehensive Meta-Analysis software version 2.2.064 (CMA; Borenstein, Hedges, Higgins, &
Rothstein, 2005). As reliability estimates such as Cronbach’s alpha are variance accounted-for
statistics, they are often considered equivalent to r2 (Thompson & Vacha-Haase, 2000). Prior to
being entered into the CMA software, these coefficients were converted into their square root,
the same metric as a correlation. To compensate for the high level of skewness and nonnormal
distribution usually found with correlations, a Fisher’s r-to-z transformation was then performed,
and all subsequent analyses were performed with these transformations (see Borenstein, Hedges,
Higgins, & Rothstein, 2009).
Each sample was assigned a relative weight, the inverse-variance, which is determined
primarily by the sample size (i.e., larger sample sizes are generally assigned more weight), and a
95% confidence interval. An overall mean effect size was then calculated, including its statistical
significance. A calculation of the standard error of the mean effect size then allowed the
computation of the upper and lower 95% confidence intervals for the mean effect size. The
Fisher’s z values were then converted back to their original internal consistency values to
facilitate interpretation.
This meta-analytic procedure was performed to produce mean internal consistency values
for each of the following scales: (a) HTQ-30; (b) HTQ-16; (c) re-experiencing subscale; (d)
avoidance/numbing subscale; and (e) arousal subscale.
59
Random-effects model. A random-effects model was chosen to account for the
variations of the true effect from sample to sample. This model allows for the true effect to be
different across samples, in contrast to the fixed-effect model that assumes that the samples have
a common true effect and that the differences in observed effects are due entirely to sampling
error alone (Borenstein, Hedges, Higgins, & Rothstein, 2010). Thus, the random-effects model
considers that the effect sizes may come from different heterogeneous populations. Considering
the vast differences in cultural origin, language spoken, trauma experienced, and other
heterogeneous characteristics of the samples to include in the analyses, a random-effects model
was considered the most suitable for this meta-analysis.
Heterogeneity of effects. Following the procedures of recent reliability generalization
studies (e.g., Graham et al., 2011; Therrien & Hunsley, 2013), a Q test of homogeneity
(Cochran’s Q; Cochran, 1954) and I2 index (Higgins & Thompson, 2002) were used to evaluate
the degree of dispersion of reliability estimates around the mean reliability coefficient for each
scale or subscale. A significant Q statistic indicates that there is true effect size variance among
the studies and suggests that moderator analyses should be performed to ascertain the sources of
heterogeneity. Although the Q statistic evaluates the null hypothesis that the studies included in
the meta-analysis have a common effect size, it does not provide information regarding the
extent to which heterogeneity plays a role in the overall reliability estimate. An I2 index was
therefore calculated to assess the proportion of the observed variance that is due to true
heterogeneity rather than being due to chance alone (Higgins & Thompson, 2002). To illustrate,
an I2 of 0 would indicate that the variability in effect sizes is due only to sampling error and not
to between-study heterogeneity; a higher I2 index implies that there is a higher proportion of
variability that is caused by between-study heterogeneity. To help gauge the different levels of
60
heterogeneity and interpret the I2 index, the following benchmarks proposed by Higgins and
Thompson (2002) were used: 25% = low, 50% = moderate, and 75% = high.
Spearman-Brown Prophecy formula. Internal consistency estimates vary according to
test length. The higher number of test items, the higher the internal consistency. Because the
different symptom scales and subscales of the HTQ have different number of items, the
Spearman-Brown Prophecy formula (Brown, 1910; Spearman, 1910) was used to estimate the
impact of adding a number of equivalent items to the scales/subscales. For example, 14 items
would need to be added to the HTQ-16 to compare it to the total 30-item scale. To compare the
PTSD subscales with each other, the subscales for re-experiencing and arousal would each be
increased to a total of 7 items, which is the number of items in the avoidance subscale.
Possibility of publication bias. To take into account the possibility that authors do not
always report reliability coefficients in their study, a Fail-Safe N for RG (Howell & Shields,
2008) was calculated for each meta-analysis. This formula examines the possible influence that
unpublished effects may have on the mean reliability values obtained in the meta-analyses.
Funnel plots were not used to assess publication bias due to the risk of visual misinterpretation
(Lau, Ioannidis, Terrin, Schmid, & Olkin, 2006).
Analysis of moderators. Moderator analyses are an important step when conducting an
RG to clarify the sources of heterogeneity (Rodriguez & Maeda, 2006), with many RG studies
finding statistically significant relations between reliability coefficients and various types of
moderators (see Vacha-Haase & Thompson, 2011, for a review). Therefore, to examine possible
sources of the heterogeneity of variance of the reliability estimates, an analysis of moderators
was performed on scales that had statistically significant heterogeneity.
61
Because the potential moderators of this RG included both continuous and categorical
variables, two different strategies were used to assess their role in accounting for heterogeneity.
First, to examine the relations between continuous variables and score reliability, mixed effects
(method of moments) meta-regression analyses were performed. These analyses determine
whether the effect sizes are moderated, or predicted, by these variables. Second, for categorical
variables, a series of random effect subgroup analyses were used to examine the relations
between the overall internal consistencies of each scale and study and sample characteristics.
Because some of the variables had an insufficient number of studies per subgroup (i.e., 2 or
more), they were merged together to create broader categories to increase the power of the
results. Specifically, “culture” was separated according to larger geographic cultural groups (e.g.,
Asian, European, African), and “types of trauma” were categorized as either “intentional trauma”
or “unintentional trauma”.
Results
The HTQ Symptom Scales in the Empirical Literature
Overall, 384 empirical articles, book chapters, and dissertations included a version of the
HTQ symptom scales in a study. After removing studies using data from the same samples, there
were 216 independent studies and 242 independent samples (67,360 individuals) that were
administered the scale. The samples came from over 59 countries, representing over 88 cultures.
The studies were conducted in over 61 countries and the scale was administered in over 74
different languages. See Table 1 for an overview of the HTQ adaptation and administration
characteristics.
62
Reporting Practices of Reliability Estimates
Of the 384 studies considered for inclusion, 34.9% made no reference to score reliability,
7% reported a reliability coefficient from another study, and 13.5% mentioned that the measure
was reliable without providing a value. Finally, 171 (44.5%) studies reported a reliability
coefficient based on data from their specific sample. Of these, 167 specifically provided the
Cronbach’s alpha values for the HTQ symptom scale of their sample. Reliability reporting
practices across journal type were variable, as 19.5 % of studies published in medical journals,
35% of studies in psychiatry journals and 51.7% of studies in general mental health journals
reported sample-specific reliability estimates. These mediocre reporting results are similar to the
rates found in other studies investigating psychometric reporting practices in education and
mental health-related fields (Barry, Chaney, Piazza-Gardner, & Chavarria, 2014; Hall, Ward, &
Comer, 1988; Qualls & Moss, 1996). This suggests that the tendency to not report psychometric
data of sample scores is longstanding and does not appear to be changing despite the practice
guidelines that have been put in place over the years, such as the Journal Article Reporting
Standards (JARS) published by the American Psychological Association (American
Psychological Association Publications and Communications Board Working Group on JARS,
2008).
Overview of Studies Included in the Meta-Analyses and Descriptive Statistics of Potential
Moderators
Because fewer than half of the studies provided reliability estimates, the present metaanalyses included 111 samples from 95 studies, representing data from 33,376 individuals. See
Appendix C for the list of studies included in the RG analyses. The mean sample size was 301,
the mean age of participants was 36 years, and, on average, 47% of study participants were male.
63
Table 2 presents information on these characteristics for each of the meta-analyses that were
conducted. The studies were conducted in over 40 countries, with samples coming from over 42
different countries. Consistent with this, the HTQ symptom scales were administered in over 43
different languages. See Table 1 for an overview of the HTQ adaptation and administration
characteristics.
Of the study samples that reported the sampling procedure (n = 108), 60.2% recruited
participants from the community, 26.9% recruited participants who were seeking mental health
services, 11.1% recruited participants who were seeking medical treatment, and 3.7% recruited
from multiple sources. Forty-six percent of the samples were composed of forcibly displaced
individuals (i.e., refugees, asylum seekers, internally displaced people) as opposed to natives or
immigrants. The most common type of traumatic events was intentional traumas, with 78.2% of
samples having all or most individuals who experienced these traumas. Unintentional traumas
were experienced by 12.7% of the samples and the remaining 9.1% of samples consisted of
individuals who had been exposed to both types of traumas.
Reliability Generalization Analyses
Descriptive statistics and results of the RG of each of the scales and subscales are
presented in Table 3. The mean reliability coefficients ranged from α = 0.74, p < 0.001, 95% CI
[0.72, 0.77] for the re-experiencing subscale to α = 0.93, p < 0.001, 95% CI [0.92, 0.94] for the
HTQ-30. All scales and subscales had a significant Q statistic, indicating that the studies in each
meta-analysis did not share a common effect size. In addition, the I2 values ranged from 79.02%
to 96.14%. The significant Q statistics and high I2 indicate that there is an important amount of
variance across samples for all RG, therefore further examination of potential study and sample
characteristics that account for this variability is warranted.
64
Spearman-Brown formula. The full HTQ symptom scale has 30 items. If, to make it
comparable in length to this scale, 14 items of similar content were added to the HTQ-16, the
mean reliability coefficient would increase from .89 to .94. The avoidance subscale has 7 items:
an addition of 3 items of similar content to the re-experiencing scale would increase the mean
alpha from .74 to .83, and an addition of 2 items of similar content to the arousal subscale would
increase the alpha from .79 to .84.
Publication bias. Based on a threshold internal consistency value of .7, the Fail-Safe N
for obtained RG values for the scales and subscales ranged from 77 to 339. Although there are no
specific guidelines regarding adequate Fail-Safe N values, in an example, Howell and Shields
(2008) considered a Fail-Safe N of 60 to be robust to publication bias. Because all reliability
estimates for the HTQ symptom scales and subscales included in the current RG are above 60,
they are unlikely to be affected by publication bias.
Moderator analyses. As all five RG analyses had a significant Q-statistic, moderator
analyses were conducted for each scale and subscale (see Table 4 for a summary of moderating
effects). Samples that had missing information for a specific variable were excluded from these
analyses. Age, being a sample for which the HTQ was validated, being forcibly displaced
participants and the type of sampling were not significant for any of the five RG analyses.
Gender was significant only for the re-experiencing subscale, β = -0.003, Qmodel = 4.66, p =
0.031, 95% CI [-0.007, 0.000]. Specifically, study samples comprised of mainly female
participants had higher reliability estimates. Subgroup analyses revealed that the summary effect
sizes for “cultural group” were significantly different from each other for the re-experiencing
subscale. “African cultures” had the highest mean Cronbach’s alpha coefficient (α = 0.77, p <
0.001, 95% CI [0.66, 0.85]) and “Asian cultures” the lowest (α = 0.56, p < 0.001, 95% CI [0.40,
65
0.70]). “Cultural group” was not a moderator for the other scales. “Type of trauma” was also a
moderator for the re-experiencing subscales. “Intentional traumas” had a significantly higher
mean reliability coefficient (α = 0.78, p < 0.001, 95% CI [0.75, 0.80]) than “unintentional
traumas” (α = 0.72, p < 0.001, 95% CI [0.67, 0.76]). “Type of trauma” was not significant for the
other scales. Finally, the type of culture (individualist/collectivist) was a moderator for the
avoidance subscale, with “Individualist country” (α = 0.78, p < 0.001, 95% CI [0.75, 0.80])
being significantly higher than “Collectivist country” (α = 0.68, p < 0.001, 95% CI [0.56, 0.77]).
This variable did not have a moderating impact on the other scales.
Discussion
The main purposes of this study were twofold: (a) to gather updated descriptive
information regarding the use of the HTQ symptom scales in the empirical literature, and (b) to
evaluate the reliability properties of this measure by calculating mean Cronbach’s alphas and
examining sample and methodological characteristics that may influence the internal consistency
of each scale and subscale.
The HTQ in the Empirical Literature
In terms of the overall usage of the HTQ symptom scales (and subscales) in the literature,
it is clear that they remain commonly used instruments among trauma researchers. In fact, since
the most recent systematic review describing the popularity of the HTQ in 2004 (Gagnon et al.,
2004), over 295 empirical studies have been published or written in dissertation format that
included the HTQ symptom scales. The HTQ has also been used globally with a range of
samples across a wide range of cultural and ethnic backgrounds, including refugee and nonrefugees, and clinical and non-clinical populations.
66
The symptom scales of the HTQ have been administered in over 74 languages and
dialects, and when the instrument was not available for the linguistic and/or cultural group at
study, researchers have adapted it to meet their research needs. Although many authors
recognized the need to adapt the HTQ, the manner in which they proceeded to make changes to
the original usually did not follow the adaptation recommendations set by the developers of this
measure, or other recommendations provided by experts in cross-cultural assessment. For
instance, only 16% of study authors reported that their adaptation included a translation and
some form of cultural adaptation. Of the studies that did not culturally adapt the instrument, none
of them reported a cultural verification process that would ensure that the measure was suitable
for their sample. Approximately 18% of translations were oral in vivo translations that make it
difficult to ensure linguistic and semantic equivalence. Ensuring measurement equivalence by
considering both linguistic and cultural elements is critical in cross-cultural assessment (e.g.,
Flaherty et al., 1988; Keane et al., 1996), and neglecting to take these factors into account can
compromise the reliability and validity of scores obtained by the sample in question. Moreover,
only one fifth of study authors reported pre-testing the adapted version before commencing data
collection. These observations reflect the frequently made observations that (a) the process of
translating and cross-culturally validating instruments is usually not considered important in
clinical research protocols (Sperber, 2004) and (b) simple forward or back translations without a
consensus approach are common in clinical cross-cultural research (Maneesriwongul & Dixon,
2004).
RG Analyses
Although there were 384 studies that used the HTQ symptom scales found in the
literature, only 167 of them provided the Cronbach’s alpha for their sample. After eliminating
67
unusable values (i.e., values of duplicate samples) and adding values provided by authors, the
final RG analyses included 183 independent reliability estimates coming from 95 studies. With
regard to the adaptation procedures reported by the authors, the studies included in the RG
analyses appear to be comparable to the overall pool of empirical studies that used the HTQ
symptom scales. The results of this RG suggest that the HTQ symptom scales included in the
analyses can produce reliable scores across a diverse set of populations. Specifically, according
to internal consistency criteria provided by Hunsley and Mash (2008), the mean internal
consistency value of the HTQ-30 is excellent (.93) and the mean estimate of the HTQ-16 is good
(.89). The unpublished internal consistency values provided by authors had comparable means
with the HTQ-30 being rated as “excellent” and the HTQ-16 as “good”. If the Spearman-Brown
prophecy formula is used to estimate the impact of adding 14 similar items to the HTQ-16 (to
compare it to the HTQ-30, which includes the Indochinese culture-related symptoms), the HTQ16’s mean reliability estimate to increases to .94, a slightly higher value than the mean reliability
of the HTQ-30. This result suggests that the addition of the cultural items, per se, does not add
more value to the HTQ-16 in terms of internal consistency. However, this does not imply that the
14 cultural symptoms are irrelevant, as these items may add value to the validity properties of the
measure. Furthermore, the majority of studies that used the HTQ-30 did not report culturally
adapting the HTQ for their specific sample. These studies thus applied the Indochinese version
of the questionnaire to non-Indochinese groups. This implies that the cultural items that are
based on clinical findings among Indochinese populations can produce reliable scores across
various populations. However, it is unknown whether the addition of other culture-related items
could improve the reliability of scores for each of these cultures. Further research is thus needed
68
to explore the impact of these cultural items on the construct validity of the HTQ across different
populations.
Even though the mean reliability estimates of the HTQ-16 and HTQ-30 are good or
excellent, the mean reliability estimates for the re-experiencing, avoidance/numbing and arousal
subscales are only adequate (.74, .78, .79 respectively). This lower performance in reliability of
the subscales may be due to the smaller number of items compared to the total scales (Cortina,
1993). Using the Spearman-Brown formula to evaluate the possible effect on reliability of having
7 items in all subscales increased the re-experiencing and arousal reliability estimates to .83 and
.84 respectively. These results indicate that, in principle, the re-experiencing and arousal
subscales perform relatively better than the avoidance/numbing subscale in terms of producing
reliable scores. This may be due to inconsistent presence of avoidance and numbing symptoms
found in trauma studies across various cultures (e.g., Marsella et al., 1996; Norris, Van
Landingham, & Vu, 2009) and issues regarding the factorial validity of the 3-factor model of
PTSD (e.g. Yufik & Simms, 2010). These issues were highlighted throughout the revision
process for the new DSM-5 criteria for PTSD (Hinton & Lewis-Fernández, 2010), during which
it was decided to separate the avoidance and numbing symptoms into separate clusters.
Moderator Analyses
The statistically significant Q values and high I2 tests for all scales and subscales suggest
that the variability among the reliability estimates is heterogeneous and is not due to sampling
error alone. The high degree of variance that is caused by true heterogeneity of the scales
indicated that examining sample and methodological characteristics may help explain this
variability. However, none of the potential moderators examined accounted for the variance in
reliability estimates of the HTQ-16 and HTQ-30. These variables included age, being a sample
69
for which the HTQ was validated, the type of sampling, and being a forcibly displaced sample. It
is possible that there may have been an insufficient number of studies available for the analyses
to test these assumptions with sufficient power. Although there may also be other sample or
study characteristics that could further explain the variability, it is not immediately obvious what
these might be, as the characteristics selected for inclusion in the analyses were those typically
shown to impact PTSD severity and/or reliability estimates.
Interestingly, three variables had a moderating effect on the typical reliability of the reexperiencing subscale. One of these moderators was gender, where samples with a higher
proportion of women tended to produce more reliable scores. This finding is consistent with
another RG study (Vassar, Knaup, Hale, & Hale, 2011) conducted with a different trauma
symptoms measure, the Impact of Event Scale (IES; Horowitz, Wilner, & Alvarez, 1979). This
RG found a positive correlation between the percentage of female participants and reliability
estimates for the intrusion subscale. As symptoms such as flashbacks are intensely emotional in
nature, Vassar et al. (2011) posited that women may be more inclined to report experiencing
these symptoms due to social norms where women are more considered to be emotionally
expressive than men. However, research is needed to determine whether this explanation may
account for the present findings with the re-experiencing subscale.
Analyses including cultural group found that reliability coefficients for the reexperiencing subscale were the highest for cultures that were grouped under “African cultures”
and the lowest for those grouped under “Asian cultures.” These findings are consistent with
suggestions that the experience of symptoms such as flashbacks and nightmares may be related
to social norms such as gender-linked expectations for emotional expressivity (Vassar et al.,
2011) and cultural valuing of dreams (Shore, Orton, & Manson, 2009). Although these
70
explanations may account for the present findings with the re-experiencing subscale, research is
needed to substantiate such interpretations.
The type of trauma also played a role in the variance of reliability estimates for the reexperiencing subscale. When the types of trauma were grouped into “intentional traumas” and
“unintentional traumas,” the former had a significantly higher typical reliability estimate than did
the latter. As intentional traumas are considered to be more severe these results may indicate that
the internal consistency of re-experiencing symptoms of the HTQ is associated with trauma
severity. However, to test these interpretations, it would be necessary to have data on trauma
severity that was independent of the HTQ scores.
The final significant moderation effect in the analyses was for the cultural orientation of
country of origin (individualist vs. collectivist) on the avoidance subscale of the HTQ.
Specifically, the reliability coefficient for individualist countries was significantly higher than
the typical estimate of collectivist countries. This may be reflective of lower rates of
avoidance/numbing symptoms reported in some collectivist cultures, such as Kalahari bushmen
(McCall & Resick, 2003) and Vietnamese Americans (Norris et al., 2009). On the other hand,
studies including other collectivist cultures have found higher rates of endorsement of
avoidance/numbing symptoms (Dyregrov et al., 2000). These variations suggest that more
research is needed regarding the role cultural orientation plays in posttraumatic symptoms.
In summary, gender, cultural group and type of trauma had a moderating effect on the
reliability of scores of the re-experiencing subscale, and culture orientation had a moderating
effect on the reliability of scores of the avoidance/numbing subscale. It is interesting to note that
cultural variables (i.e., cultural group and cultural orientation) explained some of the variability
for the re-experiencing and avoidance/numbing subscales, but did not for the arousal items. This
71
pattern offers some support for the perspective that arousal symptoms are consistent across
cultures, whereas re-experiencing and avoidance/numbing symptom expression are likely to vary
across cultures (e.g., Hinton & Lewis-Fernández, 2011). Further to this point, these results
appear to be similar to those found by Rasmussen and colleagues (2015) in a study on the
measurement invariance of the HTQ-16. Their results showed that the configural invariance, or
overall four-factor structure of the PTSD basic construct as measured by the HTQ-16, was
acceptable. However, there were significant response-style differences across cultures, as well as
variations in the clinical thresholds of PTSD.
Limitations of the Study
The results of this RG should be interpreted in light of the following limitations.
Although scientific standards require that citations be provided for instruments used in a study,
the cited reference search strategy that was used may not have identified all published studies
that have used the HTQ symptom scales. For the published studies identified with the search
strategy, data from many were not included in the RG analyses because Cronbach’s alphas were
not reported in the primary studies. In fact, only 43.5% of study authors reported a reliability
coefficient based on data from their study sample. Even though some authors provided a
coefficient alpha after being contacted, there were still many missing reliability estimates that
could have been included in the RG. Although there were no significant differences between the
unpublished estimates provided by the authors and the published ones, the remaining
unpublished estimates may have had an impact on the results. Additionally, there are likely to be
studies using the HTQ that have not been published. If these unpublished studies have lower
reliability estimates, this would alter the overall mean internal consistency values reported in this
study.
72
In addition, the coding process was difficult at times because the information reported in
the studies was often vague and challenging to interpret. There was often a lack of sufficient
detail reported about study methods and it is therefore possible that the information reported in
many of the studies did not fully represent the actual adaptation procedures that were used.
Indeed, insufficient detail in the reporting of adaptation methods appears to be a common
problem in the published literature (Maneesriwongul & Dixon, 2004). This issue continues
despite reporting standards for research that have been put in place, in part, to ensure the quality
of research synthesis and meta-analysis (Appelbaum, Cooper, Maxwell, Stone, & Sher, 2008).
Finally, RG analyses were not conducted on culturally adapted versions of the HTQ due
to a limited amount of data. Although there were numerous cultural adaptations found in the
literature, there was an insufficient amount of studies for each adapted version of the HTQ that
had reliability estimates to conduct meta-analyses. Consequently, the overall psychometric
properties of these versions and the adaptation procedures used could not be evaluated.
Implications and Recommendations
As mentioned, using strategies to ensure equivalency is essential in cross-cultural
assessment to help minimize bias. Users of the HTQ have access to multiple cross-cultural
adaptation guidelines in the literature and the HTQ developers themselves provide
recommendations for adapting the instrument for other populations. However, in reality, it
appears that researchers may not have followed these adaptation guidelines in their own use of
the HTQ, or at least, their specific adaptation procedures have not been reported in published
reports. This may be also due to a lack of awareness of adaptation guidelines or reporting
standards among researchers, or perhaps journal submission requirements that limit the amount
of information provided in the methods section of manuscripts. This, combined with the low rate
73
in the reporting of reliability estimates, is an indication that more awareness is needed with
regards to reporting guidelines for assessment instruments and assessment research. The high
cost and lengthy process of adapting instruments can be major challenges, but because the
assessment of traumatized individuals can have major clinical and societal implications (e.g.,
access to treatment, obtaining disability compensation, obtaining refugee status), must take
appropriate steps to minimize bias.
The findings of the current meta-analyses indicate that both the HTQ-16 and HTQ-30 are
likely to provide reliable scores across diverse populations. However, research results obtained
with the subscales should be interpreted cautiously due to lower overall mean reliability
estimates and inadequate performance in certain samples. Although these results have important
implications for clinicians and researchers as they weigh the strengths and weaknesses of the
HTQ, a further evaluation of the construct validity of these scales, specifically convergent and
discriminant validity is needed. This, in turn, would provide even further insight into the crosscultural applicability of this instrument.
74
References
American Educational Research Association, American Psychological Association, & National
Council on Measurement in Education (2014). Standards for educational and
psychological testing (5th ed.). Washington, DC: Author.
American Psychiatric Association. (1987). Diagnostic and statistical manual of mental
disorders (3rd ed. revised). Washington, DC: Author.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders
(4th ed.). Washington, DC: Author.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental
disorders (5th ed.). Washington, DC: Author.
American Psychological Association Publications and Communications Board Working Group
on JARS (2008). Reporting standards for research in psychology: Why do we need them?
What might they be? American Psychologist, 63, 839-851. doi: 10.1037/0003066X.63.9.839
Appelbaum, M., Cooper, H, Maxwell, S., Stone, A., & Sher, K. J. (2008). Reporting standards in
research in psychology: Why do we need them? What might they be? American
Psychologist, 63, 839-851. doi: 10.1037/0003-066X.63.9.839
Barnes, L. L. B., Harp, D., & Jung, W. S. (2002). Reliability generalization of scores on the
Spielberger State-Trait Anxiety Inventory. Educational and Psychological Measurement,
62, 603-618. doi: 10.1177/0013164402062004005
Barry, A. E., Chaney, B., Piazza-Gardner, A. K., & Chavarria, E. A. (2014). Validity and
reliability reporting practices in the field of health education and behavior: a review of
75
seven journals. Health Education & Behavior, 41, 12-18. doi:
10.1177/1090198113483139
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2005). Comprehensive
meta-analysis version 2. Englewood, NJ: Biostat.
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction
to meta-analysis. Chichester, UK: John Wiley & Sons, Ltd.
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2010). A basic
introduction to fixed-effect and random-effects models for meta-analysis. Research
Synthesis Methods, 1, 97-111. doi: 10.1002/jrsm.12
Breslau, N., Davis, G., Andreski, P., & Peterson, E. (1991). Traumatic events and posttraumatic
stress disorder in an urban population of young adults. Archives of General Psychiatry,
48, 216-222. doi:10.1001/archpsyc.1991.01810270028003
Breslau, N., Kessler, R. C., Chilcoat, H. D., Schultz, L. R., Davis, G. C., & Andreski, P. (1998).
Trauma and posttraumatic stress disorder in the community: The 1996 Detroit Area
Survey of Trauma. Archives of General Psychiatry, 55, 626-632. doi:
10.1001/archpsyc.55.7.626.
Brown, W. (1910). Some experimental results in the correlation of mental abilities. British
Journal of Psychology, 3, 296–322. doi: 10.1111/j.2044-8295.1910.tb00207.x
Cochran, W. G. (1954). The combination of estimates from different experiments. Biometrics,
10, 101–129. doi: 10.2307/3001666
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological
Measurement, 20, 37-46. doi:10.1177/001316446002000104
Cortina, J. M. (1993). What is coefficient alpha? An examination of theory and applications.
76
Journal of Applied Psychology, 78, 98-104. doi: 10.1037/0021-9010.78.1.98
Dieckmann, N., Malle, B., & Bodner, T. (2009). An empirical assessment of meta- analytic
practice. Review of General Psychology, 13, 101-115. doi:10.1037/a0015107
Dyregrov, A., Gupta, L., Gjestad, R., & Mukanoheli, E. (2000). Trauma exposure and
psychological reactions to genocide among Rwandan children. Journal of Traumatic
Stress, 13, 3–21. doi: 10.1023/A:1007759112499
Elsass, P. (2001). Individual and collective traumatic memories: A qualitative study of
posttraumatic stress disorder symptoms in two Latin American localities. Transcultural
Psychiatry, 38, 306-316. doi: 10.1177/136346150103800302
Fazel, M., Wheeler, J., & Danesh, J. (2005). Prevalence of serious mental disorder in 7000
refugees resettled in western countries: A systematic review. Lancet, 365, 1309-1314.
doi: 10.1016/S0140-6736(05)61027-6
Flaherty, J. A., Gaviria, F. M., Pathak, D., Mitchell, T., Wintrob, R., Richman, J. A., & Birz, S.
(1988). Developing instruments for cross-cultural psychiatric research. Journal of
Nervous and Mental Disease, 176, 257-263. doi: 10.1097/00005053-198805000-00001
Gagnon, A. J., Tuck, J., & Barkun, L. (2004). A systematic review of questionnaires measuring
the health of resettling refugee women. Health Care for Women International, 25, 111149. doi: 10.1080/07399330490267503
Graham, J. M., Diebels, K. J, & Barnow, Z. B. (2011). The reliability of relationship
satisfaction: A reliability generalization meta-analysis. Journal of Family Psychology, 25,
39-48. doi/10.1037/a0022441
Graham, J. M., Yenling, J. L., & Jeziorski, J. L. (2006). The Dyadic Adjustment Scale: A
reliability generalization meta-analysis. Journal of Marriage and Family, 68, 701-717.
77
doi: 10.1111/j.1741-3737.2006.00284.x
Hall, B. W., Ward, A., W. & Comer, C. B. (1988). Published educational research: An empirical
study of its quality. Journal of Educational Research, 81, 182-189. doi:
10.1080/00220671.1988.10885820
Haynes, S. N., Smith, G. T., & Hunsley, J. D. (2011). Scientific foundations of clinical
assessment. New York : Routledge.
He, J., & van de Vijver, F. (2012). Bias and equivalence in cross-cultural research. Online
Readings in Psychology and Culture, 2. Retrieved from
http://scholarworks.gvsu.edu/cgi/viewcontent.cgi?article=1111&context=orpc
Henson, R. K., Kogan, L. R., & Vacha-Haase, T. (2001). A reliability generalization of the
Teacher Efficacy Scale and related instruments. Educational and Psychological
Measurement, 61, 404-420. doi: 10.1177/00131640121971284
Higgins, J. P. T., & Thompson, S. G. (2002). Quantifying heterogeneity in a meta-analysis.
Statistics in Medicine, 21, 1539-1558. doi: 10.1002/sim.1186
Hinton, D. E., & Lewis-Fernández, R. (2010). Idioms of distress among trauma survivors:
Subtypes and clinical utility. Culture, Medicine, and Psychiatry, 34, 209–218. doi:
10.1007/s11013-010-9175-x
Hinton, D. E., & Lewis-Fernández, R. (2011). The cross-cultural validity of posttraumatic stress
disorder: Implications for DSM-5. Depression & Anxiety, 28, 783-801. doi:
10.1002/da.20753
Hofstede, G., Hofstede, G. J., & Minkov, M. (2010). Cultures and organizations: Software of the
mind (3rd ed.). New York: McGraw-Hill.
Hollifield, M., Warner, T. D., Lian, N., Krakow, B., Jenkins, J. H., Kesler, J., … Westermeyer, J.
78
(2002). Measuring trauma and health status in refugees: A critical review. Journal of the
American Medical Association, 288, 611–621. doi: 10.1001/jama.288.5.611
Horowitz, M., Wilner, N., & Alvarez, W. (1979). Impact of Event Scale: A measure of
subjective stress. Psychosomatic Medicine, 41, 209-218. doi: 10.1097/00006842197905000-00004
Howell, R. T., & Shields, A. L. (2008). The file drawer problem in reliability generalization:
A strategy to compute a Fail-Safe N with reliability coefficients. Educational and
Psychological Measurement, 68, 120-128. doi: 10.1177/0013164407301528
Hunsley, J., & Mash, E. J. (2008). Developing criteria for evidence-based assessment: An
introduction to assessments that work. In J. Hunsley & E. J. Mash (Eds.), A guide to
assessments that work (pp. 3-14). New York: Oxford University Press.
Jones, E., Vermaas, R. H., McCartney, H., Beech, C., Palmer, I., Hyams, K., & Wessely, S.
(2003). Flashbacks and post-traumatic stress disorder: The genesis of a 20th-century
diagnosis. British Journal of Psychiatry, 182, 158–163. doi: 10.1192/bjp.182.2.158
Keane, T. M., Kaloupek, D. G., & Weathers, F. W. (1996). Ethnocultural considerations in the
assessment of PTSD. In A. J. Marsella, M. J. Friedman, E. T. Gerrity, & R. M. Scurfield
(Eds.), Ethnocultural aspects of posttraumatic stress disorder: Issues, research, and
clinical applications (pp. 183-205). Washington, DC: American Psychological
Association.
Keane, T. M., Silberbogen, A. K., & Weierich, M. R. (2008). Post-traumatic stress disorder. In J.
Hunsley & E.J. Mash (Eds.), A guide to assessments that work (pp. 293-315). New York:
Oxford University Press.
79
Kessler, R. C., Sonnega, A., Bromet, E., Hughes, M., & Nelson, C. B. (1995). Posttraumatic
stress disorder in the national comorbidity survey. Archives of General Psychiatry, 52,
1048-1060. doi: 10.1002/1099-1298(200011/12)
Kessler, R. C., Sonnega, A., Bromet, E., Hughes, M., Nelson, C. B., & Breslau, N. (1999).
Epidemiological risk factors for trauma and PTSD. In R. Yehuda (Ed.), Risk factors for
posttraumatic stress disorder (pp. 23-60). Washington, DC: American Psychiatric Press.
Kleijn, W. C., Hovens, J. E. & Rodenburg, J. J. (2001). Posttraumatic stress symptoms in
refugees: Assessments with the Harvard Trauma Questionnaire and the Hopkins
Symptom Checklist-25 in different languages. Psychological Reports, 88, 527-532. doi:
10.2466/PR0.88.2.527-532
Koenen, K. C., Ratanatharathorn, A., Ng, L., McLaughlin, K. A., Bromet, E. J., Stein, D. J., …
Kessler, R. C. (2017). Posttraumatic stress disorder in the World Mental Health Surveys,
Psychological Medicine, 1-15. doi: 10.1017/S0033291717000708
Lange, A., Rietdijk, D., Hudcovicova, M., van de Ven, J. P., Schrieken, B., &
Emmelkamp,
P. M. (2003). Interapy: A controlled randomized trial of the standardized treatment of
posttraumatic stress through the internet. Journal of Consulting and Clinical Psychology,
71, 901–909. doi: 10.1037/0022-006X.71.5.901
Lau, J., Ioannidis, J. P. A., Terrin, N., Schmid, C. H., & Olkin, I. (2006). The case of the
misleading funnel plot. British Medical Journal, 333, 597-600. doi:
10.1136/bmj.333.7568.597
Maneesriwongul, W. & Dixon, J. K. (2004) Instrument translation process: A methods review.
Journal of Advanced Nursing, 48, 175– 186. doi: 10.1111/j.1365-2648.2004.03185.x
80
Marsella, A. J., Friedman, M. J., Gerrity, E. T., & Scurfield, R. M. (1996). Ethnocultural aspects
of posttraumatic stress disorder: Issues, research, and clinical applications. Washington,
DC: American Psychological Association.
Matthieu, M., & Ivanoff, A. (2006). Treatment of human-caused trauma: Attrition in the adult
outcomes research. Journal of Interpersonal Violence, 2, 1654–1664. doi:
10.1177/0886260506294243
McCall, G., & Resick, P. (2003). A pilot study of PTSD symptoms among Kalahari Bushmen.
Journal of Traumatic Stress, 16, 445–450. doi: 10.1023/A:1025702326392
Mollica, R. F., Caspi-Yavin, Y., Bollini, P., Truong, T., Tor, S., & Lavelle, J. (1992). The
Harvard Trauma Questionnaire: Validating a cross-cultural instrument for measuring
torture, trauma, and posttraumatic stress disorder in Indochinese refugees. The Journal of
Nervous and Mental Disease, 180, 111-116. doi: 10.1097/00005053-199202000-00008
Mollica, R. F., Caspi-Yavin, Y., Lavelle, J., Tor, S., Yang, T., Chan, S., ... de Marneffe, D.
(1996). The Harvard Trauma (HTQ) Manual: Cambodian, Laotian, and Vietnamese
versions. Torture, 6(1), 19-33.
Nakeyar, C., & Frewen, P. A. (2016). Evidence-based care for Iraqi, Kurdish, and Syrian asylum
seekers and refugees of the Syrian civil war: A systematic review. Canadian Psychology,
57, 233-245. doi: 10.1037/cap0000067
Norris, F. H. (1992). Epidemiology of trauma: frequency and impact of different potentially
traumatic events on different demographic groups. Journal of Consulting and Clinical
Psychology, 60, 409-418. doi: 10.1037/0022-006X.60.3.409
Norris, F. H., Van Landingham, M. J., & Vu, L. (2009). PTSD in Vietnamese Americans
following Hurricane Katrina: Prevalence, patterns, and predictors. Journal of Traumatic
81
Stress, 22, 91-101. doi: 10.1002/jts.20389
Perkonigg, A., Kessler, R. C., Storz, S., & Wittchen, H.-U. (2000). Traumatic events and posttraumatic stress disorder in the community: Prevalence, risk factors and comorbidity.
Acta Psychiatrica Scandinavica, 101, 46-59. doi: 10.1034/j.16000447.2000.101001046.x
Pietrzak, R. H., Goldstein, R. B., Southwick, S. M., & Grant, B. F. (2011). Prevalence and Axis I
comorbidity of full and partial posttraumatic stress disorder in the United States: Results
from Wave 2 of the National Epidemiologic Survey on Alcohol and Related Conditions.
Journal of Anxiety Disorders, 25, 456-465. doi: 10.1016/j.janxdis.2010.11.010
Pole, N., Gone, J. P., & Kulkarni, M (2008). Posttraumatic stress disorder among ethnoracial
minorities in the United States. Clinical Psychology: Science and Practice, 15, 35–61.
doi: 10.1111/j.1468-2850.2008.00109.x
Porter, M., & Haslam, N. (2005). Predisplacement and postdisplacement factors associated
with mental health of refugees and internally displaced persons: A meta-analysis.
Journal of the American Medical Association, 294, 602-612. doi:
10.1001/jama.294.5.602.
Qualls, A. L. & Moss, A. D. (1996). The degree of congruence between test standards and
test documentation within journal publications. Educational and Psychological
Measurement, 56, 209-214. doi: 10.1177/0013164496056002002
Rasmussen, A., Verkuilen, J., Ho, E., & Fan, Y. (2015). Posttraumatic stress disorder among
refugees: Measurement invariance of Harvard Trauma Questionnaire scores across global
regions and response patterns. Psychological Assessment, 27, 1160-1170. doi:
10.1037/pas0000115
82
Renner, W., Salem, I., & Ottomeyer, K. (2007). Posttraumatic stress in asylum seekers from
Chechnya, Afghanistan, and West Africa: Differential findings obtained by quantitative
and qualitative methods in three Austrian samples. In J. P. Wilson & C. C. Tang (Eds.),
Cross-cultural assessment of psychological trauma and PTSD (pp. 239-278). Boston,
MA: Springer.
Rodriguez, M. C., & Maeda, Y. (2006). Meta-analysis of coefficient alpha. Psychological
Methods, 11, 306-322. doi:10.1037/1082-989X.11.3.306
Shoeb, M., Weinstein, H., & Mollica, R. (2007). The Harvard Trauma Questionnaire: Adapting a
cross-cultural instrument for measuring torture, trauma and posttraumatic stress disorder
in Iraqi refugees. The International Journal of Social Psychiatry, 53, 447-463. doi:
10.1177/0020764007078362
Shore, J. H., Orton, H., & Manson, S. M. (2009). Trauma-related nightmares among
American Indian veterans: Views from the dream catcher. American Indian and
Alaska
Native Mental Health Research, 16, 25-38. doi : 10.5820/aian.1601.2009.25
Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater
reliability. Psychological Bulletin, 86, 420-428. doi: 10.1037/0033-2909.86.2.420
Spearman, C. C. (1910). Correlation calculated from faulty data. British Journal of Psychology,
3, 271-295. doi: 10.1111/j.2044-8295.1910.tb00206.x
Sperber, A. D. (2004) Translation and validation of study instruments for cross-cultural research.
Gastroenterology, 126, S124-S128. doi: 10.1053/j.gastro.2003.10.016
Steel, Z., Chey, T., Silove, D., Marmane, C., Bryant, R. A., & van Ommeren, M. (2009).
Association of torture and other potentially traumatic events with mental health outcomes
among populations exposed to mass conflict and displacement: A systematic review and
83
meta-analysis. Journal of the American Medical Association, 302, 537-549. doi:
10.1001/jama.2009.1132
Therrien, Z., & Hunsley, J. (2013). Assessment of anxiety in older adults: A
reliability generalization meta-analysis of commonly used measures. Clinical
Gerontologist 36, 171-194. doi:10.1080/07317115.2013.767871
Thompson, B. (1994). Guidelines for authors. Educational and Psychological Measurement,
54(4), 837-847.
Thompson, B. & Vacha-Haase, T. (2000). Psychometrics is datametrics: The test is
not reliable. Educational and Psychological Measurement, 60, 174-195. doi:
10.1177/0013164400602002
Vacha-Haase, T. (1998). Reliability generalization: Exploring variance in measurement error
affecting score reliability across studies. Educational and Psychological Measurement,
58, 6-20. doi: 10.1177/0013164498058001002
Vacha-Haase, T., Kogan, L.R., & Thompson, B. (2000). Sample compositions and variabilities
in published studies versus those in test manuals: Validity of score reliability inductions.
Educational and Psychological Measurement, 60, 509–522. doi:
10.1177/00131640021970682
Vacha-Haase, T., & Thompson, B. (2011). Score reliability: A retrospective look back
at 12 years of reliability generalization studies. Measurement and Evaluation in
Counseling and Development, 44, 159-168. doi: 10.1177/0748175611409845
Van der Velden, P. G., Grievink, L., Kleber, R. J., Drogendijk, A. N., Roskam, A. J.,
Marcelissen, F. G., … Gersons, B. P. (2006). Post-disaster mental health problems and
the utilization of mental health services: A four-year longitudinal comparative study.
84
Administration and Policy in Mental Health and Mental Health Services Research, 33,
279-288. doi: 10.1007/s10488-005-0027-x
Vassar, M., Knaup, K.G., Hale, W., & Hale, H. (2011). A meta-analysis of coefficient alpha for
the Impact of Event Scales: A reliability generalization study. South African Journal of
Psychology, 41, 6-16. doi: 10.1177/008124631104100102
Wood, J. (2008). Methodology for dealing with duplicate study effects in meta-analysis.
Organizational Research Methods, 11, 79-95. doi: 10.1177/1094428106296638
Yufik, T., & Simms, L. J. (2010). A Meta-analytic investigation of the structure of posttraumatic
stress symptoms. Journal of Abnormal Psychology, 119, 764-776. doi: 10.1037/a0020981
85
Figure 1
Literature Search Flow Chart
1571 entries identified
724 from P.I.L.O.T.S.
596 from Web of Science
126 from PsycINFO
125 from Medline
Studies excluded from this review
709 were duplicates
196 were not empirical studies
171 did not use HTQ symptom
scales
58 were not accessible
29 were not available in book,
article, or dissertation format
22 were not in French or English
2 did not have participants aged > 6
384 studies used the HTQsymptom scales
Studies excluded from this review
199 studies did not provide
Cronbach’s alpha for their sample
185 studies had sample reliability
estimates
167 were retrieved from the
literature search
18 were provided by authors
Studies excluded from this review
90 studies included duplicate
samples and/or modified the HTQ
(i.e., added or subtracted items)
95 studies (111 samples) were
included in the RG
86
Table 1
Descriptive Statistics of the Adaptation and Administration Procedures Reported in
Samples Using the HTQ Symptom Scales
Samples included in
Overall pool of samples
the RG analyses
n (%)
n (%)
242
111
Required an adaptation
199 (82.2)
97 (87.4)
Total adaptations of the HTQ
181 (74.8)
85 (76.6)
Reported adaptation methods
127
57
95 (74.8)
42 (73.7.4)
3 (2.4)
1 (1.8)
Translation and cultural adaptation
29 (16.0)
14 (24.6)
Used a consensus approach
60 (47.2)
30 (52.6)
Conducted a pre-test
28 (22.0)
7 (12.3)
Used HTQ adaptation guidelines
12 (9.4)
9 (15.8)
Used other adaptation guidelines
30 (23.6)
14 (24.6)
124
56
In vivo oral translation
22 (17.7)
8 (14.3)
Forward translation
25 (20.2)
13 (23.2)
Back translation
54 (43.5)
28 (50)
Blind back translation
23 (18.5)
7 (12.5)
226
108
Self-report
105 (46.5)
61 (56.5)
Non-clinician administered
50 (22.1)
20 (18.5)
Clinician administered
67 (29.6)
23 (21.3)
4 (1.8)
4 (3.7)
Total number of samples
Translation only
Cultural adaptation only
Reported translation procedures
Reported administration procedures
Mixed methods
87
Table 2
Descriptive Statistics of Studies Included in the RG
Scale
Items
Total
sample size
Mean
sample size
Mean age
(years)
HTQ-30
30
9,322
259
37.14
Gender
composition
(% males)
44.04
HTQ-16
16
23,721
339
35.47
50.62
Re-experiencing
4
6,731
306
36.69
38.28
Avoidance/numbing
7
7,910
293
35.01
38.05
Arousal
5
8,094
289
34.87
38.13
88
Table 3
Descriptive Statistics for the Reliability Values of the HTQ Symptom Scales and Subscales
95% confidence
interval
Lower
Upper
Scale
k
Mean α
HTQ-30
36
.93
.92
HTQ-16
70
.89
Re-experiencing
22
Avoidance/
numbing
Arousal
Min.
Max.
Q
I2
.94
.86
.98
684.27*
94.89
Fail-Safe
N for RG
339
.87
.90
.73
.98
1786.34*
96.14
272
.74
.72
.77
.43
.84
100.08*
79.02
77
27
.78
.75
.81
.54
.91
258.93*
89.96
222
28
.79
.76
.82
.57
.90
333.65*
91.91
198
Note. Min. = Minimum reliability estimate reported in the literature; Max. = Maximum reliability estimate reported in the literature; k
= number of samples included in the analyses; *p < .001
89
Table 4
Summary of Moderating Effects for the Internal Consistency of the HTQ Symptom
Scales
Moderator
HTQ-30
N = 36
HTQ-16
N = 70
Re-experiencing
N = 22
Avoidance
N = 27
Arousal
N = 28
Age
No (28*)
No (52)
No (21)
No (25)
No (26)
Gender
No (34)
No (62)
Yes (21)
No (26)
No (27)
Cultural
group
No (34)
No (59)
Yes (22)
No (27)
No (28)
Cultural
orientation
No (21)
No (30)
No (21)
Yes (23)
No (24)
Forcibly
displaced
sample
No (31)
No (68)
No (22)
No (27)
No (28)
Recruitment
site
No (36)
No (67)
No (22)
No (27)
No (28)
Trauma type
No (32)
No (64)
Yes (18)
No (20)
No (21)
Original
sample
No (36)
No (70)
N/A
N/A
N/A
Note. *Total samples included in the moderator analyses are in parentheses; Yes = Significant
moderating effect; No = No significant moderating effect; N/A = Insufficient information available to
conduct moderator analyses
90
Study 2
A Validity Generalization Study of the Harvard Trauma Questionnaire
91
Abstract
The current study examined the construct validity of the symptom scales of the Harvard Trauma
Questionnaire (HTQ; Mollica et al., 1992), a commonly used cross-cultural instrument that aims
to assess the psychological effects of trauma. A cited-reference search was conducted to locate
all publications and dissertation that used the symptoms scales of the HTQ. Random-effects
validity generalization (VG) meta-analyses were performed on discriminant and convergent
validity coefficients. These analyses included 125 validity coefficients from 78 studies,
representing data from 21,156 individuals. The variance of score validity was characterized by
considering sample and methodological variables that are predictive of the validity coefficients.
The findings of the VG analyses indicate that the convergent validity properties of the HTQ-16
are supported to some extent, and that this scale performs rather questionably in terms of
discriminant validity. Furthermore, there is limited support for either the convergent or
discriminant validity of the HTQ-30. Despite some encouraging results, more work is needed to
establish the validity of these scales. We discuss the complexities involved in establishing the
construct validity PTSD assessments in light of the high comorbidity rates and over lapping
symptoms with other disorders.
Keywords: posttraumatic stress disorder, assessment, cross-cultural assessment, construct
validity, meta-analysis
92
A Validity Generalization Study of the Harvard Trauma Questionnaire
Assessment instruments rely heavily on the conceptualization of the construct they are
intended to measure. If the conceptual underpinnings of the construct are not solid, then the
psychometric properties of the instrument can be seriously compromised (Haynes, Smith, &
Hunsley, 2011). One psychological construct that has been extensively debated in regards to its
conceptualization is posttraumatic stress disorder (PTSD). A central concern is that the PTSD
construct was developed according to Western cultures’ conceptualizations of trauma and its
sequelae, and may not be applicable to other cultures (e.g., Jones et al., 2003). Consequently, it is
not surprising that there has also been much questioning regarding the cross-cultural applicability
of screening tools used to assess PTSD. The objective of the current study is to shed some light
on this issue by examining the construct validity of the Harvard Trauma Questionnaire (HTQ;
Mollica et al., 1992), a PTSD measure that was designed to be adapted and used across cultures.
The Harvard Trauma Questionnaire
The HTQ was developed by Mollica and colleagues (1992) and has been validated by its
developers for various cultural and linguistic groups (Shoeb, Weinstein, & Mollica, 2007).Most
versions of the HTQ consist of four parts. Part 1 measures different traumatic events, part 2 is an
open-ended description of the most traumatic event experienced by the respondent, part 3
evaluates the circumstances surrounding possible head injury and, finally, part 4 is a list of
trauma symptoms. The first 16 items in part 4 are derived from the DSM-III-R/IV (American
Psychiatric Association, 1987, 1994) PTSD criteria, and they are the same in every version of the
HTQ. The second set of trauma symptom items in part 4 are culture-specific questions and are
tailored for each version of the HTQ. The HTQ has also been translated and adapted in many
other languages by other researchers (e.g. Kleijn, Hovens, & Rodenburg, 2001). Systematic
93
reviews of tools used to assess the health of refugees have found that the HTQ has been
extensively used by researchers to assess trauma and its sequelae (Gagnon, Tuck, & Barkun,
2004; Hollifield et al., 2002), and has been shown to be useful to exclude PTSD non-cases in
certain forcibly displaced populations (Nakeyar & Frewen, 2016). In addition, in their review of
PTSD measures, Keane, Silberbogen, and Weierich (2008) recommended the use of the HTQ for
the assessment of PTSD across cultures. They described the HTQ as having “linguistic
equivalence across the many cultures and languages with which it has been used thus far” (p.
297). However, they did not provide sufficient empirical evidence to support the cross-cultural
applicability of the instrument, or evaluate it according to established evidence-based assessment
guidelines, such as Hunsley and Mash’s (2008).
We recently conducted a reliability generalization (RG) study to evaluate the internal
consistency of HTQ scores through meta-analysis (see Darzi & Hunsley, 2017). The results of
this RG showed that the HTQ-16 and HTQ-30 are likely to provide reliable scores across a large
diversity of populations. However, the re-experiencing, avoidance/numbing and arousal
subscales should be used with caution due to lower overall mean reliability estimates and
inadequate performance in certain samples. With regards to the validity of the HTQ symptom
scale, some authors have suggested that it may not be generalizable due to having been
developed from a psychiatric outpatient population (Hollifield et al., 2002). In a more recent
study, Rasmussen and colleagues (2015) found that the configural invariance, or overall fourfactor structure of the PTSD construct as measured by the HTQ-16, was acceptable among
individuals of various cultural groups. However, there were significant response-style differences
across cultures, as well as variations in the clinical thresholds of PTSD. Another study on the
factor structure of PTSD using the HTQ-16 compared three models across three groups living in
94
non-Western low- and middle-income countries (Michalopoulos,, et al., 2015). The results
showed that all models had an adequate fit for two cultural groups (i.e. Congolese and Burmese
samples), however none of the models had an adequate fit for the other (i.e. Northern Iraqi
sample). These important findings speak to the potential limitations of the construct validity of
the HTQ symptom scales.
Although these studies provide preliminary evidence of the validity of scores provided by
the HTQ, its overall construct validity properties with different cultural groups remains
unknown. Because this instrument is intended to be used across cultures, a further analysis of
empirical studies that have used the HTQ in different populations is warranted. Aggregating
empirical findings across studies and cultures will provide pertinent information regarding the
potential cross-cultural utility of the HTQ, thereby helping clinicians and researchers decide
whether it is an appropriate assessment tool for use with individuals of different cultural
backgrounds.
Validity and Validity Generalization
Validity refers to the extent to which an assessment instrument measures what it is
supposed to measure and is considered the most fundamental element of test development and
evaluation (American Educational Research Association [AERA], American Psychological
Association [APA], & National Council on Measurement in Education [NCME], 2014). An
instrument that has been evaluated as having good validity has variation in scores that reflects
the true variation in the construct that is being measured (Haynes et al., 2011).
Although there are various kinds of validity, the overarching form of validity is construct
validity. This type refers to the degree to which a test score reflects the construct of interest, and
includes both convergent and discriminant validity. Evidence for convergent validity refers to the
95
extent to which the scores of an instrument are related to scores of a different measure that is
intended to assess the same construct, or other variables that have theoretically established
associations with the construct (Messick, 1995). Evidence for discriminant validity refers to the
degree to which the scores of an instrument are unrelated to measures of unrelated constructs
(Haynes et al., 2011). Thus, low correlations between the two instruments can be indicative of
good discriminant validity.
Validity is not a static characteristic of an instrument but can vary according to context
and different sample characteristics. This was described as “situational specificity” by Schmidt
and Hunter (1998) and was used to support their hypothesis that there are statistical artifacts
(e.g., sampling error, measurement error) that can cause instruments to demonstrate adequate
evidence of validity in one situation and inadequate evidence of validity in another. This mirrors
the concept that reliability estimates are sample-dependent. Those using the instrument are
encouraged to examine validity in their specific population and purpose in using the instrument
(Kelly, O’Malley, Kallen, & Ford, 2005).
Establishing the validity of scores produced by an instrument is important as it also
addresses the validity of theories on the construct being measured and the operationalization of
the construct (Haynes et al., 2011; Kimberlin & Winterstein, 2008). The evidence of validity of a
measure continuously evolves as the theories of the constructs being measured change according
to new research developments (Haynes et al., 2011). Thus, evaluating the validity of measures
should be an ongoing process. Due to the controversy regarding the cross-cultural applicability
of the PTSD construct, and the abundance of new research regarding posttraumatic reactions
across the globe, the evaluation of instruments that are intended to measure these symptoms
across cultures, such as the HTQ, is of utmost importance.
96
Following their initial observations regarding situation specificity in validity, Schmidt
and Hunter (1977) developed a meta-analytic method called “validity generalization” (VG) to
examine the correlations between a specific test and a criterion. Specifically, VG estimates a
mean validity coefficient of a measure based on the correlations between the measure and
validity criteria of different studies. For example, to evaluate the discriminant validity properties
of a PTSD questionnaire, one could gather relevant studies that provide correlations (or other
measures of the relations between two variables) between this questionnaire and a measure of
impression management, and then calculate a mean correlation. This method can also provide
indications of the effects that might account for variability in the values found in a given sample.
The VG method has been primarily used in the field of employment selection and testing
(e.g., Dye, Reck, & McDaniel, 1993; Ones, Viswesvaran, & Schmidt, 2003) and has been
recommended in the Standards for Educational and Psychological Testing as a general approach
to evaluate validity (AERA et al., 2014). According to these standards, VG should preferably be
used when certain conditions are met: when there is a large meta-analytic database, when these
data represent more or less the type of situation to which the use of scores on the instrument will
be generalized, and when statistical artefacts are corrected and this correction produces validity
evidence that is consistent (AERA et al., 2014).
The Current Study
The present research was designed to evaluate the extent to which the HTQ symptom
scales have produced valid scores across studies found in the literature. Specifically, the
convergent and discriminant validity of these scales across samples were investigated by using
validity generalization analyses. This meta-analytic method determined the mean convergent and
discriminant validity estimates obtained on the HTQ symptom scales and examined the
97
variability in score validity across samples through moderator analyses. Specifically, moderator
analyses were conducted to assess the potential impact of sample and methodological
characteristics on scores produced by the HTQ symptom scales and subscales, thus further
examining the potential sources of variance. Similar to Study 1, variables that have shown mean
differences between groups with and without PTSD were examined.
Six sample characteristics were included as potential moderators. First, age was
considered a relevant potential source of variance as some studies suggest that the rates of
developing PTSD after exposure to a traumatic event may decline as age increases (Kessler,
Sonnega, Bromet, Hughes, & Nelson, 1995; Norris, 1992). Second, gender was also included as
epidemiological research shows that compared to men, women tend to experience fewer
traumatic events but are more likely to develop PTSD (Breslau et al., 1998; Perkonigg, Kessler,
Storz, & Wittchen, 2000).
Because this study was designed to evaluate the cross-cultural applicability of the HTQ
the impact of the cultural background of study participants of the study was included as a third
potential moderator. In extension to this, the cultural background of the sample in terms of the
constructs of collectivism versus individualism was examined for its possible moderating effects.
Such differences in perspective may have an impact on symptom reporting, and consequently, on
score validity. To categorize cultures as collectivist or individualist in the current study,
Hofstede’s dimensions of national culture were used (Hofstede, Hofstede, & Minkov, 2010). As
a result, only the countries in Hofstede’s research studies were categorized and included in the
moderator analyses.
Another sample characteristic that was included as a possible moderator was the type of
trauma that was separated into two categories: intentional trauma (e.g., war, torture, sexual
98
assault) and unintentional trauma (e.g., natural disasters, accidents). These categories were
chosen because studies have shown that individuals who are exposed to intentional traumas had
worse health outcomes than did individuals who experienced traumas that were not deliberate
(Lange et al., 2003; Matthieu & Ivanoff, 2006; Van der Velden et al., 2006). Whether study
participants were forcibly displaced was also included as a sixth moderator variable, as
researchers tend to find higher rates of psychological difficulties among individuals who have
been displaced as the result of war as compared to non-displaced groups (Fazel, Wheeler, &
Danesh, 2005; Porter & Haslam, 2005; Steel et al., 2009).
Two methodological variables were included as potential moderators. The sites from
which the samples were recruited were examined. Psychometric properties of instruments can be
dependent on the severity of the disorder measured (Haynes et al., 2011) and, as symptoms are
generally more severe in individuals who are seeking psychotherapy or medical treatment than
they are in individuals from non-clinical samples, the effect of recruit site (i.e., clinical,
community, or mixed clinical and community) was considered. Convergent validity estimates
should be highest and discriminant validity estimates should be lowest in studies involving
samples for which the HTQ was originally intended to be used. Accordingly, moderator analyses
were conducted to determine whether having a study sample that was culturally similar to the
samples used in the development of the HTQ affects validity estimates.
Method
Literature Search and Data Collection
The details of the search strategy to identify potential studies to be included in the VG
and inclusion criteria used are reported elsewhere (Darzi & Hunsley, 2017). A flow chart of the
data reduction process can be seen in Figure 1.
99
To ensure the statistical independence of the studies to be included in the analyses,
multiple publications on the same data set were identified. The detection heuristic developed by
Wood (2008) was used as a guideline to help identify same samples. This strategy included
determining whether the study and sample characteristic were similar in the suspected multiple
publications and verifying if the same authors were included in the studies. To help decide which
study to eliminate, several factors were considered, such as the amount of descriptive data
reported and the reporting of validity coefficients. Furthermore, when studies provided validity
coefficients of different samples within the same study, each sample was identified as separate
and retained for the analyses.
In summary, there was a total of 384 studies that matched the inclusion criteria and a total
of 116 studies that had convergent and/or discriminant validity estimates. After removing studies
using the same samples, studies that included modified versions of the HTQ symptom scales
(i.e., added or removed items), and unusable validity information (such as a lack of information
necessary to compute a validity coefficient), there were a final total of 78 studies and 79 samples
to be included in the analyses.
Coding of descriptive data and moderators. The relevant articles were coded for
convergent and discriminant validity information (i.e., construct, measure, format, values),
sample characteristics, and methodological characteristics for the HTQ symptom scales. The
initial coding was performed by the first author (a clinical psychology doctoral student with
previous coding experience) and was facilitated by a coding manual (see Appendix A) that was
based on a review of the literature on trauma, cross-cultural measurement and validity
generalization. The data were entered on a coding sheet (see Appendix B). The study and
sample variables were used in the primary descriptive statistics of the overall pool of studies
100
included in the meta-analyses. Some variables were also examined as potential moderators.
Study and methodological characteristics included the following categorical variables: (a)
language of administration of the HTQ; (b) country of study; (c) type of sampling (community;
seeking mental health services; seeking medical treatment sample); (d) version of the HTQ
(original or adapted version), (e) whether an adaptation was needed for the sample (yes or no);
(f) adaptation procedure used (cultural adaptation only, translation only, or both); (g) adaptation
procedure followed the developers’ recommendations (yes or no), (h) adaptation procedure
followed other experts’ recommendations (yes or no); (i) translation procedure used (oral
translation; forward translation; back translation; blind back translation); (j) a consensus
approach was used during adaptation (yes or no); (k) the adaptation included a pre-test (yes or
no); and (l) type of administration (clinician administered; non-clinician administered; selfreport).
Sample characteristics included the following continuous variables: sample size, the
mean age of participants, and gender composition of the sample (percentage of males). Sample
characteristics also included the following categorical variables: (a) country of origin of
participants; (b) country of origin’s cultural orientation (individualist or collectivist) based on
Hofstede’s categorization (only countries classified by Hofstede were coded); (c) type of trauma
experienced (intentional or unintentional); and (d) sample consisted of forcibly displaced persons
(yes or no).
The second author (a clinical psychologist and professor with extensive experience in
meta-analysis) coded 20% of the studies that were included in the meta-analyses to assess the
reliability of the coding procedures. The inter-coder reliability analyses were performed using
SPSS version 20. The kappa (k) statistic was used for the categorical variables and intraclass
101
coefficients (ICC) were used to assess the inter-coder reliability of continuous variables. These
analyses were important to help eliminate random variation and increase the reliability and
power of the results (Dieckmann, Malle, & Bodner, 2009). The inter-coder reliability had a
rating of “good” as per Hunsley and Mash’s (2008) criteria with a mean k = 0.77 and ICC
coefficient of 0.85. Several discrepancies between coders were caused by unclear reporting of
data in the studies, and vague descriptions of the methodological and adaptation procedures.
Discrepancies were discussed until a unanimous agreement between the coders was reached.
Data Analysis
Calculating mean effect sizes. Several steps were followed to develop convergent and
discriminant validity coefficient estimates for the HTQ-16 and HTQ-30s. The first step was to
enter the validity information of each sample (sample identifier, sample size, and validity
coefficient) into the Comprehensive Meta-Analysis software version 2.2.064 (CMA; Borenstein,
Hedges, Higgins, & Rothstein, 2005). Values of indexes that reported the variance shared
between a HTQ symptom scale and a convergent or discriminant validity construct (e.g., r, d)
were included in the analyses. Effect sizes that were not reported as correlations were converted
into r with the CMA software. A Fisher’s r-to-z transformation was then performed to
compensate for the high level of skewness and non-normal distribution usually found with
correlations. All further analyses were performed with these transformations (see Borenstein,
Hedges, Higgins, & Rothstein, 2009).
A relative weight and a 95% confidence interval were calculated for each sample. This
relative weight, or inverse-variance, is primarily based on the sample size (i.e., smaller sample
sizes are generally assigned less weight). An overall mean effect size, its statistical significance
and upper and lower 95% confidence intervals were then calculated. The Fisher’s z values were
102
then converted back to their original r values to facilitate interpretation.
Random-effects model. A random-effects model was chosen to allow for the true effect
to be different across samples. This model is in contrast to the fixed-effect model that assumes
that the samples have a common true effect and that the differences in observed effects are due
entirely to sampling error alone (e.g., Borenstein, Hedges, Higgins, & Rothstein, 2010). Because
the random-effects model considers that the effect sizes may come from different heterogeneous
populations, this model was considered the most suitable to include in the VG analyses which
were comprised of highly heterogeneous samples.
Heterogeneity of effects. A Q test of homogeneity (or Cochran’s Q; Cochran, 1954) and
I2 index (Higgins & Thompson, 2002) were used to evaluate the degree of dispersion of validity
estimates around the mean validity coefficients. The Q statistic evaluates the null hypothesis that
the studies included in the meta-analysis have a common effect size and a significant Q statistic
suggests that moderator analyses should be conducted to help find the sources of heterogeneity.
The I2 index assesses the proportion of the observed variance that is caused by true heterogeneity
rather than being due to chance alone (Higgins & Thompson, 2002). The following benchmarks
proposed by Higgins and Thompson (2002) were used to assess the different levels of
heterogeneity and interpret the I2 index: 25% = low, 50% = moderate, and 75% = high.
Possibility of publication bias. To examine the possible influence that unpublished
effects may have on the mean validity values obtained in the meta-analyses, an Orwin’s fail-safe
N (Orwin, 1983) was calculated for each meta-analysis. The result of this procedure is an
estimate of the number of studies with estimate of .01 would be needed lower the average
validity estimate to under .2 (which is the cut-off for a small effect). Funnel plots were not used
103
to assess publication bias because of the risk of visual misinterpretation (Lau, Ioannidis, Terrin,
Schmid, & Olkin, 2006).
Analysis of moderators. Moderator analyses were performed to examine possible
sources of variance on estimates that had statistically significant heterogeneity. Mixed effects
(method of moments) meta-regression analyses were performed to examine the relations between
continuous variables and score validity. For categorical variables, a series of random-effects
subgroup analyses were used to examine the relations between the validity coefficients and study
and sample characteristics. Because some of the variables had an insufficient number of studies
per subgroup (i.e., less than 2) some categories were merged together to help increase the power
of the analyses. For instance, “culture” was separated according to larger geographical cultural
groups according to continent of origin (e.g., Asian, European, African), and “types of trauma”
were categorized as “intentional trauma” and “unintentional trauma”.
Results
Overview of Studies included in the Meta-Analyses and Descriptive Statistics of Potential
Moderators
The present meta-analyses included 79 samples from 78 studies, representing data from
21,156 individuals (see Appendix D for the list of studies included). The mean sample size was
280, the mean age of participants was 37 years, and, on average, 46.6% of study participants
were male. The studies were conducted in over 36 countries, with samples coming from over 35
different countries. Consistent with this, the HTQ symptom scales were administered in over 42
different languages and dialects.
Of the study samples that reported the sampling procedure (n = 79), 67.1% recruited
participants from the community, 22.8% recruited participants who were seeking mental health
104
services, 6.3% recruited participants who were seeking medical treatment, and 3.8% recruited
from multiple sources. Fifty-nine percent of the samples were composed of forcibly displaced
individuals (i.e., refugees, asylum seekers, internally displaced people) as opposed to natives or
immigrants. The most common types of traumatic events were intentional traumas, with 79.7%
of samples having all or most individuals who experienced these traumas. Unintentional traumas
were experienced by 12.7% of the samples and the remainder 7.6% of samples consisted of
individuals who had been exposed to both types of traumas or none.
In terms of how the HTQ symptom scales were used within the independent samples,
74.7% (n = 59) involved adaptations, although 86.1% (n = 68) required the use of an adapted
version of the measure because of the nature of the samples. Of those study samples that used
adaptations, 66.1% were translations only, 10.2% were both translation and cultural adaptation,
and 23.7% made no mention of how they adapted the measure. Of the studies that did not
culturally adapt the instrument, none of them reported a cultural verification process that would
ensure that the measure was suitable for their sample. Regarding translation procedures, of those
that reported translating the instrument (n = 45), 20% indicated using a forward translation,
24.4% used in vivo oral translations, 35.6% used a back translation approach, and 20% used a
blind back translation approach. Among the studies that reported their adaptation procedures (n =
45), 6.7% reported using the adaptation recommendations provided by the developers of the
HTQ and 26.7% reported following the recommendations from other authorities on measure
adaptation. With regards to steps taken to ensure the validity of the translation or adaptation,
37.8% used a consensus approach and 22.2% conducted a pre-test. Finally, of those who reported
the administration procedures used (n = 75), 40% were administered via self-report, 24% were
administrations done by non-clinicians, and 37.3% were clinician-administered.
105
Validity Generalization Analyses
Convergent validity. Altogether, there were 72 convergent validity coefficients entered
in the meta-analyses. When a study reported two or more correlations for the same construct,
they were averaged to ensure statistical independence. Consistent with procedures used in other
VG studies (e.g., Kinicki, McKee-Ryan, Schriescheim, & Carson, 2002), the various correlates
were grouped into broad correlate categories. These categories included the following constructs:
(a) cumulative trauma (e.g., number of different trauma types experienced as per the HTQ events
scale); (b) other posttraumatic stress measure (e.g., Impact of Event Scale [IES; Horowitz,
Wilner, & Alvarez, 1979); (c) cultural expressions of trauma (e.g., additional cultural items of
the HTQ symptom scales); and (d) dissociation measure (e.g., dissociation subscale of the
Trauma Symptom Checklist [TSC; Briere &Runtz, 1989]). Descriptive statistics and results of
the convergent VG of each of these categories are presented in Table 1. The mean convergent
validity coefficients for the HTQ-16 ranged from r = .32, p < 0.001, 95% CI [0.26, 0.38] for the
“cumulative trauma” category to r = .75, p < 0.001, 95% CI [0.59, 0.86] for “cultural expressions
of trauma.” For the HTQ-30, a meta-analysis could only be conducted with the “cumulative
trauma” category because of the insufficient number of correlations available for the other
constructs. The mean convergent validity estimate for this category was r = .26, p < 0.001, 95%
CI [0.17, 0.34].
All convergent VG analyses had a significant Q statistic. The significant Q statistics and
high I2 suggest that there is an important amount of variance across samples for those specific
VG results, therefore further examination of potential study and sample characteristics that
account for this variability was warranted.
106
Discriminant validity. A total of 53 discriminant validity coefficients were entered in
the meta-analyses examining discriminant validity. The various correlates were separated into
the following correlate categories: (a) depression and anxiety symptoms measures (e.g., Hopkins
Symptom Checklist-25 [HSCL-25: Mollica, Wyshak, de Marneff, & Lavelle, 1987]); (b) anxiety
symptoms measures (e.g., anxiety subscale of the General Health Questionnaire (GHQ; Goldberg
& Williams, 1988); (c) depression symptoms measures (e.g., Beck Depression Inventory-II
[BDI; Beck, Steer, & Brown, 1996]); and (d) somatization measures (e.g., HSCL somatization
scale). Descriptive statistics and results of the discriminant VG of each of these categories are
presented in Table 2. The mean discriminant validity coefficients for the HTQ-16 ranged from r
= .43, p < 0.001, 95% CI [0.28, 0.56] for the “somatization measures” category to r = .80, p <
0.001, 95% CI [0.77, 0.83] for “depression and anxiety symptoms measures.” The mean
discriminant validity coefficients for the HTQ-30 ranged from r = .45, p = 0.239, 95% CI [-0.31,
0.86] for the “somatization measures” category to r = .79, p < 0.001, 95% CI [0.61, 0.89] for
“depressive and anxiety symptoms measures.”
The “depression and anxiety symptoms measures” category for the HTQ-16 and the
“depression symptoms measures” category for the HTQ-30 had nonsignificant Q statistics. The
significant Q statistics and high I2 of the other discriminant VG analyses indicated a need to
further examine the role of potential moderators.
Publication bias. Because the CMA software cannot compute an Orwin’s fail-safe N for
meta-analyses with fewer than 3 studies, three of the VG studies could not be tested for
publication bias. A generally accepted interpretation of fail-safe N is that a meta-analytic result is
unlikely to be affected by publication bias if the calculated N ≥ 5k + 10 (Orwin, 1983). Based on
this formula, none of the VG results in this study are robust to publication bias, and should be
107
considered as preliminary results until additional data on these categories of validity coefficients
are available.
Moderator analyses. A summary of the moderating effects is presented in Table 3.
Samples that had missing information for a specific variable were excluded from these analyses.
Meta-regression analyses (mixed effects method of moments) were conducted to assess the
potential impact of age (mean age of sample) and gender (percentage of participants who were
males). Age was a significant moderator for the association between the “somatization measure”
category and HTQ-16, β = -0.092, Qmodel = 4.60, p < 0.05, 95% CI [0.00, 0.03]. Specifically,
higher associations between the HTQ-16 and somatization measures were found with lower
mean ages of study samples. Gender was also a moderator for this association, β = -0.014, Qmodel
= 10.49, p < 0.05, 95% CI [-0.02, -0.01], with study samples comprised of mainly female
participants having higher validity estimates.
With regards to the subgroup analyses, the “forcibly displaced sample” and “country of
origin’s cultural orientation” variables were not significant moderators for any VG result. The
“cultural group” variable was significant for the convergent association between the “cumulative
trauma” variable and the HTQ-30. Specifically, “African cultures” had the lowest mean validity
coefficient (r = 0.06, p = 0.100, 95% CI [-0.01, 0.13]) and “Asian cultures” had the highest mean
validity coefficient (r = 0.40, p < 0.001, 95% CI [0.32, 0.48]). The summary effect sizes for
“cultural group” were also significantly different from each other for the discriminant association
between the “somatization measures” variable and the HTQ-16. In this case, “European culture”
had the lowest coefficient (r = 0.17, p = 0.088, 95% CI [-0.03, 0.35]) and “Asian culture” the
highest (r = 0.60, p < 0.001, 95% CI [0.56, 0.64]). The “recruitment setting” variable had a
significant effect for two convergent VG results: (a) the association between “cumulative
108
trauma” and the HTQ-30, where the lowest mean coefficient was for “medical setting” (r = 0.04,
p = 0.356, 95% CI [-0.05, 0.12]) and the highest for “community setting” (r = 0.33, p < 0.001,
95% CI [0.26, 0.40]), and (b) the association between the “other posttraumatic stress measure”
correlate and the HTQ-16, with “mental health setting” having the lowest coefficient (r = 0.28, p
< 0.05, 95% CI [0.07, 0.47]) and “medical setting” the highest (r = 0.86, p < 0.001, 95% CI
[0.67, 0.94]). The “type of trauma” variable was also a moderator for both of these convergent
VG associations. Specifically, “intentional traumas” had the highest coefficient for the
association between “cumulative trauma” and the HTQ-30 (r = 0.28, p < 0.001, 95% CI [0.20,
0.35], with a lower coefficient found for “unintentional traumas” (r = 0.04, p = 0.717, 95% CI [0.17, 0.25]). However, for the association between “other posttraumatic stress measures” and the
HTQ-16, the coefficient for “intentional traumas” was significantly lower (r = 0.58, p < 0.001,
95% CI [0.43, 0.70]) than the coefficient for “unintentional traumas” (r = 0.78, p < 0.001, 95%
CI [0.65, 0.86]). Finally, for the association between the “cumulative trauma” correlate category
and the HTQ-30, the coefficient for “original HTQ sample” was higher (r = 0.43, p < 0.001, 95%
CI [0.26, 0.57]) than the coefficient for the “not original sample” (r = 0.21, p < 0.001, 95% CI
[0.12, 0.29]).
Discussion
The goals of this study were to summarize the available data regarding the convergent
and discriminant validity of the HTQ symptom scales. Of the 116 studies that provided validity
estimates, 78 of them were included in the analyses. These studies represented 79 independent
samples. The samples had very diverse compositions, with many languages, cultures, and
countries of origin represented in the data. Although most authors recognized the need to adapt
the HTQ, the manner in which they proceeded to make changes to the original did not generally
109
follow the adaptation recommendations set by the developers of this measure (i.e., culturally
adapting the instrument, using blind-back consensus approach to translation, pre-testing the
adapted version). These observations are similar to the studies included in the RG of Study 1 and
reflect the frequent findings that (a) the process of translating and cross-culturally validating
instruments is usually not a priority in clinical research protocols (Sperber, 2004) and (b) simple
forward or back translations without a consensus approach are common in clinical cross-cultural
research (Maneesriwongul & Dixon, 2004).
Overall, the results of the meta-analyses suggest that there is some evidence supporting
the convergent validity of the HTQ-16. Unfortunately, the same cannot be said for the
convergent validity of the HTQ-30. Because of the limited number of convergent correlates
found in the research literature for this scale, there was only one category (i.e., cumulative
trauma) that had sufficient data to compute effect size statistics. However, even for the HTQ-16,
the results of the publication bias analysis indicated that substantially more research is required
before one can consider that evidence for this measure’s convergent validity has been
established.
According to Hemphill’s (2003) r-value effect size benchmarks for psychology, all the
convergent validity correlates for the HTQ-16 had large effect sizes (i.e., r > .30). The strongest
association was with cultural expressions of trauma and the weakest was with cumulative
trauma. Although there were only three studies included in the cultural expressions of trauma
analyses, the relatively high mean correlation indicates that the cultural symptoms endorsed by
the samples represented in these studies (i.e., Quechuan/Peruvian, Cambodian, and West Papuan
survivors of civil war) are closely related to the DSM-IV (APA, 1994) PTSD symptoms
presented in the HTQ-16. However, because of the low power of these results we cannot assume
110
that cultural expressions of trauma in other cultures are also closely related to PTSD as measured
by this scale. On the other hand, the high mean correlations found with other posttraumatic stress
measures and dissociation measures provide evidence of the convergent validity of the HTQ-16
across various cultures. Similar to other PTSD measures (e.g., Dermichyan, Goenjian, &
Khachadourian, 2015; Ruggiero, Del Ben, Scotti, & Rabalais, 2003), “cumulative trauma” had
low correlations for both HTQ symptom scales. Because of variability in the conditional risk of
PTSD across cultures, measures of traumatic events may not provide an optimal convergent
validity correlate. In addition, it appears that trauma subtypes, rather than cumulative trauma,
may better predict mental health outcomes (Arnetz et al., 2014).
With regards to the discriminant validity properties of the HTQ symptom scales, the
results of the VG analyses are not as encouraging. More specifically, the evidence of
discriminant validity of the HTQ-30 was rather weak. The associations between these related
constructs and the HTQ-30 were very strong, especially with depressive symptoms with a mean
correlation of .78. However, these results are not entirely surprising considering that several
items of the HTQ-30 are characteristic of symptoms of depression (e.g. “hopelessness”, “feeling
guilty for having survived”). These findings suggest that the HTQ-30 may be a more general
measure of psychological distress rather than a specific measure of PTSD symptoms.
In terms of the HTQ-16, all discriminant correlates of this scale had large effect sizes,
according to Hemphill’s (2003) benchmarks. Although it is expected that the HTQ-16 would
have a certain degree of association with measures of depression and anxiety due to high rates of
comorbidity and symptom overlap, the associations between these constructs and the HTQ-16
were quite strong (i.e., mean r > .60). However, these results appear to be comparable to other
commonly used DSM-IV correspondent measures of PTSD, such as the Posttraumatic
111
Diagnostic Scale (PDS; Foa, Cashman, Jaycox, & Perry, 1997) and the PTSD Checklist-Civilian
version (PCL-C; Weathers, Litz, Herman, Huska, & Keane, 1993). These instruments have
shown variable performances in terms of discriminant validity in separate single studies, with
some showing weaker associations than the HTQ-16 with constructs such as depression and
anxiety (e.g., Adkins, Weathers, McDevitt-Murphy, & Daniels, 2008; Kornblith et al., 2003), and
others showing very strong associations, especially with depression symptoms (i.e. r > .70) (e.g.,
Carter-Visscher et al., 2010; Foa et al., 1997). However, until meta-analytic investigations are
also conducted with these instruments, it is difficult to assess whether the discriminant validity
properties of the HTQ-16 are indeed statistically comparable.
Nevertheless, questions about the discriminant validity of PTSD measures have been
raised for many instruments, as PTSD measures have been found to be highly correlated with
measures of anxiety and, especially, depression symptoms (e.g., Foa et al., 1997). This pattern
appears to be reflective of Miller, Fogler, Wolf, Kaloupek and Keane’s (2008) model of
psychiatric comorbidity that posits that PTSD shares more in common with depressive symptoms
such as melancholy, apathy and rumination than it does with anxiety and avoidance. Consistent
with this model, relatively weaker associations have been found between PTSD and alcohol
abuse and psychopathy, than anxiety and depressive symptoms (e.g., Bovin et al., 2016),
suggesting that these types of constructs may be more appropriate for examining the discriminant
validity of a PTSD measure. Unfortunately, we were unable to examine this possibility, as there
were so few reports in the literature of correlations between these specific constructs and the
HTQ symptom scales.
In order to assess the overall construct validity of the HTQ-symptoms scales, it is
important to consider the evidence of both the convergent and discriminant validity together.
112
Unfortunately, there was an insufficient amount of convergent validity correlates for the HTQ-30
to fully evaluate its construct validity. Although there are no specific benchmarks to evaluate and
compare convergent and discriminant validity coefficients in meta-analysis, it is expected that
similar constructs would have significantly higher correlations than would theoretically less
related constructs (Campbell & Fiske, 1959). This is not the case for some correlates of the
HTQ-16. For instance, the mean estimates for this scale with depressive symptom measures and
other posttraumatic stress measures were almost identical, with considerable overlap among the
confidence intervals for these two associations. The lowest mean estimates for the discriminant
validity of the HTQ symptom scales were with somatization measures, but even here the values
were relatively high. This strong association appears to be consistent with results from other
research not involving the HTQ that have found high associations between PTSD and
somatization (e.g., Beckham et al., 1998; North, Kawasaki, Spitznagel, & Hong, 2004).
However, because other measures of DSM-IV PTSD have shown to have significantly higher
correlations with other measures of posttraumatic symptoms (e.g. Ruggiero et al., 2003) than
with measures of depression and anxiety symptoms, this calls into question the overall construct
validity of the HTQ-16. Despite these findings, further data is needed before drawing definite
conclusions, as the VG results in this study are not robust to publication bias.
Moderators
The statistically significant Q and high I2 tests found with some of the mean validity
estimates suggest that the variability among them is heterogeneous and is not due to sampling
error alone. The high degree of variance that is caused by true heterogeneity of the mean
estimates indicated that the examination of sample and methodological characteristics could help
explain this variability.
113
Overall, the moderator analyses indicated that some study and sample characteristics had
a moderating or predictive effect on both the HTQ symptom scales. In summary, cultural group,
recruitment site, type of trauma, and whether the sample was a sample for which the HTQ was
originally validated affected some of the convergent validity estimates (i.e. cumulative trauma
and other posttraumatic stress measures). In terms of discriminant validity, age, gender, and
cultural group played a role in the heterogeneity of the estimates for the somatization measures
correlate category. Contrary to the hypotheses, neither having a sample comprised of forcibly
displaced individuals nor the participants’ country of origin’s cultural orientation had a
moderating effect on the convergent and discriminant validity of either scale.
Specifically regarding the HTQ-16, there were moderating effects when the scale was
correlated with other posttraumatic stress measures and measures of somatization. More
precisely, the convergent validity of the scale when correlated with other posttraumatic stress
measures was higher with samples recruited in medical settings and samples that had
experienced unintentional traumas. As for the discriminant validity of this scale when it is
compared to somatization measures, lower estimates were obtained with older participants,
samples with European backgrounds, and samples mostly comprised of men. These moderating
effects appear to be consistent with other research that has shown that age, gender, and culture
play a role in the development of PTSD (e.g., Ditlevsen & Elklit, 2010; Kessler et al., 1995;
Kessler et al., 1999; Norris 1992; Pietrzak, Goldstein, Southwick, & Grant, 2011) and
somatization (e.g., Kirmayer & Young, 1998; Leiknes, Finset, Moum, & Sandanger, 2007).
Lower estimates were also found with samples of European background, meaning that the
discriminant validity was weaker in non-European samples. Because of the high risk of
publication bias and the low power of these moderating effects due to relatively small numbers
114
of studies included in the analyses, these results should be considered as preliminary; they do not
indicate that there are significant limitations with the HTQ-16. However, further research is
needed to make inferences regarding the convergent validity of the HTQ-16, especially in
samples that have been recruited in community or mental health treatment settings, and with
individuals who have experienced intentional traumas. Similarly, more research is needed to
assess the discriminant validity of the HTQ-16 in younger samples, samples comprised mostly of
women and with individuals of non-European backgrounds that had weaker results in this study.
As for the HTQ-30, moderator analyses could only be conducted with the cumulative
trauma correlate and significant effects were found for this variable. The convergent validity
estimates were higher in samples comprised of individuals of Asian backgrounds, in samples that
were recruited in community settings, with individuals who had experienced intentional trauma,
and with samples similar to those on which the HTQ was originally validated. These results are
not surprising due to the fact that the 14 cultural items of the HTQ-30 were based on cultural
expressions of trauma in Asian samples (specifically Indochinese populations) who had
experienced intentional traumas such as war. These results further justify the importance of
adapting the cultural items for the each specific culture. On the other hand, these estimates were
high in samples recruited in community settings, whereas the original validation samples were
recruited in mental health treatment settings. Once again, because of the risk of publication bias
and low power of the moderator analyses, these findings should be interpreted with caution.
However, researchers should note that it may be possible that the convergent validity of the
HTQ-30 is compromised in non-Asian samples, samples that are recruited in clinical settings,
samples that have experienced unintentional traumas, and samples for which the HTQ was not
originally validated (i.e., non-Indochinese refugee populations).
115
Limitations of the Study
Some of the limitations of this study stem from the cited reference search strategy, that
may not have identified all of the studies and dissertations that have used the HTQ symptom
scales. In addition, the coding process was challenging at times because the information reported
in the studies was often vague and difficult to locate. More specifically for the VG, problems of
inconsistent reporting and lack of data were evident when extracting validity coefficients from
identified studies. For instance, correlation values and/or matrices were not always available, and
supplementary analyses were sometimes required to calculate the bivariate associations. In many
cases, authors reported conducting bivariate analyses but did not provide sufficient information
to include these findings in the meta-analyses.
Another limitation of this study comes from the grouping of estimates into larger
correlate categories. Although this procedure helped increase statistical power, the correlates
included in these broad categories were somewhat dissimilar because of the wide variety of
instruments that were used to measure them. For instance, there were several depression
symptom instruments that were used to measure the correlates included in the broad category of
“Depression symptoms measures,” such as the Indochinese versions of the HSCL-25 (Mollica et
al., 1987) and the BDI-II (Beck et al., 1996). Thus, the correlate categories did not represent
completely homogeneous constructs, which may have affected the results.
In addition, because the VG analyses included correlations, it is important to note that it
was the ability of the HTQ symptom scales to assess posttraumatic symptom severity and not
PTSD caseness that was evaluated. Future users of the HTQ in research settings are thus
encouraged to evaluate the ability of the scale to diagnose PTSD by conducting receiver
116
operating characteristic (ROC) curve analyses, for instance, to establish appropriate cut-off
scores for their specific sample.
Implications and Recommendations
In conclusion, several important implications and recommendations stem from this study.
First, consistent with established publication standards, authors are strongly encouraged to report
relevant and sufficient information in the methods and results sections of their study. As the
accuracy of meta-analyses is dependent on the primary studies on which they are based, missing
information and data in primary studies ultimately has a negative impact on the quality of metaanalyses.
Furthermore, the results obtained from the VG analyses focused on discriminant validity
raise doubts about the overall construct validity of HTQ symptom scales and the construct
validity of PTSD as conceptualized in the DSM. The main concern was the high degree of
similarity between the convergent and discriminant validity estimates across a wide range of
disparate variables. Although some studies support the hypothesis that removing symptoms of
other constructs theorized to overlap with PTSD, such as mood and other anxiety disorders,
increases the validity of the PTSD construct (e.g., Elhai et al., 2015; Spitzer, First, & Wakefield,
2007), others have found that removing these symptoms does not reduce comorbidity with these
other disorders (e.g., Grubaugh, Long, Elhai, Frueh, & Magruder, 2010).
Given that PTSD is highly comorbid and has several symptoms that overlap with
depression and anxiety, convergent and discriminant measures of PTSD should perhaps be
mapped on a continuum, rather than be considered as distinct categories. Westen and Rosenthal
(2003) proposed methods of quantifying construct validity based on contrast analysis that
provide effect sizes that indicate the degree to which an observed pattern of convergent-
117
discriminant correlations is similar to a theoretically based predicted pattern of correlations.
These pattern-mapping approaches to the evaluation of construct validity are particularly
relevant in the case of PTSD, where patterns of relationships with other constructs have been
identified in the literature (e.g., Bovin et al., 2015; Miller et al., 2008). Aggregating the findings
of these studies could help establish theoretically based patterns of relationships between PTSD
and other constructs and serve as benchmarks to interpret the magnitude of observed
correlations.
Because the theoretical conceptualization of PTSD and the understanding of its relation
to other constructs are in continuous development, it is difficult to evaluate whether poor
psychometric properties of an instrument are truly a function of the scale, or rather the PTSD
construct. Nonetheless, the aggregate findings of the VG analyses, although preliminary, indicate
that the convergent validity properties of the HTQ-16 are supported to some extent, and that this
scale may perform questionably in regards to discriminant validity. The present results,
combined with the findings of other studies that have evaluated the validity of the HTQ symptom
scales (e.g., Rasmussen et al., 2015), raise some doubts upon the overall construct validity of this
scale. In addition, the moderator analyses suggest that caution should be taken in designing
future studies that include variables that had significant moderating effects (e.g., convergent
validity estimates were higher in samples that experienced unintentional traumas and samples
that were recruited in medical treatment settings). Furthermore, there is limited support for either
the convergent or discriminant validity of the HTQ-30. Because of the mixed results for the
HTQ-16 and the fact that the HTQ symptom scales are commonly considered as strong crosscultural measures of trauma symptoms in the literature, considerably more work is needed to
establish the validity of these scales.
118
References
Adkins, J. W., Weathers, F. W., McDevitt-Murphy, M., & Daniels, J. B. (2008). Psychometric
properties of seven self-report measures of posttraumatic stress disorder in college
students with mixed civilian trauma exposure. Journal of Anxiety Disorders, 22, 13931402. doi: 10.1016/j.janxdis.2008.02.002
American Educational Research Association, American Psychological Association, & National
Council on Measurement in Education (2014). Standards for educational and
psychological testing. Washington, DC: AERA.
American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders
(3rd ed. revised). Washington, DC: Author.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders
(4th ed.). Washington, DC: Author.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders
(5th ed.). Washington, DC: Author.
Arnetz, B. B., Broadbridge, C. L., Jamil, H., Lumley, M. A., Pole, N., Barkho, E., … Arnetz, J.
E. (2014). Specific trauma subtypes improve the predictive validity of the Harvard
Trauma Questionnaire in Iraqi refugees. Journal of Immigrant and Minority Health, 16,
1055-1061. doi: 10.1007/s10903-014-9995-9
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory-II.
San Antonio, TX: Psychological Corporation.
Beckham, J. C., Moore, S. D., Feldman, M. E., Hertzberg, M. A., Kirby, A. C., & Fairbank, J. A.
(1998). Health status, somatization, and severity of posttraumatic stress disorder in
119
Vietnam combat veterans with posttraumatic stress disorder. American Journal of
Psychiatry, 155, 1565-1569. doi: 10.1176/ajp.155.11.1565
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2005). Comprehensive
meta-analysis version 2. Englewood, NJ: Biostat.
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction
to meta-analysis. Chichester, UK: John Wiley & Sons, Ltd.
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2010). A basic
introduction to fixed-effect and random-effects models for meta-analysis. Research
Synthesis Methods, 1, 97-111. doi: 10.1002/jrsm.12
Bovin, M. J., Marx, B. P., Weathers, F. W., Gallagher, M. W., Rodriguez, P., Schnurr, P. P., &
Keane, T. M. (2016). Psychometric properties of the PTSD Checklist for Diagnostic
and Statistical Manual of Mental Disorders–Fifth Edition (PCL-5) in veterans.
Psychological Assessment, 28, 1379-1391. doi: 10.1037/pas0000254
Briere, J., & Runtz, M. (1989). The Trauma Symptom Checklist: Early data on a new scale.
Journal of Interpersonal Violence, 4(2), 151-163.
Breslau, N., Kessler, R. C., Chilcoat, H. D., Schultz, L. R., Davis, G. C., & Andreski, P. (1998).
Trauma and posttraumatic stress disorder in the community: The 1996 Detroit Area
Survey of Trauma. Archives of General Psychiatry, 55, 626– 631. doi:
10.1001/archpsyc.55.7.626
Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the
multitrait-multimethod matrix. Psychological Bulletn, 56, 81-105. doi: 0.1037/h0046016
Carter-Visscher, R., Polusny, M. A., Murdoch, M., Thuras, P., Erbes, C. R., & Kehle, S. M.
(2010). Predeployment gender differences in stressors and mental health among U.S.
120
National Guard troops poised for Operation Iraqi Freedom deployment. Journal of
Traumatic Stress, 23, 78-85. doi: 10.1002/jts.20481
Cochran, W. G. (1954). The combination of estimates from different experiments. Biometrics,
10, 101–129. doi: 10.2307/3001666
Darzi, C. M., & Hunsley, J. (2017). The Harvard Trauma Questionnaire: A reliability
generalization study. Manuscript in preparation.
Dermichyan, A., Goenjian, A. K., & Khachadourian, V. (2015). Factor structure and
psychometric properties of the Posttraumatic Stress Disorder (PTSD) Checklist and
DSM-5 PTSD symptom set in a long-term post-earthquake cohort in Armenia.
Assessment, 22, 594-606. doi: 10.1177/1073191114555523
Dieckmann, N., Malle, B., & Bodner, T. (2009). An empirical assessment of meta- analytic
practice. Review of General Psychology, 13, 101-115. doi:10.1037/a0015107
Ditlevsen, D. N., & Elklit, A. (2010). The combined effect of gender and age on posttraumatic
stress disorder: Do men and women show differences in the lifespan distribution of the
disorder? Annals of General Psychiatry, 9. doi: 10.1186/1744-859X-9-32
Dye, D. A., Reck, M., & McDaniel, M. A. (1993). The validity of job knowledge measures.
International Journal of Selection and Assessment, 1(3), 153-157.
Elhai, J. D., Contractor, A. A., Tamburrino, M., Fine, T. H., Cohen, G., Shirley, E., …, Galea, S.
(2015). Structural relations between DSM-5 PTSD and major depression symptoms in
military soldiers. Journal of Affective Disorders, 175, 373-378. doi:
10.1016/j.jad.2015.01.034
Fazel, M., Wheeler, J., & Danesh, J. (2005). Prevalence of serious mental disorder in 7000
refugees resettled in western countries: A systematic review. Lancet, 365, 1309-1314.
121
doi: 10.1016/S0140-6736(05)61027-6
Foa, E., Cashman, L., Jaycox, L., & Perry, K. (1997). The validation of a self-report measure of
PTSD: The Posttraumatic Diagnostic Scale. Psychological Assessment, 9, 445-451. doi:
10.1037/1040-3590.9.4.445
Gagnon, A. J., Tuck, J., & Barkun, L. (2004). A systematic review of questionnaires measuring
the health of resettling refugee women. Health Care for Women International, 25, 111149. doi: 10.1080/07399330490267503
Goldberg, D., & Williams, P. (1988). A user’s guide to the General Health
Questionnaire.
Windsor, UK: NFER-Nelson.
Grubaugh, A. L., Long, M. E., Elhai, J. D., Frueh, B. C., & Magruder, K. M. (2010). An
examination of the construct validity of posttraumatic stress disorder with veterans using
a revised criterion set. Behavior, Research and Therapy. 48, 909–914. doi:
10.1016/j.brat.2010.05.019
Haynes, S. N., Smith, G. T., & Hunsley, J. D. (2011). Scientific foundations of clinical
assessment. New York: Routledge.
Hemphill, J. (2003). Interpreting the magnitudes of correlation coefficients.
American Psychologist, 58, 78-79. doi: 10.1037/0003-066X.58.1.78
Higgins, J. P. T., & Thompson, S. G. (2002). Quantifying heterogeneity in a meta-analysis.
Statistics in Medicine, 21, 1539-1558. doi: 10.1002/sim.1186
Hofstede, G., Hofstede, G. J., & Minkov, M. (2010). Cultures and organizations: Software of the
Mind, (3rd ed.). New York: McGraw-Hill.
Hollifield, M., Warner, T. D., Lian, N., Krakow, B., Jenkins, J. H., Kesler, J., Stevenson, J., &
Westermeyer, J. (2002). Measuring trauma and health status in refugees: A critical
122
review. Journal of the American Medical Association, 288, 611–621. doi:
10.1001/jama.288.5.611
Horowitz, M., Wilner, N., & Alvarez, W. (1979). Impact of Event Scale: A measure of
subjective stress. Psychosomatic Medicine, 41, 209-218. doi: 10.1097/00006842197905000-00004
Hunsley, J., & Mash, E. J. (2008). Developing criteria for evidence-based assessment: An
introduction to assessments that work. In J. Hunsley & E. J. Mash (Eds.), A guide to
assessments that work (pp. 3-14). New York: Oxford University Press.
Jones, E., Vermaas, R. H., McCartney, H., Beech, C., Palmer, I., Hyams, K., & Wessely, S.
(2003). Flashbacks and posttraumatic stress disorder: The genesis of a 20th-century
diagnosis. British Journal of Psychiatry, 182, 158–163. doi: 10.1192/bjp.182.2.158
Keane, T. M., Silberbogen, A. K., & Weierich, M. R. (2008). Posttraumatic stress disorder. In J.
Hunsley & E.J. Mash (Eds.), A guide to assessments that work (pp. 293-315). New York:
Oxford University Press.
Kelly, P. A., O’Malley, K. J., Kallen, M. A., & Ford, M. E. (2005). Integrating validity theory
with use of measurement instruments in clinical settings. Health Services Research, 40,
1605-1619. doi: 10.1111/j.1475-6773.2005.00445.x
Kessler, R. C., Sonnega, A., Bromet, E., Hughes, M., & Nelson, C. B. (1995). Posttraumatic
stress disorder in the national comorbidity survey. Archives of General Psychiatry, 52,
1048-1060. doi: 10.1002/1099-1298(200011/12)
Kessler, R. C., Sonnega, A., Bromet, E., Hughes, M., Nelson, C. B., & Breslau, N. (1999).
Epidemiological risk factors for trauma and PTSD. In R. Yehuda (Ed.), Risk factors for
posttraumatic stress disorder (pp. 23-60). Washington, DC: American Psychiatric Press.
123
Kimberlin, C. L., & Winterstein, A. G. (2008). Validity and reliability of measurement
instruments in research. American Journal of Health-System Pharmacists, 65, 2276-2284.
doi: 10.2146/ajhp070364
Kinicki, A. J., McKee-Ryan, F. M., Schriescheim, C. A., & Carson, K. P. (2002). Assessing the
construct validity of the Job Descriptive Index: A review and meta-analysis. Journal of
Applied Psychology, 87, 14-32. doi: 10.1037//0021-9010.87.1.14
Kirmayer, L. J., & Young, A. (1998). Culture and somatization: Clinical, epidemiological, and
ethnographic perspectives. Psychosomatic Medicine, 60(4), 420-430.
Kleijn, W. C., Hovens, J. E. & Rodenburg, J. J. (2001). Posttraumatic stress symptoms in
refugees: Assessments with the Harvard Trauma Questionnaire and the Hopkins
Symptom Checklist-25 in different languages. Psychological Reports, 88, 527-532. doi:
10.2466/PR0.88.2.527-532
Kornblith, A. B., Herndon, J. E., Weiss, R. B., Zhang, C., Zuckerman, E. L., Rosenburg,
S.,…Holland, J. C. (2003). Long-term adjustment of survivors of early-stage breast
carcinoma, 20 years after adjuvant chemotherapy. Cancer, 98, 679-689. doi:
10.1002/cncr.11531
Lange, A., Rietdijk, D., Hudcovicova, M., van de Ven, J. P., Schrieken, B., & Emmelkamp, P.
M. (2003). Interapy: A controlled randomized trial of the standardized treatment of
posttraumatic stress through the internet. Journal of Consulting and Clinical Psychology,
71, 901–909. doi: 10.1037/0022-006X.71.5.901
Lau, J., Ioannidis, J. P. A., Terrin, N., Schmid, C. H., & Olkin, I. (2006). The case of the
misleading funnel plot. British Medical Journal, 333, 597-600. doi:
10.1136/bmj.333.7568.597
124
Leiknes, K. A., Finset, A., Moum, T., & Sandanger, I. (2007). Course and predictors of
medically unexplained pain symptoms in the general population. Journal of
Psychosomatic Research, 62, 119-128. doi: 10.1016/j.jpsychores.2006.08.009
Maneesriwongul, W. & Dixon, J. K. (2004) Instrument translation process: A methods review.
Journal of Advanced Nursing, 48, 175– 186. doi: 10.1111/j.1365-2648.2004.03185.x
Matthieu, M., & Ivanoff, A. (2006). Treatment of human-caused trauma: Attrition in the adult
outcomes research. Journal of Interpersonal Violence, 2, 1654–1664. doi:
10.1177/0886260506294243
Messick, S. (1995). Validity of psychological assessment: Validation of inferences from persons’
response and performances as scientific inquiry into score meaning. American
Psychologist, 50, 741-749. doi: 10.1037/0003-066X.50.9.741
Michalopoulos, L. M., Unick, G. J., Haroz, E. E., Bass, J., Murray, L. K., & Bolton, P. A.
(2015). Exploring the fit of western PTSD models across three non-western low- and
middle-income countries. Traumatology, 21, 55-63. doi: 10.1037/trm0000020
Miller, M. W., Fogler, J. M., Wolf, E. J., Kaloupek, D. G., & Keane, T. M. (2008). The
internalizing and externalizing structure of psychiatric comorbidity in combat veterans.
Journal of Traumatic Stress, 21, 58–65. doi: 10.1002/jts.20303
Mollica, R. F., Caspi-Yavin, Y., Bollini, P., Truong, T., Tor, S., & Lavelle, J. (1992). The
Harvard Trauma Questionnaire: Validating a cross-cultural instrument for measuring
torture, trauma, and posttraumatic stress disorder in Indochinese refugees. Journal of
Nervous and Mental Disease, 180, 111-116. doi: 10.1097/00005053-199202000-00008
Mollica, R. F., Wyshak, G., de Marneffe, D., & Lavelle, J. (1987). Indochinese versions of the
Hopkins Symptom Checklist-25: A screening instrument for the psychiatric care of
125
refugees. American Journal of Psychiatry, 144, 497-500. doi: 10.1176/ajp.144.4.497
Nakeyar, C., & Frewen, P. A. (2016). Evidence-based care for Iraqi, Kurdish, and Syrian asylum
seekers and refugees of the Syrian civil war: A systematic review. Canadian Psychology,
57, 233-245. doi: 10.1037/cap0000067
North, C. S., Kawasaki, A., Spitznagel, E. L., & Hong, B. A. (2004). The course of PTSD, major
depression, substance abuse, and somatization after a natural disaster. Journal of Nervous
and Mental Disease, 192, 823-829. doi: 10.1097/01.nmd.0000146911.52616.22
Norris, F. H. (1992). Epidemiology of trauma: Frequency and impact of different potentially
traumatic events on different demographic groups. Journal of Consulting and Clinical
Psychology, 60, 409-418. doi: 10.1037/0022-006X.60.3.409
Ones, D. S., Viswesvaran, C., & Schmidt, F. L. (2003). Personality and absenteeism: A metaanalysis of integrity tests. European Journal of Personality, 17, S19-S38. doi:
10.1002/per.487
Orwin, R. G. (1983). A fail-safe N for effect sizes in meta-analysis. Journal of Educational
Statistics, 8, 157-159. doi: 10.3102/10769986008002157
Perkonigg, A., Kessler, R. C., Storz, S., & Wittchen, H.-U. (2000). Traumatic events and posttraumatic stress disorder in the community: Prevalence, risk factors and comorbidity.
Acta Psychiatrica Scandinavica, 101, 46-59. doi: 10.1034/j.16000447.2000.101001046.x
Pietrzak, R. H., Goldstein, R. B., Southwick, S. M., & Grant, B. F. (2011). Medical comorbidity
of full and partial posttraumatic stress disorder in US adults: Results from Wave 2 of the
National Epidemiologic Survey on Alcohol and Related Conditions. Psychosomatic
Medicine, 73, 697-707. doi: 10.1097/PSY.0b013e3182303775
126
Porter, M., & Haslam, N. (2005). Pre-displacement and post-displacement factors associated
with mental health of refugees and internally displaced persons: A meta-analysis.
Journal of the American Medical Association, 294, 602-612. doi:
10.1001/jama.294.5.602.
Rasmussen, A., Verkuilen, J., Ho, E., & Fan, Y. (2015). Posttraumatic stress disorder among
refugees: Measurement invariance of Harvard Trauma Questionnaire scores across global
regions and response patterns. Psychological Assessment, 27, 1160-1170. doi:
10.1037/pas0000115
Ruggiero, K. J., Del Ben, K., Scotti, J. R., & Rabalais, A. E. (2003). Psychometric properties of
the PTSD Checklist-Civilian version. Journal of Traumatic Stress, 16, 495-502. doi:
10.1023/A:1025714729117
Schmidt, F. L., & Hunter, J. E. (1977). Development of a general solution to the problem of
validity generalization. Journal of Applied Psychology, 62, 529-540. doi: 10.1037/00219010.62.5.529
Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in
personnel psychology: Practical and theoretical implications of 85 years of research
findings. Psychological Bulletin, 124, 262-274. doi: 10.1.1.172.1733
Shoeb, M., Weinstein, H., & Mollica, R. (2007). The Harvard Trauma Questionnaire:
Adapting a cross-cultural instrument for measuring torture, trauma and posttraumatic
stress disorder in Iraqi refugees. The International Journal of Social Psychiatry, 53,
447-463. doi: 10.1177/0020764007078362
Sperber, A. D. (2004) Translation and validation of study instruments for cross-cultural research.
Gastroenterology, 126, S124–S128. doi: 10.1053/j.gastro.2003.10.016
127
Spitzer, R. L., First, M. B., & Wakefield, J. C. (2007). Saving PTSD from itself in DSM-V.
Journal of Anxiety Disorders, 21, 233-241. doi: 10.1016/j.janxdis.2006.09.006
Steel, Z., Chey, T., Silove, D., Marmane, C., Bryant, R. A., & van Ommeren, M. (2009).
Association of torture and other potentially traumatic events with mental health outcomes
among populations exposed to mass conflict and displacement: A systematic review and
meta-analysis. Journal of the American Medical Association, 302, 537-549. doi:
10.1001/jama.2009.1132
Van der Velden, P. G., Grievink, L., Kleber, R. J., Drogendijk, A. N., Roskam, A. J.,
Marcelissen, F. G., … Gersons, B. P. (2006). Post-disaster mental health problems and
the utilization of mental health services: A four-year longitudinal comparative study.
Administration and Policy in Mental Health, 33, 279–288. doi: 10.1007/s10488-0050027-x
Weathers, F. W., Litz, B. T., Herman, D. S., Huska, J. A., & Keane, T. M. (1993, October). The
PTSD checklist (PCL): Reliability, validity, and diagnostic utility. Poster presented at the
9th annual meeting of the International Society for Traumatic Stress Studies, San
Antonio, TX.
Westen, D., & Rosenthal, R. (2003). Quantifying construct validity: Two simple measures.
Journal of Personality and Social Psychology, 84, 608-618. doi: 10.1037/00223514.84.3.608
Wood, J. (2008). Methodology for dealing with duplicate study effects in meta-analysis.
Organizational Research Methods, 11, 79-95. doi: 10.1177/1094428106296638
128
Figure 1
Literature Search Flow Chart
1571 entries identified
724 from P.I.L.O.T.S.
596from Web of Science
126 from PsycINFO
125 from Medline
Studies excluded from this review
709 were duplicates
196 were not empirical studies
171 did not use HTQ-symptom
scales
58 were not accessible
29 were not available in book, article
or dissertation format
22 were not in French or English
2 did not have participants aged > 6
384 studies used the HTQsymptom scales
Studies excluded from this review
268 studies did not provide
convergent or discriminant validity
coefficients
116 studies had sample validity
estimates
78 studies (79 samples) were
included in the VG
Studies excluded from this review
38 studies included duplicate
samples, provided estimates that did
not fit into the correlate categories
and/or modified the HTQ (i.e. added
or subtracted items)
129
Table 1
Descriptive Statistics for the Convergent Validity of the HTQ Symptom Scales
95% confidence
interval
Validity estimate
k
Mean Lower
Upper Min. Max.
Q
r
Cumulative
trauma
HTQ-16 43
.32
.26
.38
-.28
.90
546.53*
I2
Orwin’s
N
92.32
17
HTQ-30
12
.26
.17
.34
.04
.53
48.16*
77.16
1
Posttraumatic
stress measures
HTQ-16
11
.66
.52
.76
.24
.86
154.92*
93.55
41
Cultural
expressions of
trauma
HTQ-16
3
.75
.59
.86
.55
.83
29.38*
93.19
15
Dissociation
Measures
HTQ-16
3
.53
.32
.70
.18
.69
33.46*
94.03
8
Note. Min. = Minimum correlation reported in the literature; Max. = Maximum correlation reported in the literature; k =
number of samples included in the analyses; N/A = Insufficient number of samples to compute Orwin’s N
*p < .001, **p < .01
130
Table 2
Descriptive Statistics for the Discriminant Validity of the HTQ Symptom Scales
95% confidence
interval
Validity
k
Mean
Lower Upper Min. Max.
Q
estimate
r
Depressive and
anxiety
symptoms
measures
HTQ-16 3
.80
.77
.83
.79
.86
1.73
I2
Orwin’s
N
0
15
HTQ-30
2
.79
.61
.89
.71
.85
12.84*
92.21
N/A
Anxiety
symptoms
measures
HTQ-16
12
.62
.56
.68
.44
.78
61.47*
82.10
31
HTQ-30
2
.66
.16
.89
.43
.80
18.49*
94.60
N/A
Depression
symptoms
measures
HTQ-16
21
.64
.57
.69
.16
.84
199.78*
90.00
58
HTQ-30
5
.78
.72
.82
.64
.82
9.02
55.67
23
Somatization
measures
HTQ-16
6
.43
.28
.56
.17
.62
47.11*
89.37
11
HTQ-30
2
.45
-.31
.86
.08
.72
37.24*
97.32
N/A
Note. Min. = Minimum correlation reported in the literature; Max. = Maximum correlation reported in the literature; k
= number of samples included in the analyses; N/A = Insufficient number of samples to compute Orwin’s N
*p < .001
131
Table 3
Summary of Moderating Effects for the Construct Validity of the HTQ Symptom Scales
Cumulative trauma
PTSD
measures
Anxiety
symptoms
measures
Depressive
symptoms
measures
Somatization
measures
Moderator
HTQ-16
N = 43
HTQ-30
N = 12
HTQ-16
N = 11
HTQ-16
N = 12
HTQ-16
N = 21
HTQ-16
N=6
Age
No (34*)
No (10)
No (8)
No (10)
No (18)
Yes (5)
Gender
No (40)
No (12)
No (9)
No (12)
No (20)
Yes (6)
Cultural
group
No (34)
Yes (9)
No (10)
No (11)
No (19)
Yes (5)
Cultural
orientation
No (19)
No (5)
No (5)
No (4)
No (9)
N/A
Forcibly
displaced
sample
No (43)
No (12)
No (11)
No (11)
No (20)
No (5)
Recruitment
site
No (41)
Yes (12)
Yes (10)
No (11)
No (20)
No (6)
Trauma type
No (39)
Yes (12)
Yes (11)
N/A
No (19)
N/A
Original
sample
No (43)
Yes (12)
Yes (12)
No (12)
No (21)
N/A
Note. *Total samples included in the moderator analyses are in parentheses; Yes = Significant moderating effect;
No = No significant moderating effect; N/A = Insufficient information available to conduct moderator analyses
132
General Discussion
The Harvard Trauma Questionnaire (HTQ; Mollica et al., 1992) is a measure that was
intended to assess traumatic events and related symptoms of trauma across cultures. The HTQ
has been found to be commonly used in the literature (Gagnon et al., 2004; Hollifield et al.,
2002), and has been shown to be useful to exclude PTSD non-cases in certain forcibly displaced
populations (Nakeyar & Frewen, 2016). This instrument also has a reputation of being a
culturally sensitive measure of PTSD symptoms (Keane et al., 2008). However, to date, there has
been no thorough review that supports this reputation and evaluates the overall psychometric
strength of the HTQ symptom scales as measures of PTSD across cultures. This dissertation was
designed to fill this gap by: (a) gathering descriptive information regarding the use of the HTQ
symptom scales in the empirical literature, (b) evaluating the reliability properties of this
measure by calculating mean internal consistency estimates, (c) evaluating whether the HTQ
symptom scales have produced valid scores across studies found in the literature by calculating
mean construct validity estimates, and (d) examining sample and methodological characteristics
that may influence the reliability and construct validity of scores produced by these scales. The
following sections include a summary of both meta-analytic studies included in this dissertation,
as well as the implications and recommendations stemming from their findings while taking into
account potential limitations.
The Psychometric Properties of the HTQ
The literature search of studies that used the HTQ symptom scales in Study 1 shows no
signs of a decline in the frequency of use of the scales. Researchers across the globe have used
the HTQ to assess posttraumatic stress symptoms in refugee and nonrefugees, clinical and
nonclinical populations, and across a wide range of cultural and ethnic backgrounds. Researchers
133
have also made attempts at adapting the HTQ for their specific sample. However, in the majority
of studies, the reported adaptation procedures they used did not follow the translation and
cultural adaptation guidelines provided by the original developers, nor did they follow
commonly known adaptation recommendations (e.g., International Test Commission, 2005).
Although the current study could not evaluate the efficacy of following these adaptation
procedures, several studies offer support for the use of carefully implemented adaptation
strategies in cross-cultural assessment to help minimize bias (e.g. Dolnicar & Grün, 2007,
Harzing, 2006). Furthermore, those who used the HTQ-30 used the Indochinese version of the
scale. This scale was based on the cultural expressions of trauma among Cambodian, Lao and
Vietnamese refugees that may not be appropriate for individuals of other cultures. Although
some researchers had translated the cultural items for use in their studies (e.g., Vloeberghs, van
der Kwaak, Knipscheer, & ven den Muijsenbergh, 2012), the sense or meaning of the cultural
items remained the same, even when participants did not have Cambodian, Lao, or Vietnamese
backgrounds. However, because there was no opportunity to evaluate the 14 cultural items of the
HTQ-30 due to insufficient data, it is difficult to determine if the inclusion of these items
enhances the use of the HTQ across cultures.
To assess the reliability properties of the HTQ symptom scales, I conducted reliability
generalization analyses on Cronbach’s alpha coefficients in Study 1. These coefficients were
retrieved from the published literature and unpublished dissertations, or were provided by
authors after requesting this information. There was sufficient data to conduct RG analyses for
the HTQ-16, HTQ-30, and the re-experiencing, avoidance and arousal subscales. Publication
bias analyses revealed that all RG results were likely to be unaffected by this form of bias.
Taking into account the results of the Spearman-Brown formula, the HTQ-16 and HTQ-30 had
134
comparable mean coefficients that would be described as “excellent” (i.e., ≥ .90) according to
Hunsley and Mash’s (2008) guidelines. Although both had excellent ratings, the addition of the
cultural expressions of trauma items did not add value to the HTQ-16 in terms of increasing
internal consistency. Regarding the subscales of the HTQ-16, the Spearman-Brown formula was
used to compare the subscales with each other. The re-experiencing and arousal subscales had a
rating of “good” (i.e., mean values in the.80-.89 range) and the avoidance/numbing subscale a
rating of “adequate” (i.e., mean values in the .70-.79 range). In summary, the RG analyses
revealed that the reliability properties of the HTQ-30, HTQ-16, and the re-experiencing and
arousal subscales are strong. However, the reliability properties of the avoidance/numbing
subscale are weaker. The lower level of reliability for this subscale may be reflective of an
inadequate set of items chosen to represent avoidance/numbing symptoms, or even the poor
factorial validity of the DSM-IV 3-factor model of PTSD (e.g. Yufik & Simms, 2010). Another
potential cause for the suboptimal performance of the numbing subscale may be the inconsistent
presence of avoidance and numbing symptoms across cultures (Marsella, Friedman, Gerrity, &
Scurfield, 1996; Norris, Van Landingham, & Vu, 2009). This issue was highlighted throughout
the revision process for the new DSM-5 criteria for PTSD (Hinton & Lewis-Fernández, 2010),
during which it was decided to separate the avoidance and numbing symptoms into separate
clusters. To date, there is no updated HTQ available that reflects the changes made to the PTSD
construct in DSM-5. If these changes were made, perhaps the reliability of the subscales would
improve. Therefore in terms of reliability properties, the two HTQ symptom scales as a whole
appear to be strong, however some caution should be used when interpreting the subscales.
To assess the construct validity of the HTQ symptom scales in Study 2, I conducted
validity generalization analyses on convergent and discriminant validity coefficients retrieved
135
from the published literature and unpublished dissertations. The coefficients were grouped
according to broader correlate categories. The convergent validity correlate categories were: (a)
cumulative trauma; (b) other posttraumatic stress measures; (c) cultural expressions of trauma;
and (d) dissociation measures. The discriminant validity categories were: (a) depression and
anxiety symptoms measures; (b) anxiety symptoms measures; (c) depression symptoms
measures; and (d) somatization measures. Publication bias analyses revealed that the overall VG
findings should be only considered as preliminary until additional data are available.
Based on Hemphill’s (2003) r-value effect-size benchmarks, there was strong evidence
for the convergent validity of the HTQ-16, as all correlates had large effect sizes (i.e., > .30). The
strongest association was with cultural expressions of trauma and the weakest with cumulative
trauma. The mean correlation between the HTQ-30 and cumulative trauma was considered a
medium effect size (i.e., r = .26). There were insufficient data to compute VG estimates for the
other convergent correlate categories for the HTQ-30. Similar to other PTSD measures (e.g.,
Dermichyan, Goenjian, & Khachadourian, 2015; Ruggiero, Del Ben, Scotti, & Rabalais, 2003),
the cumulative trauma category demonstrated the lowest convergent validity for the HTQ-16 and
only had a medium effect for the HTQ-30. Because of variability in the conditional risk of PTSD
across cultures, traumatic events may not be an ideal convergent validity correlate. In addition, it
appears that trauma subtypes can predict mental health outcomes better than cumulative trauma
(Arnetz et al., 2014), which could be a reason for the weaker association with PTSD. Because of
these reasons, future validation studies of PTSD measures should refrain from using this variable
as a convergent validity correlate.
The results of Study 2 call into question the overall discriminant validity properties of the
HTQ symptom scales. Based on Hemphill’s (2003) guidelines, all discriminant correlates of both
136
scales had large effect sizes. Similar to other PTSD measures (e.g., Bovin et al., 2016;
Dermichyan et al., 2015; Lauterbach, Vrana, King, & King, 1997), the associations between the
HTQ symptom scales and depression and anxiety symptom measures were rather strong (i.e., >
.60). Although large correlations were expected due to the high rates of comorbidity and
symptom overlap with depression and anxiety, PTSD measures should have significantly higher
correlations with similar constructs than with theoretically less related constructs. This was not
the case for the HTQ-16 where some mean discriminant validity coefficients were comparable in
size to convergent validity coefficients. Furthermore, although somatization measures had lower
mean correlations with both HTQ symptom scales, they were still relatively high. This strong
association is consistent with the results of other studies showing high associations between
PTSD and somatization (e.g., Beckham et al., 1998; North, Kawasaki, Spitznagel, & Hong,
2004). The present results, combined with the findings of other studies that have evaluated the
validity of the HTQ symptom scales (e.g., Rasmussen et al., 2015), indicate that the construct
validity of both scales appears to be “adequate,” based on Hunsley and Mash’s (2008) criteria for
psychometric ratings. In other words, there is some independently replicated evidence of
construct validity for the scales.
In summary, the findings of the RG and VG analyses indicate that the HTQ-16 is likely
to produce reliable scores across various types of samples. This scale also has some evidence of
convergent validity, but performs rather poorly with regards to discriminant validity. In terms of
the subscales of the HTQ-16, the reliability properties of the re-experiencing and arousal
subscales were good, but the avoidance/numbing subscale was only considered adequate. With
regard to the HTQ-30, it is important to note that the majority of studies that used this scale did
not report culturally adapting the HTQ for their specific sample. These studies thus applied the
137
Indochinese version of the questionnaire to non-Indochinese groups. Although the cultural items
that are based on clinical findings among Indochinese populations provided reliable scores across
various populations, the ability to discriminate from measures of anxiety and depression
symptoms was weak. There was insufficient information to evaluate the overall convergent
validity, which prevented from fully assessing the construct validity of this scale.
The Role of Moderators
The majority of mean reliability and validity coefficients calculated from the RG and VG
analyses had statistically significant Q and high I2 tests. These results indicated that these
estimates had heterogeneous variability, which was not due to sampling error alone, and that the
examination of sample and methodological characteristics could perhaps explain this variability.
Meta-regression and subgroup analyses were thus performed on reliability and validity
coefficients that had a statistically significant Q-statistic and sufficient amount of available data.
There were a total of eight variables that were selected as potential moderators, and seven of
these variables played a moderating role with the reliability and/or validity estimates. These
included age, gender, cultural group, cultural orientation, recruitment site, trauma type and
whether the study sample was a sample for which the HTQ was originally validated.
Having the experience of being a forcibly displaced person did not have a significant
moderating effect, suggesting that the HTQ symptom scales and subscales should perform rather
equally in terms of psychometric properties in both forcibly displaced samples and not forcibly
displaced samples. This is somewhat surprising given that the HTQ symptom scales were
originally validated for forcibly displaced samples (i.e., refugees). However, there may be
underlying factors that account for the lack of a moderating effect. For instance, studies have
found that several elements may a role in the association between the experience of being
138
forcibly displaced and psychological distress, such as pre-migratory potentially traumatic events
and post-migration stressors (e.g. Laban, Gernaat, Komproe, Van, & de Jong, 2005; SchwarzNielsen & Elklit, 2009), as well as the length of stay in the host country/area (Guajardo, SlewaYounan, Smith, Eagar, & Stone, 2016). Because the forcibly displaced samples included in the
current analyses varied greatly with regards to these factors, the heterogeneous characteristic of
these samples may explain the lack of a moderating effect of this variable.
In terms of significant moderating effects, the mean age of samples played a role in the
discriminant validity of the HTQ-16, specifically with the associations between this scale and
somatization measures. Because higher associations were found with lower mean ages of study
samples, it appears that the discriminant validity of the HTQ-16 may be compromised when it is
compared to somatization measures in younger age groups. This moderating effect appears to
support other findings that have shown significant associations between age and PTSD (e.g.,
Kessler et al., 1995; Norris 1992), as well as age and somatization (e.g., Leiknes, Finset, Moum,
& Sandanger, 2007).
Gender also played a moderating role with some reliability and discriminant validity
estimates. Specifically, study samples comprised of mainly female participants had higher
reliability estimates on the re-experiencing subscale than did samples comprised mainly of male
participants, as well as with the associations between the HTQ-16 and somatization measures.
This is seemingly consistent with the findings of other studies that have shown gender
differences in somatization (e.g., Leiknes et al., 2007) and PTSD (e.g., Ditlevsen & Elklit, 2010;
Pietrzak et al., 2011), including re-experiencing symptoms specifically (Vassar, Knaup, Hale, &
Hale, 2011). These results suggest that the discriminant validity of the HTQ-16 and the reliability
of the re-experiencing subscale may be weaker in male samples.
139
Although it was expected that the cultural group would have a significant moderating
effect on all of the dimensions of reliability and construct validity examined, this variable only
had an impact on some aspects. Specifically, the cultural group had a predictive effect on the
reliability of scores on the re-experiencing subscale, the convergent validity of the HTQ-30 (i.e.,
association with cumulative trauma) and the discriminant validity of the HTQ-16 (i.e.,
association with somatization measures). These findings support the existing evidence of cultural
and ethnic variations in PTSD symptomatology (e.g., Kessler et al., 1999), exposure to traumatic
events (e.g., Roberts, Gilman, Breslau, Breslau, & Koenen, 2011) and somatization (e.g.,
Kirmayer & Young, 1998).
As an extension to this, the cultural background of the samples’ country in terms of the
construct of collectivism versus individualism played a moderating role in the reliability of
scores of the HTQ. Specifically, the reliability coefficient of the avoidance/numbing subscale for
samples from individualist countries was significantly higher than the typical estimate for
samples from collectivist countries. This may be reflective of lower rates of avoidance/numbing
symptoms reported in collectivist cultures (e.g., McCall & Resick, 2003). These results also
support Jayawickreme, Jayawickreme, and Foa’s (2012) hypothesis that individuals from
collectivist cultures may have less avoidant behaviours because of social roles and obligations
that require them to consistently engage with others and their environment. This variable
however, did not have a significant impact on the validity coefficients. Given the small amount
of empirical evidence to date on this matter, it is clear that further research is needed to makes
inferences regarding the role of cultural orientation on posttraumatic symptoms.
Similar to the cultural group variable, the type of trauma experienced (i.e., intentional
versus unintentional) played a role in both the reliability and construct validity of the HTQ. The
140
reliability of scores produce by the re-experiencing subscale was stronger for “intentional
traumas.” Furthermore, the experience of “intentional traumas” was also associated with stronger
convergent validity properties of the HTQ-30. As intentional traumas are considered to be more
severe than non-intentional traumas, these results may indicate that the internal consistency of reexperiencing symptoms and the convergent validity of the HTQ-30 are associated with trauma
severity. On the other hand, conflicting results were obtained for the HTQ-16 that had stronger
convergent validity properties with “unintentional traumas.” These results indicate that further
research is needed regarding the role of trauma types and trauma severity to further test these
interpretations.
The sites from which the samples were recruited had a significant effect on the construct
validity of the HTQ symptom scales. The convergent validity of the HTQ-30 when associated
with cumulative trauma was stronger for samples recruited in community settings. In addition,
the convergent validity of the HTQ-16 when associated with other posttraumatic stress measures
was stronger for samples recruited in medical treatment settings. Although the underlying factors
contributing to these differences are unclear, these findings provide further support regarding the
questionable generalizability of the scales due to variability in the sensitivity across recruitment
settings (e.g., Mollica et al. 1992; Silove et al., 2007). Further studies are needed to clarify these
findings.
Finally, being a sample for which the HTQ was originally validated had a positive effect
on the association between cumulative trauma and the HTQ-30. These results provide some
(although rather limited) support for the convergent validity of this scale in Indochinese refugee
populations.
141
In summary, no moderators played a role in the overall reliability of scores of the HTQ16 and HTQ-30. This finding suggests that these scales perform relatively consistently across
various sample type and methodological procedures. The scales did differ, however, in terms of
moderators for the construct validity. Because of the high risk of publication bias and the low
power of these moderating effects due to relatively small numbers of studies included in the
analyses, these results do not indicate that there are significant limitations with either HTQ
symptom scale and should only be considered as preliminary findings.
Limitations of the Studies
There are several limitations to these studies. As highlighted in both studies, the cited
reference search strategy used to identify published studies may not have identified all published
studies and unpublished dissertations that have used the HTQ symptom scales. In addition, this
strategy could not identify studies, other than dissertations, that have used this measure but are
unpublished. This is important to bear in mind given that the publication bias analyses for the
VG indicated that few studies would be needed to significantly lower the mean coefficients
reported in the analyses. Thus, the findings of the VG in particular should be considered as
preliminary until considerably more construct validity data is available on the HTQ scales.
For the studies identified with the search strategy, data from many were not included in
the RG and VG analyses because reliability and validity estimates were either not reported in the
primary studies or, as in the case for the VG, authors reported conducting bivariate analyses but
did not provide sufficient information to include these findings in the meta-analyses. Although
some authors provided a coefficient alpha after being contacted, there were still many missing
reliability estimates that could have been included in the RG.
142
Furthermore, the coding process used to extract information for the analyses was
challenging at times given that the information reported in the studies was often vague,
challenging to interpret, and difficult to locate. A large amount of the information coded was
retrieved in the methods sections where the adaptation and administration procedures of the HTQ
would normally be described. However, there was often a lack of sufficient detail reported on
these procedures. It is therefore possible that the information reported in these studies did not
fully represent the actual methodological procedures, including cross-cultural adaptations that
were used in the research. Indeed, the lack of details in the reporting of adaptation methods
appears to be a common theme in the published literature (Maneesriwongul & Dixon, 2004).
These problems continue despite reporting standards for research that been put in place, in part,
to ensure the quality of research synthesis and meta-analysis (Appelbaum, Cooper, Maxwell,
Stone, & Sher, 2008).
Another possible limitation lies in the merging of estimates of potential moderators and
VG correlates into larger categories. Although this procedure helped increased statistical power,
there may have been an important loss of information, and perhaps less accurate results. For
instance, combining individual cultural groups into a broader category (e.g., combining
Cambodian and Tibetan samples into the Asian category) considers these groups as homogenous
when in fact there can be important differences across subgroups. This procedure thus reduces
the generalizability of the findings. Moreover, the merging of VG estimates into broad correlate
categories involved combining estimates provided by several different instruments that may not
have been measuring the same underlying construct. For instance, there were several
instruments, such as the Indochinese versions of the Hopkins Symptom Checklist-25 (Mollica et
al., 1987) and the Beck Depression Inventory-II (Beck, Steer, & Brown, 1996) that were
143
included in the broad category of “depression symptoms measures.” One way of ensuring that
these instruments were indeed measuring the same construct was to evaluate the psychometric
properties of each instrument. However, because of the numerous measures included, the
psychometric evaluation of each of them was not feasible.
Furthermore, only 20% of the studies included in each of the RG and VG analyses were
coded by a second coder. Although the interrater reliability was rated as “good” for both studies,
a higher percentage of studies coded by the second coder may have increased the accuracy of the
overall RG and VG results.
Finally, the decision to include samples consisting of children and adolescents by using a
lower age limit of 7 years may not have been the most judicious. The variations in the symptom
expression of PTSD due to developmental factors, especially in school-aged children (e.g. Cohen
& the Work Group on Quality Issues, 1998), could have impacted the accuracy of the results.
However, there was only one sample in the RG and two samples in the VG that included children
as young as 9 years. Although approximately 23% of the RG samples and 20% of the VG
samples included adolescents, PTSD symptoms during this developmental stage are more likely
to be similar to symptoms experienced in adulthood (Cohen & the Work Group on Quality
Issues, 1998). Because there were only a few studies that included younger children, the possible
negative impact on the overall meta-analytic results is likely to be minimal.
Implications
In addition to the information regarding the psychometric properties of the HTQ
symptom scales, several other important implications stem from the current studies. For instance,
the findings of both studies add fuel to the debate regarding etic versus emic approaches for the
assessment of PTSD. The findings revealed that an etic approach to assessment across cultures
144
can be justified given that the HTQ-16 produced reliable scores and had sound convergent
validity properties across diverse samples. In addition, the variability in the reliability of scores
among the subscales (i.e., re-experiencing, avoidance/numbing, arousal symptoms) indicates that
the etic approach to the assessment of PTSD can help capture the cultural nuances in the
expression of symptom sets when the same measure is used cross-culturally. On the other hand,
the practice of incorporating an emic perspective to the etic approach is also supported, as the
HTQ-30 had excellent reliability properties as well as some evidence of convergent validity
(especially in samples similar to those on which the HTQ was originally validated). Although
there were no data available to compare strictly emic approaches to assessment, the findings of
the current studies suggest that there is no right or wrong answer to this debate, and that each
approach can be justified to be used for specific sets of purposes. For instance, the etic approach
may be more useful to evaluate cross-cultural differences in research settings, whereas an emic
approach would be more appropriate to help better understand the expressions of trauma in a
certain cultural group. A combination of emic and etic approaches to the assessment of PTSD
would be particularly useful in clinical settings. The etic approach could help guide clinicians in
establishing a diagnosis comparing symptom patterns to normative data, whereas the information
provided by emic approaches could help tailor treatment interventions according to the specific
needs of a client.
Furthermore, the results obtained from the VG analyses focused on discriminant validity
raise doubts about the overall construct validity of HTQ symptom scales and the construct
validity of PTSD as conceptualized in the DSM-IV (APA, 1994). The main concern was the high
degree of similarity between the convergent and discriminant validity estimates across a wide
range of disparate variables. Although some studies support the hypothesis that removing
145
symptoms of other constructs theorized to overlap with PTSD, such as mood and other anxiety
disorders, increases the validity of the PTSD construct (e.g., Elhai et al., 2015; Spitzer, First, &
Wakefield, 2007), others have found that removing these symptoms does not reduce comorbidity
with these other disorders (e.g., Grubaugh, Long, Elhai, Frueh, & Magruder, 2010). In attempts
to find a solution to this debate, some researchers have gone further and suggested that taking a
dimensional approach to PTSD assessment is more appropriate than conceptualizing the disorder
into a category (e.g., Antony & Rowa, 2005). For instance, cluster and factor analytic studies
have found interesting interactions between PTSD and the internalizing and externalizing
dimensions to psychopathology (Cox, Clara, & Enns, 2002; Miller, Fogler, Wolf, Kaloupek, &
Keane, 2008). Clearly, more research is needed to resolve this key validity issue for both the
construct of PTSD and measures of PTSD across diverse populations.
Although the HTQ symptom scales were based on DSM-IV PTSD criteria, the findings
of the current studies also have some implications for the PTSD construct as conceptualized in
DSM-5. This most recent version includes an even broader set of symptoms with an addition of
three symptoms that were not included in DSM-IV and an increased emphasis on depressive
mood and cognition (Friedman, 2013). Because the current findings question the DSM-IV PTSD
construct due to a lack of ability to discriminate PTSD symptoms from symptoms of depression
and other anxiety disorders, it is possible that an increase in dysphoria symptoms in the DSM-5
would render the new conceptualization of PTSD even more questionable. Recent studies that
have used DSM-5 based measures suggest that the dysphoria symptoms of PTSD may be the
underlying reason for the observed high level of comorbidity with major depressive disorder
(Contractor et al., 2014; Elhai et al., 2015). However, these studies were conducted with North
American samples and it is therefore not appropriate to generalize these findings to other
146
cultures. On the other hand, one study comparing the DSM-IV and DSM-5 conceptualization of
PTSD found no significant differences in comorbidity rates with major depressive episode
(O’Donnell et al., 2014) revealing no great impact of the addition of mood related items to PTSD
in the DSM-5. This study would also need to be replicated across diverse cultural groups to
evaluate the generalizability of these findings. Also, because of the limitations of Study 2 (i.e.,
the potential for publication bias, low power), these interpretations should be considered
tentative. Nevertheless, the findings of the current studies further justify the need for a deeper
investigation into the cross-cultural applicability of the PTSD construct and of PTSD measures.
Recommendations for Researchers
Several important recommendations for researchers stem from this study. First, consistent
with established publication standards, authors are encouraged to report relevant and sufficient
information in the methods and results sections of their research publications. The inconsistent
and vague reporting of adaptation procedures, insufficient information to compute validity
coefficients, and low rate in the reporting of reliability estimates are indications that more
awareness is needed with regards to reporting guidelines for assessment instruments and
assessment research. As the accuracy of meta-analyses is dependent on the primary studies on
which they are based, missing information and data in primary studies ultimately has a negative
impact on the quality of meta-analyses. Researchers are thus encouraged to follow established
reporting guidelines such as the journal article reporting standards (JARS) when writing
manuscripts. Cooper (2011) has provided a practical guide on understanding and implementing
these standards. Although researchers may be aware of these standards, editorial policies can
prevent them from following these standards by limiting the amount of information allowed in
manuscripts. If there are space limitations, editors are encouraged to follow the example of some
147
journals, such as the journals of the American Psychological Association, that provide external
links to websites where supplementary information can be viewed (American Psychological
Association Publications and Communications Board Working Group on Journal Article
Reporting Standards, 2008). In light of the challenges in the data retrieval of the current metaanalyses, researchers are particularly encouraged to provide the reliability estimates of their
sample, detailed methodological procedures (including instrument adaptation procedures, if
applicable), and correlation matrices. The reporting of this information is important not only for
meta-analytic purposes, but also for helping readers to evaluate the research and to decide
whether a measure is appropriate for their specific use.
The current studies also highlighted significant problems in the utilization practices of the
HTQ symptom scales among researchers. Users of the HTQ (and any other assessment
instrument, for that matter) must be vigilant of the potential sources of bias, especially in
samples of different cultural backgrounds for which the original version of the measure was
intended to be used, and carefully address bias by ensuring measurement equivalency. This
includes evaluating whether an instrument can be applicable to a specific culture and if
necessary, using appropriate translation and cultural adaptation strategies Although there is a
lack of consensus regarding the specific adaptation procedures to be used (Epstein et al., 2015),
all experts in cross-cultural measurement agree that the adaptation process goes beyond a mere
translation. Common recommendations include using a consensus approach to translation, using
focus groups and validating the adaptation (Epstein et al., 2015). Cross-cultural validation
strategies can include multi-group confirmatory analyses and differential item functioning
(Milfont & Fischer, 2010). Although the factorial validity of the HTQ was not evaluated in the
current studies, the variable reliability properties of the HTQ-16 subscales, further support the
148
need for factor analyses in different cultural groups. Future users of the HTQ in research settings
are also encouraged to evaluate the ability of the scale to diagnose PTSD by conducting receiver
operating characteristic (ROC) curve analyses to establish appropriate cut-off scores for their
specific sample. The high cost and lengthy process of adapting and validating instruments can be
major obstacles in conducting research, but because the assessment of traumatized individuals
can have major clinical and societal implications (e.g., access to treatment, obtaining disability
compensation, obtaining refugee status), these procedures should not be ignored.
Recommendations for Clinicians
The aggregate nature of the findings in the current meta-analytic studies makes it
challenging to provide precise clinical recommendations regarding the use of the HTQ symptom
scales with specific cultural or linguistic groups. However, the moderating effects of cultural
variables do provide some insight on this matter and can help clinicians as they weigh the
strengths and weaknesses of this measure. Clinicians can also refer to the primary studies
included in the meta-analyses for more specific psychometric information relevant to the cultural
or linguistic group of interest. Clinicians are also encouraged to weigh the costs and benefits of
using the HTQ as compared to other PTSD measures that have been validated for use with
individuals who have similar characteristics to the client in question. For instance, if another
PTSD measure has more sound psychometric evidence for a specific cultural/linguistic group
than the HTQ, the other measure would be more appropriate to use.
Furthermore, because of the overlapping symptoms and high rates of comorbidity with
anxiety and mood disorders, it is recommended that clinicians include the assessment of anxiety
and mood-related symptoms in their differential diagnosis of PTSD. As the results from Study 2
also highlighted an important association between the HTQ symptom scales and somatization
149
measure, a further investigation of somatization symptoms may also be appropriate. Several
other factors need to be considered in the assessment of PTSD in diverse cultures, including the
evaluation of local idioms of distress, other culture-bound syndromes and culture-specific
response styles.
In terms of the specific uses of the HTQ-symptom scales, in addition to diagnostic
purposes, these scales may be particularly useful for the monitoring of symptoms, as they are
relatively short and easy to administer. However, future studies are needed to evaluate these
scales’ sensitivity to change, by administering the scales at different time points and compare the
scores obtained with another PTSD instrument shown to be sensitive to change, such as the
CAPS (Keane et al., 2008). Although there was insufficient information available to evaluate the
revised version of the HTQ (HTQ-R) in the current studies, the addition of items that assess daily
functioning in this newer version may provide relevant information when monitoring the
progress of treatment.
Conclusion
Common knowledge among psychologists dictates that good assessment practices
include ensuring that an assessment instrument has good psychometric properties. This is true,
however evidence-based assessment (EBA) goes far beyond this. In addition to using
psychometrically sound measures the overall assessment process entails a complex decisionmaking strategy that involves a constant back and forth between the formulation and testing of
hypotheses, which is often based on incomplete or contradicting information (Hunsley & Mash,
2007). Strategies to help clinicians and researchers make informed and practical decisions that
are grounded in EBA practices are available in the literature (e.g., Christon, McLeod, & Jensen-
150
Doss, 2015). However useful these strategies are, EBA is in continuous development and is not
without its struggles.
The current studies highlighted challenges in the development of EBA described by
Hunsley and Mash (2007). These difficulties included: (a) defining psychometric adequacy; (b)
addressing comorbidity; and (c) addressing diversity. The challenges in defining psychometric
adequacy were particularly evident in Study 2 where the absence of benchmarks for validity
coefficients made it difficult to fully evaluate the construct validity of the HTQ symptom scales.
Further research is thus needed to establish more specific and quantitative guidelines related to
the evaluation of convergent and discriminant validity. The difficulties in addressing comorbidity
were also highlighted in Study 2 where the ability of the HTQ symptom scales to discriminate
PTSD from other constructs, such as mood and anxiety disorders, was compromised. This issue
has been an ongoing debate in the literature where some researchers argue that a dimensional or
transdiagnostic approach to PTSD diagnosis may be more appropriate than conceptualizing the
disorder as a discrete category (e.g., Elhai et al., 2015). Because there is evidence for both
arguments, further data are needed to address the issue. Finally, addressing diversity was the
most salient challenge of the three and was also the connecting thread throughout this
dissertation. For instance, questions regarding the cross-cultural applicability of the HTQ trauma
symptom scales (and consequently the cross-cultural applicability of the PTSD construct) arose
as significant cultural variations were found in both the reliability and validity properties of the
scales. Other factors pertaining to diversity such as age and gender also had moderating effects in
both studies.
Despite the need for further investigation into the psychometric properties of the HTQ
symptom scales, the current dissertation has shed some light on the overall psychometric strength
151
of this measure. Researchers and clinicians can use the information provided in both studies to
help them decide whether to use this instrument for the assessment of PTSD with a specific
client or sample. Potential users of the HTQ are also encouraged to complement their knowledge
base with EBA guidelines to establish whether the HTQ is appropriate as is, or needs to be
adapted. If an adaptation is required, researchers need to refer to best practice cross-cultural
adaptation guidelines to help them in this process. In the end, following established EBA
guidelines while using the HTQ will not only help maximize the accuracy of the assessment
process, but also lead towards the ultimate goal of providing the best (and most deserving) care
for survivors of trauma.
152
References (General Introduction and Discussion)
Aboraya, A. (2009). Use of structured interviews by psychiatrists in real clinical settings: Results
of an open-question survey. Psychiatry, 6(6), 24-28.
Alcántara, C., Casement, D. M., & Lewis-Fernández, R. (2013). Conditional risk for PTSD
among Latinos: A systematic review of racial/ethnic differences and sociocultural
explanations. Clinical Psychology Review, 33, 107-119. doi: 10.1016/j.cpr.2012.10.005.
American Educational Research Association (AERA), American Psychological Association
(APA), & National Council on Measurement in Education (NCME) (2014). Standards
for educational and psychological testing. Washington, DC: AERA.
American Psychiatric Association (APA). (1952). Diagnostic and statistical manual of mental
disorders (1st ed.). Washington, DC: Author.
American Psychiatric Association. (1968). Diagnostic and statistical manual of mental disorders
(2nd ed.). Washington, DC: Author.
American Psychiatric Association. (1980). Diagnostic and statistical manual of mental disorders
(3rd ed.). Washington, DC: Author.
American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders
(3rd ed. revised). Washington, DC: Author.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders
(4th ed.). Washington, DC: Author.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders
(4th ed., text rev.). Washington, DC: Author.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental
disorders (5th ed.). Washington, DC: Author.
153
American Psychological Association Presidential Task Force on Evidence Based Practice
(2006). Evidence-based practice in psychology. American Psychologist, 61, 271-285. doi:
10.1037/0003-066X.61.4.271
American Psychological Association Publications and Communications Board Working Group
on JARS (2008). Reporting standards for research in psychology: Why do we need them?
What might they be? American Psychologist, 63, 839-851. doi: 10.1037/0003066X.63.9.839
Antony, M. M., & Rowa, K. (2005). Evidence-based assessment of anxiety disorders in adults.
Psychological Assessment, 17, 256-266. doi: 10.1037/1040-3590.17.3.256
Appelbaum, M., Cooper, H, Maxwell, S., Stone, A., & Sher, K. J. (2008). Reporting standards in
research in psychology: Why do we need them? What might they be? American
Psychologist, 63, 839-851. doi: 10.1037/0003-066X.63.9.839
Armour, C., Műllerová, J., & Elhai, J. D. (2016). A systematic literature review of PTSD’s latent
structure in the Diagnostic and Statistical Manual of Mental Disorders: DSM-IV to DSM5. Clinical Psychology Review, 44, 60-74. doi: 10.1016/j.cpr.2015.12.003
Arnetz, B. B., Broadbridge, C. L., Jamil, H., Lumley, M. A., Pole, N., Barkho, E., … Arnetz, J.
E. (2014). Specific trauma subtypes improve the predictive validity of the Harvard
Trauma Questionnaire in Iraqi refugees. Journal of Immigrant and Minority Health, 16,
1055-1061. doi: 10.1007/s10903-014-9995-9
Bahari, R., Alwi, M. N. M., Ahmad, M. R., & Saiboon, I. M. (2015). Translation and validation
of the Malay Posttraumatic Stress Disorder Checklist for Civilians. ASEAN Journal of
Psychiatry, 16(12), 203-211.
154
Barnes, L. L. B., Harp, D., & Jung, W. S. (2002). Reliability generalization of scores on the
Spielberger State-Trait Anxiety Inventory. Educational and Psychological Measurement,
62, 603-618. doi: 10.1177/0013164402062004005
Baschnagel, J. S., O’Connor, R. M., Colder, C. R., & Hawk, L. W. (2005). Factor structure of
posttraumatic stress among Western New York undergraduates following the September
11th terrorist attack on the World Trade Center. Journal of Traumatic Stress, 18, 677684. doi: 10.1002/jts.20076
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory-II.
San Antonio, TX: Psychological Corporation.
Beckham, J. C., Moore, S. D., Feldman, M. E., Hertzberg, M. A., Kirby, A. C., & Fairbank, J. A.
(1998). Health status, somatization, and severity of posttraumatic stress disorder in
Vietnam combat veterans with posttraumatic stress disorder. American Journal of
Psychiatry, 155, 1565-1569. doi: 10.1176/ajp.155.11.1565
Berry, J. W. (1989). Imposed etics-emics-derived etics: The operationalization of a compelling
idea. International Journal of Psychology, 24, 721–735. doi:
10.1080/00207598908247841
Bickman, L., Nurcombe, B., Townsend, C., Belle, M., Schut, J., & Karver, M. (1999).
Consumer measurement systems in child and adolescent mental health. Canberra,
Australia: Department of Health and Family Services.
Blake, D. D., Weathers, F., Nagy, L. M., Kaloupek, D. G., Klauminzer, G., Charney, D. S., &
Keane, T. M. (1990). A clinician rating scale for assessing current and lifetime PTSD:
The CAPS-1. Behavior Therapist, 13, 187-188.
Bovin, M. J., Marx, B. P., Weathers, F. W., Gallagher, M. W., Rodriguez, P., Schnurr, P. P., &
155
Keane, T. M. (2016). Psychometric properties of the PTSD Checklist for Diagnostic
and Statistical Manual of Mental Disorders–Fifth Edition (PCL-5) in veterans.
Psychological Assessment, 28, 1379-1391. doi: 10.1037/pas0000254
Brackbill, R. M., Hadler, J. L., DiGrande, L., Ekenga, C.C., Farfel, M.R., Friedman S.,
….Thorpe, L.E. (2009). Asthma and posttraumatic stress symptoms 5 to 6 years
following exposure to the World Trade Center terrorist attack. Journal of the American
Medical Association, 302, 502–516. doi: 10.1001/jama.2009.1121
Bracken, P. J., Giller, J. E., & Summerfield, D. (1995). Psychological responses to war and
atrocity: the limitations of current concepts. Social Science & Medicine, 40, 1073-1082.
doi: 10.1016/0277-9536(94)00181-R
Breslau, N., Davis, G., Andreski, P., & Peterson, E. (1991). Traumatic events and posttraumatic
stress disorder in an urban population of young adults. Archives of General Psychiatry,
48, 216-222. doi:10.1001/archpsyc.1991.01810270028003
Breslau, N., Kessler, R. C., Chilcoat, H. D., Schultz, L. R., Davis, G. C., & Andreski, P. (1998).
Trauma and posttraumatic stress disorder in the community: The 1996 Detroit Area
Survey of Trauma. Archives of General Psychiatry, 55, 626-631. doi:
10.1001/archpsyc.55.7.626
Breslau, N., Troost, J. P., Bohnert, K., & Luo, Z. (2013). Influence of predispositions on posttraumatic stress disorder: does it vary by trauma severity? Psychological Medicine, 43,
381-390. doi: 10.1017/S0033291712001195
Brewin, C. R., Andrews, B., & Rose, S. (2003). Diagnostic overlap between acute stress disorder
and PTSD in victims of violent crime. The American Journal of Psychiatry, 160, 783785. doi: 10.1176/appi.ajp.160.4.783
156
Briere, J. (1995). Trauma Symptom Inventory: Professional Manual. Odessa, FL: Psychological
Assessment Resources.
Brislin, R. W. (1970). Back-translation for cross-cultural research. Journal of Cross-Cultural
Psychology, 1, 185-216. doi: 10.1177/135910457000100301
Brislin, R. W. (1986). The wording and translation of research instruments. In W. J. Lonner & J.
W. Berry (Eds.), Field methods in cross-cultural research (pp. 137-164). Thousand Oaks,
CA: Sage Publications.
Calhoun, P. S., Hertzberg, J. S., Kirby, A. C., Dennis, M. F., Hair, L. P., Dedert, E. A., &
Beckham, J. C. (2012). The effect of draft DSM-V criteria on posttraumatic stress
disorder prevalence. Depression and Anxiety, 29, 1032-1042. doi: 10.1002/da.22012
Canadian Psychological Association Task Force on Evidence-Based Practice of Psychological
Treatments (2012). Evidence-based practice of psychological treatments: A Canadian
perspective. Retrieved from Canadian Psychological Association website:
http://www.cpa.ca/docs/File/Practice/Report_of_the_EBP_Task_Force_FINAL_Board_A
pproved_2012.pdf
Carlson, E. B., & Rosser-Hogan, R. (1991). Trauma experiences, posttraumatic stress,
dissociation, and depression in Cambodian refugees. American Journal of Psychiatry,
148, 1548-1551. doi: 10.1176/ajp.148.11.1548
Charney, M. E., & Keane, T. M. (2007). Psychometric analyses of the Clinician-Administered
PTSD Scale (CAPS) – Bosnian translation. Cultural Diversity & Ethnic Minority
Psychology, 13, 161-168. doi: 10.1037/1099-9809.13.2.161
Chhim, S. (2012). Baksbat (broken courage): The development and validation of the inventory to
measure baksbat, a Cambodian trauma-based cultural syndrome of distress. Culture,
157
Medicine and Psychiatry, 36, 640-659. doi: 10.1007/s11013-012-9279-6
Christon, L. M., McLeod, B. D., & Jensen-Doss, A. (2015). Evidence-based assessment meets
evidence-based treatment: An approach to science-informed case conceptualization.
Cognitive and Behavioral Practice, 22, 36-48. doi: 10.1016/j.cbpra.2013.12.004
Chun, K. M., Organista, P. B., & Marín G. (Eds.). (2003). Acculturation: Advances in theory,
measurement, and applied research. Washington, DC: American Psychological
Association.
Churchill, G. A., & Peter, J. P. (1984). Research design effects on the reliability of rating scales:
A meta-analysis. Journal of Marketing Research, 21(4), 360-374.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159. doi: 10.1037/00332909.112.1.155
Cohen, J. A., & the Work Group on Quality Issues (1998). Practice parameters for the
assessment and treatment of children and adolescents with posttraumatic stress disorder.
Journal of the American Academy of Child and Adolescent Psychiatry, 37, 4S-26S. doi:
10.1097/00004583-199810001-00002
Contractor, A. A., Durham, T. A., Brennan, J. A., Armour, C., Wutrick, H. R., Frueh, B. C., &
Elhai, J. D. (2014). DSM-5 PTSD's symptom dimensions and relations with major
depression's symptom dimensions in a primary care sample. Psychiatry Research, 215,
146-153. doi: 10.1016/j.psychres.2013.10.015
Cooper, H. M. (2011). Reporting research in psychology: How to meet journal article reporting
standards (6th ed.). Washington, DC: American Psychological Association.
Cox, B. J., Clara, I. P., & Enns, M. W. (2002). Posttraumatic stress disorder and the structure of
common mental disorders. Depression & Anxiety, 15, 168-171. doi: 10.1002/da.10052
158
Creamer, M., & Parslow, R. (2008). Trauma exposure and posttraumatic stress disorder in the
elderly: a community prevalence study. The American Journal of Geriatric Psychiatry,
16, 853-856. doi: 10.1097/01.JGP.0000310785.36837.85
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16,
297-334. doi:10.1007/BF02310555
da Mota Falcão, D., Ciconelli, R. M., & Ferraz, M. B. (2003). Translation and cultural adaptation
of quality of life questionnaires: An evaluation of methodology. The Journal of
Rheumatology, 30(2), 379-385.
de Fouchier, C., Blanchet, A., Hopkins, W., Bui, E., Ait-Aoudia, M., & Jehel, L. (2012).
Validation of a French adaptation of the Harvard Trauma Questionnaire among torture
survivors from sub-Saharan African countries. European Journal of Psychotraumatology,
3. doi: 10.3402/ejpt.v3i0.19225
de Jong, J. T., Komproe, I. H., Van Ommeren, M., El Masri, M., Araya, M., Khaled, N., …
Somasundaram, D. (2001). Lifetime events and posttraumatic stress disorder in 4
postconflict settings. Journal of the American Medical Association, 286, 555-562. doi:
10.1001/jama.286.5.555
Dermichyan, A., Goenjian, A. K., & Khachadourian, V. (2015). Factor structure and
psychometric properties of the Posttraumatic Stress Disorder (PTSD) Checklist and
DSM-5 PTSD symptom set in a long-term post-earthquake cohort in Armenia.
Assessment, 22, 594-606. doi: 10.1177/1073191114555523
Ditlevsen, D. N., & Elklit, A. (2010). The combined effect of gender and age on posttraumatic
stress disorder: Do men and women show differences in the lifespan distribution of the
disorder? Annals of General Psychiatry, 9. doi: 10.1186/1744-859X-9-32
159
Dolnicar, S., & Grün, B. (2007). Cross-cultural differences in survey response patterns.
International Marketing Review, 24, 127-143. doi: 10.1108/02651330710741785
Dunn, T. J., Baguley, T., & Brunsden, V. (2014). From alpha to omega: A practical solution to
the pervasive problem of internal consistency estimation. British Journal of Psychology,
105, 399-412. doi:10.1111/bjop.12046
Dye, D. A., Reck, M., & McDaniel, M. A. (1993). The validity of job knowledge measures.
International Journal of Selection and Assessment, 1(3), 153-157.
Elhai, J. D., Contractor, A. A., Tamburrino, M., Fine, T. H., Cohen, G., Shirley, E., …, Galea, S.
(2015). Structural relations between DSM-5 PTSD and major depression symptoms in
military soldiers. Journal of Affective Disorders, 175, 373-378. doi:
10.1016/j.jad.2015.01.034
Elhai, J. D., Gray, M. J., Kashdan, T. B., & Franklin, C. L. (2005). Which instruments are most
commonly used to assess traumatic event exposure and posttraumatic effects? A survey
of traumatic stress professionals. Journal of Traumatic Stress, 18, 541-545. doi:
10.1002/jts.20062
Elklit, A., & Shevlin, M. (2007). The structure of PTSD symptoms: A test of alternative models
using confirmatory factor analysis. British Journal of Clinical Psychology, 46, 299-313.
doi: 10.1348/014466506X171540
Elsass, P. (2001). Individual and collective traumatic memories: A qualitative study of
posttraumatic stress disorder symptoms in two Latin American localities. Transcultural
Psychiatry, 38, 306-316. doi: 10.1177/136346150103800302
Epstein, J., Osborne, R. H., Elsworth, G. R., Beaton, D. E., & Guillemin, F. (2013). Crosscultural adaption of the Health Education Impact Questionnaire: experimental study
160
showed expert committee, not back-translation, added value. Journal of Clinical
Epidemiology, 68, 360-369. doi: 10.1016/j.jclinepi.2013.07.013
Epstein, J., Santo, R. M., & Guillemin, F. (2015). A review of guidelines for cross-cultural
adaptation of questionnaires could not bring out a consensus. Journal of Clinical
Epidemiology, 68, 435-441. doi: 10.1016/j.jclinepi.2014.11.021
Eshun, S., & Gurung, R. A. (Eds.). (2009). Culture and mental health: Sociocultural influences,
theory and practice. Malden, MA: Blackwell Publishing Ltd.
Fazel, M., Wheeler, J., & Danesh, J. (2005). Prevalence of serious mental disorder in 7000
refugees resettled in western countries: A systematic review. Lancet, 365, 1309-1314.
doi: 10.1016/S0140-6736(05)61027-6
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1996). Structured Clinical
Interview for DSM-IV Axis I disorders, Clinician version. Washington, DC: American
Psychiatric Press, Inc.
First, M. B., Williams, J. B. W., Karg, R. S., & Spitzer, R. L. (2016). Structured Clinical
Interview for DSM-5 Disorders – Clinician version. Arlington, WA: American
Psychiatric Association Publishing.
Flaherty, J. A., Gaviria, F. M., Pathak, D., Mitchell, T., Wintrob, R., Richman, J. A., & Birz, S.
(1988). Developing instruments for cross-cultural psychiatric research. Journal of
Nervous and Mental Disease, 176, 257-263. doi: 10.1097/00005053-198805000-00001
Foa, E. B., Cashman, L., Jaycox, L., & Perry, K. (1997). The validation of a self-report measure
of posttraumatic stress disorder: The Posttraumatic Diagnostic Scale. Psychological
Assessment, 9, 445– 451. doi: 10.1037/1040-3590.9.4.445
161
Friedman, M. J. (2013). Finalizing PTSD in DSM-5: Getting here from there and where to go
next. Journal of Traumatic Stress, 26, 548-556. doi: 10.1002/jts.21840
Gagnon, A. J., Tuck, J., & Barkun, L. (2004). A systematic review of questionnaires measuring
the health of resettling refugee women. Health Care for Women International, 25, 111149. doi: 10.1080/07399330490267503
Gentes, E. L., Dennis, P. A., Kimbrel, N. A., Rissling, M. B., Beckham, J. C., VA Mid-Atlantic
MIRECC Workgroup, & Calhoun, P. S. (2014). DSM-5 posttraumatic stress disorder:
Factor structure and rates of diagnosis. Journal of Psychiatric Research, 59, 60-67. doi:
10.1016/j.jpsychires.2014.08.014
Gorst-Unsworth, C., & Goldenberg, E. (1998). Psychological sequelae of torture and organised
violence suffered by refugees from Iraq. Trauma-related factors compared with social
factors in exile. The British Journal of Psychiatry, 172, 90-94. doi: 10.1192/bjp.172.1.90
Graham, J. M., Yenling, J. L., & Jeziorski, J. L. (2006). The Dyadic Adjustment Scale: A
reliability generalization meta-analysis. Journal of Marriage and Family, 68, 701-717.
doi: 10.1111/j.1741-3737.2006.00284.x
Grubaugh, A. L., Long, M. E., Elhai, J. D., Frueh, B. C., & Magruder, K. M. (2010). An
examination of the construct validity of posttraumatic stress disorder with veterans using
a revised criterion set. Behavior, Research and Therapy. 48, 909-914. doi:
10.1016/j.brat.2010.05.019
Guajardo, M. G. U., Slewa-Younan, S., Smith, M., Eagar, S., & Stone, G. (2016). Psychological
distress is influenced by length of stay in resettled Iraqi refugees in Australia.
International Journal of Mental Health Systems, 10, 4. doi: 10.1186/s13033-016-0036-z
Hagengimana, A., & Hinton, D. (2009). Ihahamuka , a Rwandan syndrome of response to the
162
genocide: Blocked flow, spirit assault, and shortness of breath. In D. E. Hinton & B. J.
Good (Eds.), Culture and panic disorder (pp. 205-229). Stanford, CA: Stanford
University Press.
Harvard Program in Refugee Trauma (n.d.). Harvard Trauma Questionnaire: Kurdish version.
Harvard Program in Refugee Trauma (n.d.). Harvard Trauma Questionnaire: Russian version.
Harvard Program in Refugee Trauma (1996). Harvard Trauma Questionnaire: Japanese version.
Harvard Program in Refugee Trauma (1998). Harvard Trauma Questionnaire: Bosnian version.
Harvard Program in Refugee Trauma (1998). Harvard Trauma Questionnaire: Croatian version.
Harvard Program in Refugee Trauma (1998). Harvard Trauma Questionnaire: revised English
version.
Harvard Program in Refugee Trauma (2000). Harvard Trauma Questionnaire: revised
Cambodian version.
Harvard Program in Refugee Trauma (2000). Harvard Trauma Questionnaire: Peruvian version.
Harvard Program in Refugee Trauma (2000). Harvard Trauma Questionnaire: Spanish version.
Harvard Program in Refugee Trauma (2006). Harvard Trauma Questionnaire: Arabic version.
Harzing, A.-W. (2006). Response styles in cross-national survey research: A 26-country study.
International Journal of Cross-Cultural Management, 6, 243-266. doi:
10.1177/1470595806066332
Hauff, E., & Vaglum, P. (1994). Chronic posttraumatic stress disorder in Vietnamese refugees. A
prospective community study of prevalence, course, psychopathology, and stressors.
Journal of Nervous and Mental Disease, 182(2), 85-90.
Haynes, S. N., Richard, D. C. S., & Kubany, E. S. (1995). Content validity in psychological
assessment: A functional approach to concepts and methods. Psychological Assessment,
163
7, 238-247. doi: 10.1037/1040-3590.7.3.238
Haynes, S. N., Smith, G. T., & Hunsley, J. D. (2011). Scientific foundations of clinical
assessment. New York, NY: Routledge.
He, J., & van de Vijver, F. (2012). Bias and equivalence in cross-cultural research. Online
Readings in Psychology and Culture, 2. Retrieved from
http://scholarworks.gvsu.edu/cgi/viewcontent.cgi?article=1111&context=orpc
Hemphill, J. (2003). Interpreting the magnitudes of correlation coefficients.
American Psychologist, 58, 78-79. doi: 10.1037/0003-066X.58.1.78
Henchy, A. M. (2012). Review and evaluation of reliability generalization research
(Unpublished doctoral dissertation). University of Kentucky, Lexington, Kentucky.
Henson, R. K., Kogan, L. R., & Vacha-Haase, T. (2001). A reliability generalization of the
Teacher Efficacy Scale and related instruments. Educational and Psychological
Measurement, 61, 404-420. doi: 10.1177/00131640121971284
Hinton, D. E., Chhean, D., Pich, V., Um, K., Fama, J. M., & Pollack, M. H. (2006). Neckfocused panic attacks among Cambodian refugees: A logistic and linear regression
analysis. Journal of Anxiety Disorders, 20, 119-138. doi: 10.1016/j.janxdis.2005.02.001
Hinton, D. E., Hinton, A., Chhean, D., Pich, V., Loeum, J. R., & Pollack, M. H. (2009).
Nightmares among Cambodian refugees: The breaching of concentric ontological
security. Culture, Medicine, and Psychiatry, 33, 219-265. doi: 10.1007/s11013-009-91319
Hinton, D. E., Hofmann, S. G., Pitman, R. K., Pollack, M. H., & Barlow, D. H. (2008). The
panic attack-PTSD model: Applicability to orthostatic panic among Cambodian refugees.
Cognitive Behaviour Therapy, 37, 101-116. doi: 10.1080/16506070801969062
164
Hinton, D. E., & Lewis-Fernández, R. (2010). Idioms of distress among trauma survivors:
Subtypes and clinical utility. Culture, Medicine, and Psychiatry, 34, 209–218. doi:
10.1007/s11013-010-9175-x
Hinton, D. E., & Lewis-Fernández, R. (2011). The cross-cultural validity of posttraumatic stress
disorder: Implications for DSM-5. Depression & Anxiety, 28, 783-801. doi:
10.1002/da.20753
Hofstede, G., Hofstede, G. J., & Minkov, M. (2010). Cultures and organizations: Software of the
Mind, (3rd ed.). New York: McGraw-Hill.
Hollifield, M., Hewage, C., Gunawardena, C. N., Kodituwakku, P., Bopagoda, K., &
Weerarathnege, K. (2008). Symptoms and coping in Sri Lanka 20-21 months after the
2004 tsunami. The British Journal of Psychiatry, 192, 39-44. doi:
10.1192/bjp.bp.107.038422
Hollifield, M., Warner, T. D., Lian, N., Krakow, B., Jenkins, J. H., Kesler, J., … Westermeyer, J.
(2002). Measuring trauma and health status in refugees: A critical review. Journal of the
American Medical Association, 288, 611-621. doi: 10.1001/jama.288.5.611
Howell, R. T., & Shields, A. L. (2008). The file drawer problem in reliability generalization: A
strategy to compute a fail-safe N with reliability coefficients. Educational and
Psychological Measurement, 68, 120-128. doi:10.1177/0013164407301528
Hunsley, J., & Mash, E. J. (2007). Evidence-based assessment. Annual Review of Clinical
Psychology, 3, 29-51. doi: 10.1146/annurev.clinpsy.3.022806.091419
Hunsley, J., & Mash, E. J. (2008). Developing criteria for evidence-based assessment: An
introduction to assessments that work. In J. Hunsley & E. J. Mash (Eds.), A guide to
assessments that work (pp. 3-14). New York: Oxford University Press.
165
Hunsley, J., & Meyer, G. J. (2003). The incremental validity of psychological testing and
assessment: Conceptual, methodological, and statistical issues. Psychological
Assessment, 15, 446-455. doi: 10.1037/1040-3590.15.4.446
Ichikawa, M., Nakahara, S., & Wakai, S. (2006). Effect of post-migration detention on mental
health among Afghan asylum seekers in Japan. Australian and New Zealand Journal of
Psychiatry, 40, 341-346. doi: 10.1080/j.1440-1614.2006.01800.x
International Test Commission (2005). International Guidelines on Test Adaptation. Retrieved
from: http://www.intestcom.org
Jakobsen, M., Thoresen, S., & Johansen, L. E. E. (2011). The validity of screening for posttraumatic stress disorder and other mental health problems among asylum seekers from
different countries. Journal of Refugee Studies, 24, 171-186. doi: 10.1093/jrs/feq053
Jayawickreme, N., Jayawickreme, E., & Foa, E. B. (2012). Using the individualism-collectivism
construct to understand cultural differences in posttraumatic stress disorder. In K. M.
Gow, & M. J. Celinski (Eds.), Wayfinding through life’s challenges: Coping and
survival. New York: Nova Science Publishers.
Jenkins, J. H., & Valiente, M. (1994). Bodily transactions of the passions: El calor (the heat)
among Salvadoran women. In T. Csordas (Ed.), The body as existential ground: Studies
in culture, self, and experience (pp. 163-182). Cambridge, UK: Cambridge University
Press.
Jobson, L., & O’Kearney, R. (2008). Cultural differences in personal identity in posttraumatic
stress disorder. British Journal of Clinical Psychology, 47, 95-109. doi:
10.1348/014466507X235953
166
Jones, E., Vermaas, R. H., McCartney, H., Beech, C., Palmer, I., Hyams, K., & Wessely, S.
(2003). Flashbacks and posttraumatic stress disorder: The genesis of a 20th-century
diagnosis. British Journal of Psychiatry, 182, 158–163. doi: 10.1192/bjp.182.2.158
Kawakami, N., Takeshima, T., Ono, Y., Uda, H., Hata, Y., Nakane, Y., … Kikkawa, T. (2005).
Twelve-month prevalence, severity, and treatment of common mental disorders in
communities in Japan: preliminary finding from the World Mental Health Japan Survey
2002–2003. Psychiatry and Clinical Neurosciences, 59, 441-452. doi: 10.1111/j.14401819.2005.01397.x
Keane, T. M. (1995). Guidelines for the forensic psychological assessment of posttraumatic
stress disorder claimants. In R. I. Simon (Ed.), Posttraumatic stress disorder in litigation:
Guidelines for forensic assessment (pp. 99-115). Washington, DC: American Psychiatric
Press, Inc.
Keane, T. M., Buckley, T., & Miller, M. (2003). Guidelines for the forensic psychological
assessment of posttraumatic stress disorder claimants. In R. I. Simon (Ed.), Posttraumatic
stress disorder in litigation: Guidelines for forensic assessment (2nd ed., pp. 119-140).
Washington, DC: American Psychiatric Association Press, Inc.
Keane, T. M., Fairbank, J. A., Caddell, J. M., Zimering, R. T., & Bender, M. E. (1985). A
behavioral approach to assessing and treating posttraumatic stress disorder in Vietnam
veterans. In C. R. Figley (Ed.), Trauma and its wake (pp. 257–294). New York:
Brunner/Mazel.
Keane, T. M., Kaloupek, D. G., & Weathers, F. W. (1996). Ethnocultural considerations in the
assessment of PTSD. In A. J. Marsella, M. J. Friedman, E. T. Gerrity & R. M. Scurfield
(Eds.), Ethnocultural aspects of posttraumatic stress disorder: Issues, research, and
167
clinical applications (pp. 183-205). Washington, DC: American Psychological
Association.
Keane, T. M., Kolb, L. C., Kaloupek, D. G., Orr, S. P., Blanchard, E. B., Thomas, R. G., …
Lavori, P. W. (1998). Utility of psychophysiological measurement in the diagnosis of
posttraumatic stress disorder: results from a Department of Veterans Affairs Cooperative
Study. Journal of Consulting and Clinical Psychology, 66, 914-923. doi: 10.1037/0022006X.66.6.914
Keane, T. M., Silberbogen, A. K., & Weierich, M. R. (2008). Posttraumatic stress disorder. In J.
Hunsley & E.J. Mash (Eds.), A guide to assessments that work (pp. 293-315). New York:
Oxford University Press.
Kessler, R. C., Sonnega, A., Bromet, E., Hughes, M., & Nelson, C. B. (1995). Posttraumatic
stress disorder in the national comorbidity survey. Archives of General Psychiatry, 52,
1048-1060. doi: 10.1002/1099-1298(200011/12)
Kessler, R. C., Sonnega, A., Bromet, E., Hughes, M., Nelson, C. B., & Breslau, N. (1999).
Epidemiological risk factors for trauma and PTSD. In R. Yehuda (Ed.), Risk factors for
posttraumatic stress disorder (pp. 23-60). Washington, DC: American Psychiatric Press.
Kilpatrick, D. G., Resnick, H. S., Milanak, M. E., Miller, M. W., Keyes, K. M., & Friedman, M.
J. (2013). National estimates of exposure to traumatic events and PTSD prevalence using
DSM-IV and DSM-5 criteria. Journal of Traumatic Stress, 26, 537-547. doi:
10.1002/jts.21848
Kimberlin, C. L., & Winterstein, A. G. (2008). Validity and reliability of measurement
instruments in research. American Journal of Health-System Pharmacists, 65, 2276-2284.
doi: 10.2146/ajhp070364
168
King, D. W., Leskin, G. A., King, L. A., & Weathers, F. W. (1998). Confirmatory factor analysis
of the clinician-administered PTSD scale: evidence for the dimensionality of
posttraumatic stress disorder. Psychological Assessment, 10, 90-96. doi: 10.1037/10403590.10.2.90
Kirmayer, L. J., & Young, A. (1998). Culture and somatization: Clinical, epidemiological, and
ethnographic perspectives. Psychosomatic Medicine, 60(4), 420-430.
Kleijn, W. C., Hovens, J. E. & Rodenburg, J. J. (2001). Posttraumatic stress symptoms in
refugees: Assessments with the Harvard Trauma Questionnaire and the Hopkins
Symptom Checklist-25 in different languages. Psychological Reports, 88, 527-532. doi:
10.2466/PR0.88.2.527-532
Koenen, K. C., Ratanatharathorn, A., Ng, L., McLaughlin, K. A., Bromet, E. J., Stein, D. J., …
Kessler, R. C. (2017). Posttraumatic stress disorder in the World Mental Health Surveys,
Psychological Medicine, 1-15. doi: 10.1017/S0033291717000708
Krause, E. D., Kaltman, S., Goodman, L. A., & Dutton, M. A. (2007). Longitudinal factor
structure of posttraumatic stress disorder symptoms related to intimate partner violence.
Psychological Assessment, 19, 165-175. doi: 10.1037/1040-3590.19.2.165
Laban, C. J., Gernaat, H. B., Komproe, I. H., van der Tweel, I., & de Jong, J. T. (2005).
Postmigration living problems and common psychiatric disorders in Iraqi asylum seekers
in the Netherlands. Journal of Nervous and Mental Disease, 193, 825-832. doi:
00005053-200412000-00007
Lange, A., Rietdijk, D., Hudcovicova, M., van de Ven, J. P., Schrieken, B., & Emmelkamp, P.
M. (2003). Interapy: A controlled randomized trial of the standardized treatment of
169
posttraumatic stress through the internet. Journal of Consulting and Clinical Psychology,
71, 901–909. doi: 10.1037/0022-006X.71.5.901
Lauterbach, D., Vrana, S., King, D. W., & King, L. A. (1997). Psychometric properties of the
Civilian Version of the Mississippi PTSD Scale. Journal of Traumatic Stress, 10, 499513. doi: 10.1002/jts.2490100313
Leiknes, K. A., Finset, A., Moum, T., & Sandanger, I. (2007). Course and predictors of
medically unexplained pain symptoms in the general population. Journal of
Psychosomatic Research, 62, 119-128. doi: 10.1016/j.jpsychores.2006.08.009
Lewis-Fernández, R., Hinton, D. E., & Marques, L. (2014). Culture and PTSD. In M. J.
Friedman, T. M. Keane & P. A. Resick (Eds.) Handbook of PTSD (2nd ed., pp. 522-539).
New York, NY: Guilford Press.
López- López, J. A., Botella, J., Sánchez-Meca, J. & Marín-Martínez, F. (2013). Alternatives for
mixed-effects meta-regression models in the reliability generalization approach: A
simulation study. Journal of Educational and Behavioral Statistics, 38, 443-469. doi:
10.3102/1076998612466142
Maneesriwongul, W. & Dixon, J. K. (2004) Instrument translation process: A methods review.
Journal of Advanced Nursing, 48, 175-186. doi: 10.1111/j.1365-2648.2004.03185.x
Marsella, A. J., Friedman, M. J., Gerrity, E. T. & Scurfield, R. M. (1996). Ethnocultural aspects
of posttraumatic stress disorder: Issues, research, and clinical applications. Washington,
DC: American Psychological Association.
Matthieu, M., & Ivanoff, A. (2006). Treatment of human-caused trauma: Attrition in the adult
outcomes research. Journal of Interpersonal Violence, 2, 1654–1664. doi:
10.1177/0886260506294243
170
McCall, G., & Resick, P. (2003). A pilot study of PTSD symptoms among Kalahari Bushmen.
Journal of Traumatic Stress, 16, 445–450. doi: 10.1023/A:1025702326392
McDonald, R. P. (1999). Test theory: A unified treatment. Mahwah, NJ: Lawrence Erlbaum.
Medina-Mora, M., Borges, G., Lara, C., Benjet, C., Blanco, J., Fleiz, C., … Zambrano, J. (2005).
Prevalence, service use, and demographic correlates of 12-month DSM-IV psychiatric
disorders in Mexico: Results from the Mexican National Comorbidity Survey.
Psychological Medicine, 35, 1-11. doi: 10.1017/S0033291705005672
Messick, S. (1995). Validity of psychological assessment: Validation of inferences from persons’
response and performances as scientific inquiry into score meaning. American
Psychologist, 50, 741-749. doi: 10.1037/0003-066X.50.9.741
Michalopoulos, L. M., Unick, G. J., Haroz, E. E., Bass, J., Murray, L. K., & Bolton, P. A.
(2015). Exploring the fit of western PTSD models across three non-western low- and
middle-income countries. Traumatology, 21, 55-63. doi: 10.1037/trm0000020
Miles, J. N., Marshall, G. N., & Schell, T. L. (2008). Spanish and English versions of the PTSD
Checklist-Civilian version (PCL-C): Testing for differential item functioning. Journal
of Traumatic Stress, 21, 369-376. doi: 10.1002/jts.20349
Milfont, T. L., & Fischer, R. (2010). Testing measurement invariance across groups:
Applications in cross-cultural research. International Journal of Psychological Research,
3(1), 111-121.
Miller, M. W., Fogler, J. M., Wolf, E. J., Kaloupek, D. G., & Keane, T. M. (2008). The
internalizing and externalizing structure of psychiatric comorbidity in combat veterans.
Journal of Traumatic Stress, 21, 58–65. doi: 10.1002/jts.20303
Miller, M. W., Kaloupek, D. G., Dillon, A. L., & Keane, T. M. (2004). Externalizing and
171
internalizing subtypes of combat-related PTSD: A replication and extension using the
PSY-5 scales. Journal of Abnormal Psychology, 113, 636–645. doi: 10.1037/0021843X.113.4.636
Miller, K. E., Omidian, P., Kulkarni, M., Yaqubi, A., Daudzai, H., & Rasmussen, A. (2009). The
validity and clinical utility of posttraumatic stress disorder in Afghanistan. Transcultural
Psychiatry, 46, 219–237. doi: 10.1177/1363461509105813
Miller, K. E., Omidian, P. A., Quraishy, A. S., Quraishy, N., Nasiry, M. N., Nasiry, S., …
Yaqubi, A. A. (2006). The Afghan Symptom Checklist: A culturally grounded approach
to mental health assessment in a conflict zone. American Journal of Orthopsychiatry, 76,
423–433. doi: 10.1037/0002-9432.76.4.423
Mills, J., & Clark, M. S. (1982). Communal and exchange relationships. Review of
Personality and Social Psychology, 3, 121-144.
Mollica, R. F., Caspi-Yavin, Y., Bollini, P., Truong, T., Tor, S., & Lavelle, J. (1992). The
Harvard Trauma Questionnaire: Validating a cross-cultural instrument for measuring
torture, trauma, and posttraumatic stress disorder in Indochinese refugees. Journal of
Nervous and Mental Disease, 180, 111-116. doi: 10.1097/00005053-199202000-00008
Mollica, R. F., Caspi-Yavin, Y., Lavelle, J., Tor, S., Yang, T., Chan, S., ... de Marneffe, D.
(1994). The Harvard Trauma (HTQ) Manual: Cambodian, Laotian, and Vietnamese
versions (Unpublished manual). Cambridge, MA.
Mollica, R. F., McDonald, L. S., Massagli, M. P., & Silove, D. M. (2004). Measuring trauma,
measuring torture: Instructions and guidance on the utilization of the Harvard Program
in Refugee Trauma’s versions of the Hopkins Symptom Checklist-25 (HSCL-25) and the
Harvard Trauma Questionnaire (HTQ) (Unpublished manual). Cambridge, MA.
172
Mollica, R. F., McInness, K., Poole, C., & Tor, S. (1998). Dose–effect relationships
of trauma to symptoms of depression and post-traumatic stress disorder among
Cambodian survivors of mass violence. British Journal of Psychiatry, 173, 482-488. doi:
10.1192/bjp.173.6.482
Mollica, R. F., Wyshak, G., de Marneffe, D., & Lavelle, J. (1987). Indochinese versions of the
Hopkins Symptom Checklist-25: A screening instrument for the psychiatric care of
refugees. American Journal of Psychiatry, 144, 497-500. doi: 10.1176/ajp.144.4.497
Murphy, H. B. M. (1982) Comparative psychiatry: The International and intercultural
distribution of mental illness. Berlin, Germany: Springer-Verlag.
Nakeyar, C., & Frewen, P. A. (2016). Evidence-based care for Iraqi, Kurdish, and Syrian asylum
seekers and refugees of the Syrian civil war: A systematic review. Canadian Psychology,
57, 233-245. doi: 10.1037/cap0000067
Nock, M. K., Hwang, I., Sampson, N. A., & Kessler, R. C. (2010). Mental disorders,
comorbidity and suicidal behavior: Results from the National Comorbidity Survey
Replication. Molecular Psychiatry, 15, 868–876. doi:10.1038/mp.2009.29
North, C. S., Kawasaki, A., Spitznagel, E. L., & Hong, B. A. (2004). The course of PTSD, major
depression, substance abuse, and somatization after a natural disaster. Journal of Nervous
and Mental Disease, 192, 823-829. doi: 10.1097/01.nmd.0000146911.52616.22
Norris, A. E., & Aroian, K. J. (2008). Avoidance symptoms and assessment of posttraumatic
stress disorder in Arab immigrant women. Journal of Traumatic Stress, 21, 471-478. doi:
10.1002/jts.20363
173
Norris, F. H. (1992). Epidemiology of trauma: frequency and impact of different potentially
traumatic events on different demographic groups. Journal of Consulting and Clinical
Psychology, 60, 409-418. doi: 10.1037/0022-006X.60.3.409
Norris, F. H., & Slone, L. B. (2014). Epidemiology of trauma and PTSD. In M. J. Friedman, T.
M. Keane & P. A. Resick (Eds.), Handbook of PTSD: Science and practice (2nd ed., pp.
100-120). New York: The Guilford Press.
Norris, F. H., Van Landingham, M. J., & Vu, L. (2009). PTSD in Vietnamese Americans
following Hurricane Katrina: Prevalence, patterns, and predictors. Journal of Traumatic
Stress, 22, 91-101. doi: 10.1002/jts.20389
O’Donnell, M. L., Alkemade, N., Nickerson, A., Creamer, M., McFarlane, A. C., Silove,
D.,…Forbes, D. (2014). Impact of the diagnostic changes to post-traumatic stress
disorder for DSM-5 and the proposed changes to ICD-11. British Journal of Psychiatry,
205, 230-235. doi: 10.1192/bjp.bp.113.135285
Ones, D. S., Viswesvaran, C., & Schmidt, F. L. (2003). Personality and absenteeism: A metaanalysis of integrity tests. European Journal of Personality, 17, S19-S38. doi:
10.1002/per.487
Orr, S. P., Metzger, L. J., Miller, M. W., & Kaloupek, D. G. (2004). Psychophysiological
assessment of PTSD. In J. P. Wilson & T. M. Keane (Eds.), Assessing psychological
trauma and PTSD (2nd ed., pp. 289-343). New York: Guilford Press.
Palinkas, L. A., Petterson, J. S., Russell, J. C., & Downs, M. A. (2004). Ethnic differences in
symptoms of posttraumatic stress after the Exxon Valdez oil spil. Prehospital and
Disaster Medicine, 19, 102-112. doi: 10.1017/S1049023X00001552
174
Parry-Jones, B., & Parry-Jones, W. L. L. (1994). Posttraumatic stress disorder: Supportive
evidence from and eighteenth century natural disaster. Psychological Medicine, 24, 1527. doi: 10.1017/S0033291700026799
Paunović, N., & Öst, L.-G. (2005). Psychometric properties of a Swedish translation of the
Clinician Administered PTSD Scale-Diagnostic version. Journal of Traumatic Stress, 18,
161-164. doi: 10.1002/jts.20013
Perkonigg, A., Kessler, R. C., Storz, S., & Wittchen, H.-U. (2000). Traumatic events and posttraumatic stress disorder in the community: Prevalence, risk factors and comorbidity.
Acta Psychiatrica Scandinavica, 101, 46-59. doi: 10.1034/j.16000447.2000.101001046.x
Peter, J. P., & Churchill, G. A. (1986). Relationships among research design choices and
psychometric properties of rating scales: A meta-analysis. Journal of Marketing
Research, 23, 1–10. doi: 10.2307/3151771
Pietrzak, R. H., Goldstein, R. B., Southwick, S. M., & Grant, B. F. (2011). Prevalence and Axis I
comorbidity of full and partial posttraumatic stress disorder in the United States: Results
from Wave 2 of the National Epidemiologic Survey on Alcohol and Related Conditions.
Journal of Anxiety Disorders, 25, 456-465. doi: 10.1016/j.janxdis.2010.11.010
Pole, N., Gone, J. P., & Kulkarni, M. (2008). Posttraumatic stress disorder among ethnoracial
minorities in the United States. Clinical Psychology: Science and Practice, 15, 35–
61.
doi: 10.1111/j.1468-2850.2008.00109.x
Porter, M., & Haslam, N. (2005). Predisplacement and postdisplacement factors associated
with mental health of refugees and internally displaced persons: A meta-analysis.
Journal of the American Medical Association, 294, 602-612. doi:
175
10.1001/jama.294.5.602.
Rasmussen, A., Verkuilen, J., Ho, E., & Fan, Y. (2015). Posttraumatic stress disorder among
refugees: Measurement invariance of Harvard Trauma Questionnaire scores across global
regions and response patterns. Psychological Assessment, 27, 1160-1170. doi:
10.1037/pas0000115
Renner, W., Salem, I., & Ottomeyer, K. (2007). Posttraumatic stress in asylum seekers from
Chechnya, Afghanistan, and West Africa: Differential findings obtained by quantitative
and qualitative methods in three Austrian samples. In J. P. Wilson & C. C. Tang (Eds.),
Cross-cultural assessment of psychological trauma and PTSD (pp. 239-278). Boston,
MA: Springer.
Roberts, A. L., Gilman, S. E., Breslau, J., Breslau, N., & Koenen, K. C. (2011). Race/ethnic
differences in exposure to traumatic events, development of post-traumatic stress
disorder, and treatment-seeking for post-traumatic stress disorder in the United States.
Psychological Medicine, 41, 71-83. doi: 10.1017/S0033291710000401
Robins, L. N., Wing, J., Wittchen, H. U., Helzer, J. E., Babor, T. F, Burke, J., … Towle, M. S.
(1988). The Composite International Diagnostic Interview: An epidemiologic instrument
suitable for use in conjunction with different diagnostic systems and in different cultures.
Archives of General Psychiatry, 45, 1069-1077. doi:
10.1001/archpsyc.1988.01800360017003
Robinson, J. P., Shaver, P. R., & Wrightsman, L. S. (1991). Measures of personality and social
psychological attitudes. San Diego, CA: Elsevier.
Ruggiero, K. J., Del Ben, K., Scotti, J. R., & Rabalais, A. E. (2003). Psychometric properties of
the PTSD Checklist-Civilian version. Journal of Traumatic Stress, 16, 495-502. doi:
176
10.1023/A:1025714729117
Salkind, N. J. (2012). Tests and measurement for people who (think they) hate tests and
measurement (2nd ed.). Los Angeles, CA: Sage Publications.
Sam, D. L., & Moreira, V. (2002). The mutual embeddedness of culture and mental illness. In
W. J. Lonner, D. L. Dinnel, S. A. Hayes, & D. N. Sattler (Eds.), Online readings in
psychology and culture (Unit 9, Chapter 1). Bellingham, WA: Center for Cross-Cultural
Research, Western Washington University.
Schmidt, F. L., & Hunter, J. E. (1977). Development of a general solution to the problem of
validity generalization. Journal of Applied Psychology, 62, 529-540. doi: 10.1037/00219010.62.5.529
Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in
personnel psychology: Practical and theoretical implications of 85 years of research
findings. Psychological Bulletin, 124, 262-274. doi: 10.1.1.172.1733
Schnyder, U., & Moergeli, H. (2002). German version of Clinician-Administered PTSD Scale.
Journal of Traumatic Stress, 15, 487-492. doi: 10.1023/A:1020922023090
Schwarz-Nielsen, K. H., & Elklit, A. (2009). An evaluation of the mental status of
rejected asylum seekers in two Danish asylum centers. Torture, 19(1), 51-59.
Shoeb, M., Weinstein, H., & Mollica, R. (2007). The Harvard Trauma Questionnaire: Adapting a
cross-cultural instrument for measuring torture, trauma and posttraumatic stress disorder
in Iraqi refugees. The International Journal of Social Psychiatry, 53, 447-463. doi:
10.1177/0020764007078362
Shore, J. H., Orton, H., & Manson, S. M. (2009). Trauma-related nightmares among American
Indian veterans: Views from the dream catcher. American Indian and Alaska Native
177
Mental Health Research, 16, 25-38. doi: 10.5820/aian.1601.2009.25
Silove, D., Manicavasagar, V., Mollica, R., Thai, M., Khiek, D., Lavelle, J., & Tor, S. (2007).
Screening for depression and PTSD in a Cambodian population unaffected by war:
Comparing the Hopkins Symptom Checklist and Harvard Trauma Questionnaire with the
structured clinical interview. Journal of Nervous and Mental Disease, 195, 152-157. doi:
10.1097/01.nmd.0000254747.03333.70
Speroff, T., Sinnot, P. L., Marx, B., Owen, R. R., Jackson, J. C., Greevy, R., … Friedman, M. J.
(2012). Impact of evidence-based standardized assessment on the disability clinical
interview for diagnosis of service-connected PTSD: A cluster-randomized trial. Journal
of Traumatic Stress, 25, 607-615. doi: 10.1002/jts.21759
Spitzer, R. L., First, M. B., & Wakefield, J. C. (2007). Saving PTSD from itself in DSM-V.
Journal of Anxiety Disorders, 21, 233-241. doi: 10.1016/j.janxdis.2006.09.006
Steel, Z., Chey, T., Silove, D., Marmane, C., Bryant, R. A., & van Ommeren, M. (2009).
Association of torture and other potentially traumatic events with mental health outcomes
among populations exposed to mass conflict and displacement: A systematic review and
meta-analysis. Journal of the American Medical Association, 302, 537-549. doi:
10.1001/jama.2009.1132
Summerfield, D. (1999). A critique of seven assumptions behind psychological trauma
programmes in war-affected areas. Social Science and Medicine, 48, 1449–1462. doi:
10.1016/S0277-9536(98)00450-X
Summerfield, D. (2004). Cross-cultural perspectives on the medicalization of human suffering.
In G. M. Rosen (Ed.), Posttraumatic stress disorder: Issues and controversy (pp. 233–
245). Chichester, UK: John Wiley & Sons, Ltd.
178
Tang, S. S., & Fox, S. H. (2001). Traumatic experiences and the mental health of Senegalese
refugees. Journal of Nervous and Mental Disease, 189, 507-512.
Tavakol, M., & Dennick, R. (2011). Making sense of Cronbach’s alpha. International Journal of
Medical Education, 2, 53-55. doi: 10.5116/ijme.4dfb.8dfd
Thompson, B. (1994). Guidelines for authors. Educational and Psychological Measurement,
54(4), 837-847.
Thompson, B. (1999, January). Reliability generalization: An important meta-analysis method,
because it is incorrect to say “the test is unreliable.” Paper presented at the meeting of
the Southwest Educational Research Association, San Antonio, TX.
Triandis, H. C. (2001). Individualism-collectivism and personality. Personality, 69, 907-924.
doi: 10.1111/1467-6494.696169
Trimble, M. R. (1981). Post traumatic neuroses. Chichester, UK: John Wiley & Sons.
Vacha-Haase, T. (1998). Reliability generalization: Exploring variance in measurement error
affecting score reliability across studies. Educational and Psychological Measurement,
58, 6-20. doi: 10.1177/0013164498058001002
Vacha-Haase, T., Kogan, L.R., & Thompson, B. (2000). Sample compositions and variabilities
in published studies versus those in test manuals: Validity of score reliability inductions.
Educational and Psychological Measurement, 60, 509–522. doi:
10.1177/00131640021970682
Vacha-Haase, T., & Thompson, B. (2011). Score reliability: A retrospective look back
at 12 years of reliability generalization studies. Measurement and Evaluation in
Counseling and Development, 44, 159-168. doi: 10.1177/0748175611409845
Van der Velden, P. G., Grievink, L., Kleber, R. J., Drogendijk, A. N., Roskam, A. J.,
179
Marcelissen, F. G., … Gersons, B. P. (2006). Post-disaster mental health problems and
the utilization of mental health services: A four-year longitudinal comparative study.
Administration and Policy in Mental Health, 33, 279–288. doi: 10.1007/s10488-0050027-x
van de Vijver, F., & Tanzer, N. K. (2004). Bias and equivalence in cross-cultural assessment: An
overview. Revue européenne de psychologie appliquée, 54, 119-135. doi:
10.1016/j.erap.2003.12.004
Vassar, M., Knaup, K.G., Hale, W., & Hale, H. (2011). A meta-analysis of coefficient alpha for
the Impact of Event Scales: A reliability generalization study. South African Journal of
Psychology, 41, 6-16. doi: 10.1177/008124631104100102
Vloeberghs, E., van der Kwaak, A., Knipscheer, J., & van den Muijsenbergh (2012). Coping and
chronic psychosocial consequences of female genital mutilation in the Netherlands.
Ethnicity and Health, 17, 677-695. doi: 10.1080/13557858.2013.771148
Wang, R., Su, J., Bi., X., Wei, Y., & Mo, L. (2012). Application of the Chinese Posttraumatic
Stress Disorder Checklist to adolescent earthquake survivors in China. Social Behavior
and Personality: An International Journal, 40, 415-423. doi: 10.2224/sbp.2012.40.3.415
Warne, R. (2008, February). Reliability generalization (RG) analysis: The test is not reliable.
Paper presented at the meeting of the Southwest Educational Research Association, New
Orleans, LA.
Weathers, F. W., Blake, D. D., Schnurr, P. P., Kaloupek, D. G., Marx, B. P., & Keane, T. (2013).
Clinician-Administered PTSD Scale for DSM-5 (CAPS-5). Boston, MA: National Center
for PTSD.
180
Weathers, F. W., Litz, B. T., Herman, D. S., Huska, J. A., & Keane, T. M. (1993, October). The
PTSD checklist (PCL): Reliability, validity, and diagnostic utility. Poster presented at the
9th annual meeting of the International Society for Traumatic Stress Studies, San
Antonio, TX.
White, J., Pearce, J., Morrison, S., Dunstan, F., Bisson, J. I., & Fone, D. L. (2015). Risk of posttraumatic stress disorder following traumatic events in a community sample.
Epidemiology and Psychiatric Sciences, 24, 249-257. doi: 10.1017/S2045796014000110
Wilkinson, L., & American Psychological Association Task Force on Statistical Inference.
(1999). Statistical methods in psychology journals: Guidelines and explanations.
American Psychologist, 54, 594–604. doi:10.1037//0003-066X.54.8.594
Williams, D. R., Herman, A., Stein, D. J., Heeringa, S. G., Jackson, P. B., Moomal, H., &
Kessler, R. C. (2008). Twelve-month mental disorders in South Africa: Prevalence,
service use and demographic correlates in the population-based South African Stress and
Health Study. Psychological Medicine, 38, 211-220. doi: 10.1017/S0033291707001420
Yeomans, P. D., & Foreman, E. M. (2009). Cultural factors in traumatic stress. In S. Eshun, & R.
A. R. Gurung (Eds.), Culture and mental health: Sociocultural influences, theory, and
practice (pp. 221-244). Chichester, U.K.: John Wiley & Sons Ltd.
Yufik, T., & Simms, L. J. (2010). A Meta-analytic investigation of the structure of posttraumatic
stress symptoms. Journal of Abnormal Psychology, 119, 764-776. doi: 10.1037/a0020981
Zinbarg, R., Revelle, W., Yovel, I., & Li, W. (2005). Cronbach’s α, Revelle’s β and McDonald’s
ωη: Their relations with each other and two alternative conceptualizations of reliability.
Psychometrika, 70, 123–133. doi:10.1007/s11336-003-0974-7
181
Appendix A
Reliability and Validity Generalizations Coding Manual
182
Reliability and Validity Generalizations Coding Manual
Identification information
1) Document identifier: The document identifier is comprised of two sets of numbers. The first
set is the year of publication and the second set is the document number (e.g., 2010-18).
2) Sample of study: Indicate which sample of the study is used to calculate the reliability or
validity coefficients. If there are coefficients for multiple sub-samples (e.g., control group,
treatment group) within the same study complete separate coding sheets for each sub-sample.
Sample and sampling characteristics
3) Sample size: Code the sample size used to calculate the reliability/validity coefficients
(sometimes this number is not the total sample size of the study).
4-5) Age and gender: Code age & gender provided (even if this information is only provided for
the total sample and not the sample that is used for the reliability/validity coefficients).
6) Culture/Ethnicity: Indicate how the sample was described. There can be multiple
cultures/ethnicities. If not specified, code as culture/ethnicity of country of study (e.g.: a study
conducted in Denmark with no ethnicities/cultures specified, code “Danish”), unless it was
evident that there were various cultures/ethnicities such as sample comprised of refugees.
7) Country of origin (COO): Indicate each country if there are multiple. Code “various” if there
are too many to code (i.e., more than five).
8) Cultural orientation of COO: Please refer to the following website: http://geerthofstede.com/countries.html. Select whether the country’s culture is considered individualist
(score > 50) or collectivist (score < 50). If the country in question is not listed on the website,
select “information unavailable”.
183
9) Country of study: Indicate each country if there are multiple. Code “various” if there are too
many to code (i.e., more than five).
10) Recruitment setting: Indicate whether the sample was recruited in a mental health setting
(also includes psychiatric treatment and counselling), medical treatment setting (e.g., hospital,
primary care setting), community setting, or mixed sampling procedure.
11) Trauma category: Indicate which of the following traumatic events category best fits the
trauma experienced by the sample (select as many as necessary): (a) war, combat or terrorism
(e.g., war, civil conflict, political instability, political violence, ethnic persecution, human rights
violations); (b) torture (i.e., sample was described as torture survivors [including political
prisoners] – select only this category even if the sample also experienced war or combat); (c)
human trafficking (e.g., sex trafficking, trafficking for labour exploitation); (d) sexual assault
(e.g., sexual assault, rape, childhood sexual abuse, sexual violence); (e) nonsexual assault (e.g.,
robbery, crime victim, physical assault); (f) personal illness or injury (i.e., participants had an
illness [e.g., cancer] or injury [e.g., whiplash from car accident]; (g) witnessing death or injury
(e.g., witnessing a stabbing, witnessing a murder); (h) death, injury or illness of loved one; (i)
man-made disaster or fire (e.g., explosions, house fire); (j) natural disasters (i.e., disasters from
nature [e.g., tsunami, earthquake]); (k) various (i.e., participants experienced a range of traumas
[e.g., probability surveys]; (l) none (e.g., control group with no prior history of trauma). Code
“N/R” if the type of trauma is not indicated.
12) Forcibly displaced persons: Code whether the sample consisted of participants that were
forcibly displaced (e.g., refugees, asylum seekers, internally displaced) at any point in time (i.e.,
participants do not need to be forcibly displaced at the time of study).
Description of the HTQ administration and adaptation procedures
184
13) Language of administration of the Harvard Trauma Questionnaire (HTQ): Indicate the
language that the HTQ was administered in, not the language of the paper version of the HTQ
(e.g., the English paper version of the HTQ was used, but was administered orally in Urdu
through the use of an interpreter, code “Urdu”). Code each language that the instrument was
administered in. If there were more than five languages, code “Various”.
14) Original or adapted HTQ: Code whether the HTQ was used as is, or translated and/or
culturally adapted in any way. If there was no mention of this, but it was obvious that there
would have been an adaptation (i.e., different number of items, or they administered the HTQ to
individuals who didn’t speak the language of one of the original versions) select “adaptation”.
15) Was the sample of the study a sample for which the HTQ was originally validated? Original
samples include: Vietnamese, Cambodian (Khmer), Lao, Arabic (Iraqi) and Russian-speaking
refugees, Bosnian, Kurdish (Iraqi), Peruvian or Spanish (Peruvian) civilian survivors of war,
Croatian combat soldiers, and Japanese earthquake survivors.
16) Would an adaptation have been required for the sample? In some cases, an adaptation is not
explicitly described but it is obvious that an adaptation would have been needed. For example,
the HTQ was administered to a Swedish-speaking sample, but no Swedish version of the HTQ
has been approved by the developers. Furthermore, in some instances a cultural adaptation (or at
least a cultural verification with a pre-test) would have been needed even if the language of the
original HTQ was spoken by the population. For example, a sample of Liberian English speaking
political prisoners completed an English version of the HTQ.
17) Type of adaptation procedure described: Only code what was reported by the authors.
18) Authors stated they followed the guidelines recommended by the developers of the HTQ?
Only code what was reported by the authors.
185
19) Authors stated they followed guidelines developed by experts in translation/adaptation of
instruments? Only code what was reported by the authors.
20) Translation procedure used: Only select one of the following: (a) oral translation (i.e., the
HTQ was orally translated by the administrator or interpreters during the interviews); (b) forward
translation (code this when they simply say “was translated” or described as a one-way, simple
translation), (c) back translation (when translation was then translated back into source
language), or (d) blind back translation (the back translation was done by someone who did not
see the source document).
21) Did the authors report that a committee approach used during adaptation? Code “yes” if
there is mention of a consensus done among a group of people regarding the final version of the
adaptation (2 or more people).
22) Did the authors report that there was a cultural adaptation of the HTQ? Code “yes” if any
there is any evidence of this. Code “no” if they didn’t report anything about verifying cultural
differences or culturally adapting.
23) Did the authors report that the adapted HTQ was pre-tested? Code “yes” if there is any
evidence of a pre-test of the translation and/or cultural adaptation. Code “no” otherwise.
24) Administration of the HTQ was done by___: Indicate whether the administration of the HTQ
was done by: (a) self-report; (b) nonclinicians (e.g. lay interviewers, teachers); or (c) clinicians
(including research assistants/students in clinical psychology, psychiatry residents). If the
administration procedures are not clearly specified, code “N/R”.
Cronbach’s alpha values
25-31) Cronbach’s alpha values: Note the exact value reported by authors. Code “N/R” if the
value was not reported.
186
Construct validity coefficients
32) Version of HTQ symptom scale used: Select the version of the HTQ that was used to
calculate the validity coefficients.
33) Correlate: Select the construct that best described the variable being measured
34) Measure/Variable: Indicate which measure or variable the HTQ was compared to in the
bivariate analysis
35) Format & values: Provide the relevant values needed to compute a correlation.
36) Other validity coefficients: If there are more than one validity coefficient, please input
information in this section according to the various subheadings.
187
Appendix B
Reliability and Validity Generalizations Coding Sheet
188
Reliability and Validity Generalizations Coding Sheet
Identification information
1.
Document identifier:
Sample and sampling characteristics
3. Sample size:
4. Age (mean):
2. Sample of study:
5. Gender (% males):
6. Culture/ethnicity of sample:
7. Country of origin (COO) of sample:
8. Cultural orientation of COO ( ):
___ Individualist
___ Collectivist
___ Information unavailable
9. Country of study (COS):
10. Recruitment setting ( ):
___ mental health services
___ medical treatment
___ community
___ mixed
___ not reported (N/R)
11. Trauma category ( ):
___ war, combat or terrorism
___ torture
___ human trafficking
___ sexual assault
___ nonsexual assault
___ personal illness or injury
___ witnessing death or injury
___ death, injury or illness of loved one
___ man-made disaster or fire
___ natural disaster
___ various
___ other (specify:__________________)
___ none
___ N/R
12. Forcibly displaced persons (FDP) ( )?
___ yes
___ no
___ some
___ N/R
189
Description of the HTQ administration and adaptation procedures
13. Language(s) of administration of HTQ:
14. Original or adapted HTQ ( )?
___ original
___ adaptation
___ N/R
15. Was the sample of the study a sample for
which the HTQ was originally validated? ( )
___ yes
___ no
16. Would an adaptation/cultural verification
have been required for the sample? ( )
___ yes
___ no
17. Type of adaptation procedure described
( ):
___ translation only
___ cultural adaptation only
___ translation & cultural adaptation
___ N/R
___ not applicable (N/A) – original
HTQ
18. Authors stated they followed the guidelines
recommended by the developers of the HTQ?
( )
___ yes
___ no
___ N/A (original)
19. Authors stated they followed guidelines
developed by experts in translation/adaptation
of instruments? ( )
___ yes
___ no
___ N/A (original)
20. Translation procedure used ( ):
___ oral translation
___ forward translation
___ back translation
___ blind back translation
___ N/R
___ N/A (original)
21. Did the authors report that a committee
approach was used during adaptation? ( )
___ yes
___ no
___ N/A (original)
23. Did the authors report that the adapted
HTQ was pre-tested? ( )
___ yes
___ no
___ N/A (original)
22. Did the authors report that there was a
cultural adaptation of the HTQ? ( ):
___ yes
___ no
___ N/A (original)
24. Administration of the HTQ was done by
( ):
___ self-report
___ clinician
___ non-clinician
___ N/R
190
Cronbach’s alpha values of the HTQ symptom scales
25. HTQ-30 α:
26. HTQ-16 α:
28. Intrusion/Re-experiencing
subscale (4 items) α:
29. Avoidance/numbing
subscale (7 items) α:
27. HTQ-14 cultural
symptoms α:
30. Hypervigilance/arousal
subscale (5 items) α:
31. Other scale α (specify the scale):
Construct validity coefficients
32. Version of HTQ symptom scale used ( ):
___ 16 (PTSD – original HTQ)
___ 30 (incl. 16 PTSD – original HTQ)
33. Correlate ( ):
34. Measure/Variable (e.g.: HSCL-25 Anxiety,
BDI-II, witnessing violence)
Other PTSD measure___
Cumulative trauma___
Cultural symptoms___
Dissociation___
Anxiety & Depression___
Anxiety___
Depression___
Somatization___
Other:
________________________________
35. Format & values
Correlation (r or rho): ______ Sample size: _______ p-value: __________
Kappa (k): ________ Sample size: ________
ANOVA (comparison of two independent groups) F-value: _______ Group #1 n: ________
Group #2 n: _______
Odds ratio: ______ Confidence limits: ______
191
T-test (t-value): ________ Sample size: _________
Cohen’s d: ________ Sample size: ____________
Other format & values:
36. Other validity coefficients: If there is more than one validity coefficient, please input
information here :
Version of HTQ
Values
Correlate
Measure/variable
Format
192
Appendix C
Studies Included in the Reliability Generalization of the HTQ (Study 1)
193
Studies Included in the Reliability Generalization of the HTQ (Study 1)
Ahern, J., Galea, S., Fernandez, W. G., Koci, B., Waldman, R., & Vlahov, D. (2004). Gender,
social support, and posttraumatic stress in postwar Kosovo. Journal of Nervous and
Mental Disease, 192, 762-770. doi: 10.1097/01.nmd.0000144695.02982.41
Ahmad, A., von Knorring, A. L., & Sundelin-Wahlsten, V. (2008). Traumatic experiences and
post-traumatic stress disorder in Kurdistanian children and their parents in homeland and
exile: An epidemiological approach. Nordic Journal of Psychiatry, 62, 457-463. doi:
10.1080/08039480801984305
Andersen, T. E., Andersen, L.-A. C., & Andersen, P. G. (2014). Chronic pain patients with
possible co-morbid posttraumatic stress disorder admitted to multidisciplinary pain
rehabilitation: A 1-year cohort study. European Journal of Psychotraumatology, 5. doi:
10.3402/ejpt.v5.23235
Andersen, T. E., Elklit, A., & Brink, O. (2013). PTSD symptoms mediate the effect of
attachment on pain and somatization after whiplash injury. Clinical Practice and
Epidemiology in Mental Health, 9, 75-83. doi: 10.2174/1745017901309010075
Anwar, J., Mpofu, E., Matthews, L. R., & Brock, K. E. (2013). Risk factors of posttraumatic
stress disorder after an earthquake disaster. Journal of Nervous and Mental Disease, 201,
1045-1052. doi: 10.1097/NMD.0000000000000060
Badri, A., Crutzen, R., & Van den Borne, H. W. (2012). Exposures to war-related traumatic
events and post-traumatic stress disorder symptoms among displaced Darfuri female
university students: An exploratory study. BioMed Central Public Health, 12. doi:
10.1186/1471-2458-12-603
194
Bentley, J. A. (2011). Cross-cultural assessment of psychological symptoms among Somali
refugees (Unpublished doctoral dissertation). Seattle Pacific University, Seattle, WA.
Bödvarsdóttir, I., & Elklit, A. (2004). Psychological reactions in Icelandic earthquake survivors.
Scandinavian Journal of Psychology, 45, 3-13. doi: 10.1111/j.14679450.2004.00373.x
Bödvarsdóttir, I., & Elklit, A. (2007). Victimization and PTSD-like states in an Icelandic youth
probability sample. BMC Psychiatry, 7, doi: 10.1186/1471-244x-7-51
Cardozo, B. L., Holtz, T .H., Kaiser, R., Gotway, C. A., Ghitis, F., Toomey, E., & Salama, P.
(2005). The mental health of expatriate and Kosovar Albanian humanitarian aid workers.
Disasters, 29, 152-170. doi: 10.1111/j.0361-3666.2005.00278.x
Carlsson, J. M., Mortensen, E. L., & Kastrup, M. (2006). Predictors of mental health and quality
of life in male tortured refugees. Nordic Journal of Psychiatry, 60, 51-57. doi:
10.1080/08039480500504982
Christiansen, D., Bak, R., & Elklit, A. (2012). Secondary victims of rape. Violence and Victims,
27, 246-262. doi: 10.1891/0886-6708.27.2.246
Christiansen, D., Elklit, A., & Olff, M. (2013). Parents bereaved by infant death: PTSD
symptoms up to 18 years after the loss. General Hospital Psychiatry, 35, 605-611. doi:
10.1016/j.genhosppsych.2013.06.006
Cleveland, J., & Rousseau, C. (2013). Psychiatric symptoms associated with brief detention of
adult asylum seekers in Canada. Canadian Journal of Psychiatry, 58, 409-416. doi:
10.1177/070674371305800706
Correa-Velez, I., Barnett, A. G., Gifford, S. M., & Sackey, D. (2011). Health status and use of
health services among recently arrived men with refugee backgrounds: a comparative
195
analysis of urban and regional settlement in South-east Queensland. Australian Journal of
Primary Health, 17, 66-71. doi: 10.1071/py10051
Crescenzi, A., Ketzer, E., Van Ommeren, M., Phuntsok, K., Komproe, I., & De Jong, J. T. V. M.
(2002). Effect of political imprisonment and trauma history on recent Tibetan refugees in
India. Journal of Traumatic Stress, 15, 369-375. doi: 10.1023/A:1020129107279
Culhane, M. A. (2009). Suffering strangers: The social construction of the torture survivor in the
American healing community (Unpublished doctoral dissertation). Boston University,
Boston, MA.
de Fouchier, C., Blanchet, A., Hopkins, W., Bui, E., Ait-Aoudia, M., & Jehel, L. (2012).
Validation of a French adaptation of the Harvard Trauma Questionnaire among torture
survivors from sub-Saharan African countries. European Journal of Psychotraumatology,
3. doi: 10.3402/ejpt.v3i0.19225
Domanskaité-Gota, V., Elklit, A., & Christiansen, D. M. (2009). Victimization and PTSD in a
Lithuanian national youth probability sample. Nordic Psychology, 61, 66-81. doi:
10.1027/1901-2276.61.3.66
Doran, J. M., Kalayjian, A., Toussaint, L., & DeMucci, J. (2012). The relationship between
trauma and forgiveness in post-conflict Sierra Leone. Psychological Trauma-Theory,
Research, Practice, and Policy, 4, 614-623. doi: 10.1037/a0025470
Eisen, E. M. (2016). The impact on post-migration factors on PTSD and depressive symptoms
among asylum seekers in the United States (Unpublished doctoral dissertation). George
Washington University, Washington, DC.
Elklit, A. (2002a). Acute stress disorder in victims of robbery and victims of assault. Journal of
Interpersonal Violence, 17, 872-887. doi: 10.1177/0886260502017008005
196
Elklit, A. (2002b). Victimization and PTSD in a Danish national youth probability sample.
Journal of the American Academy of Child and Adolescent Psychiatry, 41, 174-181. doi:
10.1097/00004583-200202000-00011
Elklit, A., & Blum, A. (2011). Psychological adjustment one year after the diagnosis of breast
cancer: A prototype study of delayed post-traumatic stress disorder. British Journal of
Clinical Psychology, 50, 350-363. doi: 10.1348/014466510x527676
Elklit, A., & Brink, O. (2003). Acute stress disorder in physical assault victims visiting a Danish
emergency ward. Violence and Victims, 18, 461-472. doi: 10.1891/vivi.2003.18.4.461
Elklit, A., & Kurdahl, S. (2013). The psychological reactions after witnessing a killing in public
in a Danish high school. European Journal of Psychotraumatology, 4. doi:
10.3402/ejpt.v4i0.19826
Elklit, A., & O'Connor, M. (2005). Post-traumatic stress disorder in a Danish population of
elderly bereaved. Scandinavian Journal of Psychology, 46, 439-445. doi: 10.1111/j.14679450.2005.00475.x
Elklit, A., Christiansen, D. M., Palic, S., Karsberg, S., & Eriksen, S. B. (2014). Impact of
traumatic events on posttraumatic stress disorder among Danish survivors of sexual abuse
in childhood. Journal of Child Sexual Abuse, 23, 918-934. doi:
10.1080/10538712.2014.964440
Elklit, A., Hartvig, T., & Christiansen, M. (2007). Psychological sequelae in parents of extreme
low and very low birth weight infants. Journal of Clinical Psychology in Medical
Settings, 14, 238-247. doi: 10.1007/s10880-007-9077-4
Ellis, B. H., Abdi, S. M., Lazarevic, V., White, M. T., Lincoln, A. K., Stern, J. E., & Horgan, J.
G. (2016). Relation of psychosocial factors to diverse behaviors and attitudes among
197
Somali refugees. American Journal of Orthopsychiatry, 86, 393-408. doi:
10.1037/ort0000121
Eng, S., Mulsow, M., Cleveland, H., & Hart, S. L. (2009). Academic achievement among
adolescents in Cambodia: Does caregiver trauma matter? Journal of Community
Psychology, 37, 754-768. doi: 10.1002/jcop.20329
Farhood, L. F., & Dimassi, H. (2012). Prevalence and predictors for post-traumatic stress
disorder, depression and general health in a population from six villages in South
Lebanon. Social Psychiatry and Psychiatric Epidemiology, 47, 639-649. doi:
10.1007/s00127-011-0368-6
Finklestein, M., & Solomon, Z. (2009). Cumulative trauma, PTSD and dissociation among
Ethiopian refugees in Israel. Journal of Trauma and Dissociation, 10, 38-56. doi:
10.1080/15299730802485151
Gordon, J. S., Staples, J. K., Blyta, A., Bytyqi, M., & Wilson, A. T. (2008). Treatment of
posttraumatic stress disorder in postwar Kosovar adolescents using mind-body skills
groups: A randomized controlled trial. Journal of Clinical Psychiatry, 69(9), 1469-1476.
Gordon, J. S., Staples, J. K., He, D. Y., Atti, J. A. A. (2016). Mind–body skills groups for
posttraumatic stress disorder in Palestinian adults in Gaza. Traumatology, 22, 155-164.
doi: 10.1037/trm0000081
Gupta, J., Falb, K. L., Carliner, H., Hossain, M., Kpebo, D., & Annan, J. (2014). Associations
between exposure to intimate partner violence, armed conflict, and probable PTSD
among women in rural Côte d'Ivoire. PLOS One, 9, e96300. doi:
10.1371/journal.pone.0096300
198
Halvorsen, J. O., & Kagee, A. (2010). Predictors of psychological sequelae of torture among
South African former political prisoners. Journal of Interpersonal Violence, 25, 9891005. doi: 10.1177/0886260509340547
Hansen, N. B., Eriksen, S. B., & Elklit, A. (2014). Effects of an intervention program for female
victims of intimate partner violence on psychological symptoms and perceived social
support. European Journal of Psychotraumatology, 5. doi: 10.3402/ejpt.v5.24797
Harper, G. W., Wade, R. M., Onyango, D. P., Abuor, P. A., Bauermeister, J. A., Odero, W. W.,
& Bailey, R. C. (2015). Resilience among gay/bisexual young men in Western Kenya:
Psychosocial and sexual health outcomes. AIDS, 29, S261-S269. doi:
10.1097/QAD.0000000000000905
Hensen, M., Andersen, T. E., Armour, C., Elklit, A., Palic, S., & Mackrill, T. (2010). PTSD-8: A
short PTSD inventory. Clinical Practice and Epidemiology in Mental Health, 6, 101-108.
doi: 10.2174/1745017901006010101
Hossain, M., Zimmerman, C., Abas, M., Light, M., & Watts, C. (2010). The relationship of
trauma to mental disorders among trafficked and sexually exploited girls and women.
American Journal of Public Health, 100, 2442-2449. doi: 10.2105/AJPH.2009.173229
Huijts, I., Kleijn, W. C., van Emmerik, A. A. P., Noordhof, A., & Smith, A. J. M. (2012).
Dealing with man-made trauma: The relationship between coping style, posttraumatic
stress, and quality of life in resettled, traumatized refugees in the Netherlands. Journal of
Traumatic Stress, 25, 71-78. doi: 10.1002/jts.21665
Hussain, D., & Bhushan, B. (2009). Development and validation of the Refugee Trauma
Experience Inventory. Psychological Trauma: Theory, Research, Practice and Policy, 1,
107-117. doi: 10.1037/a0016120
199
Ichikawa, M., Nakahara, S., & Wakai, S. (2006). Effect of post-migration detention on mental
health among Afghan asylum seekers in Japan. Australian and New Zealand Journal of
Psychiatry, 40, 341-346. doi: 10.1080/j.1440-1614.2006.01800.x
Jakobsen, M., Thoresen, S., & Johansen, L. E. E. (2011). The validity of screening for posttraumatic stress disorder and other mental health problems among asylum seekers from
different countries. Journal of Refugee Studies, 24, 171-186. doi: 10.1093/jrs/feq053
Jind, L., Elklit, A., & Christiansen, D. (2010). Cognitive schemata and processing among parents
bereaved by infant death. Journal of Clinical Psychology in Medical Settings, 17, 366377. doi: 10.1007/s10880-010-9216-1
Jones, L., & Kafetsios, K. (2002). Assessing adolescent mental health in war-affected societies:
The significance of symptoms. Child Abuse & Neglect, 26, 1059-1080. doi:
10.1016/S0145-2134(02)00381-2
Kagee, K. (2005). Symptoms of distress and posttraumatic stress among South African former
political detainees. Ethnicity & Health, 10, 169-179. doi: 10.1080/13557850500071244
Kanninen, K., Punamaki, S. L., & Qouta, S. (2002). The relation of appraisal, coping efforts, and
acuteness of trauma to PTS symptoms among former political prisoners. Journal of
Traumatic Stress, 15, 245-253. doi: 10.1023/A:1015211529584
Kanninen, K., Salo, J., & Punamaki, R. L. (2000). Attachment patterns and working alliance in
trauma therapy for victims of political violence. Psychotherapy Research, 10, 435-449.
doi: 10.1093/ptr/10.4.435
Karsberg, S., Armour, C., & Elklit, A. (2014). Patterns of victimization, suicide attempt, and
posttraumatic stress disorder in Greenlandic adolescents: A latent class analysis. Social
200
Psychiatry and Psychiatric Epidemiology, 49, 1389-1399. doi: 10.1007/s00127-0140890-4
Keller, A., Lhewa, D., Rosenfeld, B., Sachs, E., Aladjem, A., Cohen, I., . . . Portetfield, K.
(2006). Traumatic experiences and psychological distress in an urban refugee population
seeking treatment services. Journal of Nervous and Mental Disease, 194, 188-194. doi:
10.1097/01.nmd.0000202494.75723.83
Kleijn, W. C., Hovens, J. E., & Rodenburg, J .J. (2001). Posttraumatic stress symptoms in
refugees: Assessments with the Harvard Trauma Questionnaire and the Hopkins
Symptom Checklist-25 in different languages. Psychological Reports, 88, 527-532. doi:
10.2466/PR0.88.2.527-532
Kristensen, T. E., Elklit, A., Karstoff, K.-A., & Palic, S. (2014). Predicting chronic posttraumatic
stress disorder in bereaved relatives: A 6-month follow-up study. American Journal of
Hospice and Palliative Medicine, 31, 396-405. doi: 10.1177/1049909113490066
Kuo, C., Reddy, M. K., Operario, D., Cluver, L., & Stein, D. J. (2013). Posttraumatic stress
symptoms among adults caring for orphaned children in HIV-endemic South Africa.
AIDS and Behavior, 17, 1755-1763. doi: 10.1007/s10461-013-0461-9
Leaman, S. C. (2009). Risk factors for psychological distress and uses of religious coping
(Unpublished doctoral dissertation). The George Washington University, Washington,
DC.
Lhewa, D. W. (2010). Coping and distress among Tibetan survivors of torture and refugee
trauma (Unpublished doctoral dissertation). Boston University, Boston, MA.
201
Machisa, M. T., Christofides, N., & Jewkes, R. (2016). Structural pathways between child abuse,
poor mental health outcomes and male-perpetrated intimate partner violence (IPV). PLoS
ONE, 11, e0150986. doi:10.1371/journal.pone.0150986
Mai, L. G. (2010). Exploration of the intergenerational transmission of historical trauma in
Vietnamese Americans (Unpublished doctoral dissertation). California Institute of
Integral Studies, San Francisco, CA.
Michalopoulos, L. (2013). Exploration of cross-cultural adaptability of PTSD among trauma
survivors in Northern Iraq, Thailand, and Democratic Republic of Congo: Application of
item response theory and classical test theory (Unpublished doctoral dissertation).
University of Maryland, Baltimore, MD.
Mollica, R. F., Caspi-Yavin, Y., Bollini, P., Truong, T., Tor, S., & Lavelle, J. (1992). The
Harvard Trauma Questionnaire: Validating a cross-cultural instrument for measuring
torture, trauma, and posttraumatic stress disorder in Indochinese refugees. Journal of
Nervous and Mental Disease, 180, 111-116. doi: 10.1097/00005053-199202000-00008
Mollica, R. F., Chernoff, M. C., Berthold, S. M., Lavelle, J., Lyoo, I. K., & Renshaw, P. (2014).
The mental health sequelae of traumatic head injury in South Vietnamese ex-political
detainees who survived torture. Comprehensive Psychiatry, 55, 1626-1638. doi:
10.1016/j.comppsych.2014.04.014
Mollica, R. F., McInnes, K., Sarajlic, N., Lavelle, J., Sarajlic, J., & Massagli, M. P. (1999).
Disability associated with psychiatric comorbidity and health status in Bosnian refugees
living in Croatia. Journal of the American Medical Association, 282, 433-439. doi:
10.1001/jama.282.5.433
202
Morina, N., & Ford, J. D. (2008). Complex sequelae of psychological trauma among Kosovar
civilian war victims. International Journal of Social Psychiatry, 54, 425-436. doi:
10.1177/0020764008090505
Nickerson, A., Bryant, R. A., Brooks, R., Steel, Z., Silove, D., & Chen, J. (2011). The familial
influence of loss and trauma on refugee mental health: A multilevel path analysis.
Journal of Traumatic Stress, 24, 25-33. doi: 10.1002/jts.20608
O'Connor, M. (2010). PTSD in older bereaved people. Aging and Mental Health, 14, 670-678.
doi: 10.1080/13607860903311725
O'Connor, M., & Elklit, A. (2008). Attachment styles, traumatic events, and PTSD: a crosssectional investigation of adult attachment and trauma. Attachment & Human
Development, 10, 59-71. doi: 10.1080/14616730701868597
Opaas, M., & Hartmann, E. (2013). Rorschach assessment of traumatized refugees: An
exploratory factor analysis. Journal of Personality Assessment, 95, 457-470. doi:
10.1080/00223891.2013.781030
Oruc, L., Kapetanovic, A., Pojskic, N., Miley, K., Forstbauer, S., Mollica, R. F., & Henderson,
D. C. (2008). Screening for PTSD and depression in Bosnia and Herzegovina: Validating
the Harvard Trauma Questionnaire and the Hopkins Symptom Checklist. International
Journal of Culture and Mental Health, 1, 105-116. doi: 10.1080/17542860802456620
Petersen, T., Elklit, A., & Olesen, J. G. (2010). Victimization and PTSD in a Faroese youth totalpopulation sample. Scandinavian Journal of Psychology, 51, 56-62. doi: 10.1111/j.14679450.2009.00728.x
203
Pham, K. T. (2013). Trauma transmission between first-generation and second-generation
Vietnamese refugees and immigrants in the United States: A study of mother and
daughter (Unpublished doctoral dissertation). Palo Alto University, Palo Alto, CA.
Renner, W., Salem, I., & Ottomeyer, K. (2006). Cross-cultural validation of measures of
traumatic symptoms in groups of asylum seekers from Chechnya, Afghanistan, and West
Africa. Social Behavior and Personality, 34, 1101-1114. doi:
10.2224/sbp.2006.34.9.1101
Roberts, B., Damundu, E. Y., Lomoro, O., & Sondorp, E. (2009). Post-conflict mental health
needs: A cross-sectional survey of trauma, depression and associated factors in Juba,
Southern Sudan. BioMed Central Psychiatry, 9, 7. doi: 10.1186/1471-244X-9-7
Roberts, B., Ocaka, K. F., Browne, J., Oyok, T., & Sondorp, E. (2008). Factors associated with
post-traumatic stress disorder and depression amongst internally displaced persons in
northern Uganda. BMC Psychiatry, 8, 38. doi: 10.1186/1471-244x-8-38
Roth, G., Ekbad, S., & Agren, H. (2006). A longitudinal study of PTSD in a sample of adult
mass-evacuated Kosovars, some of whom returned to their home country. European
Psychiatry, 21, 152-159. doi: 10.1016/j.eurpsy.2005.11.004
Rothon, C., Stansfeld, S. A., Mathews, C., Kleinhans, A., Clark, C., Lund, C., & Flisher, A. J.
(2011). Reliability of self-report questionnaires for epidemiological investigations of
adolescent mental health in Cape Town, South Africa. Journal of Child and Adolescent
Mental Health, 23, 119-128. doi: 10.2989/17280583.2011.634551
Sachs, E., Rosenfeld, B., Lhewa, D., Rasmussen, A., & Keller, A. (2008). Entering exile:
Trauma, mental health, and coping among Tibetan refugees arriving in Dharamsala,
India. Journal of Traumatic Stress, 21, 199-208. doi: 10.1002/jts.20324
204
Schubert, S. J., Lee, C. W., de Araujo, G., Butler, S. R., Taylor, G., & Drummond, P. D. (2016).
The effectiveness of Eye Movement Desensitization and Reprocessing Therapy to treat
symptoms following trauma in Timor Leste. Journal of Traumatic Stress, 29, 141-148.
doi: 10.1002/jts.22084
Semb, O., Henningsson, M., Fransson, P., & Sundbom, E. (2009). Trauma-related symptoms
after violent crime: The role of risk factors before, during and eight months after
victimization. The Open Psychology Journal, 2, 77-88. doi:
10.2174/1874350100902010077
Silove, D., Brooks, R., Steel Bateman, C., Steel, Z., Fonseca, Z., Amaral, C.,… Soosay, I.
(2010). Social and trauma-related pathways leading to psychological distress and
functional limitations four years after the humanitarian emergency in Timor Leste.
Journal of Traumatic Stress, 23, 151-160. doi: 10.1002/jts.20499
Silove, D., Rees, S., Tay, A. K., da Costa, Z. M., Soares Savio, E., Soares, C., & Tol, W. (2015).
Pathways to perinatal depressive symptoms after mass conflict in Timor-Leste: A
modelling analysis using cross-sectional data. Lancet Psychiatry, 2, 161-167. doi:
10.1016/S2215-0366(14)00054-6
Slewa-Younan, S., Chippendale, K., Heriseanu, A., Lujic, S., Atto, J., & Raphael, B. (2012).
Measures of psychophysiological arousal among resettled traumatized Iraqi refugees
seeking psychological treatment. Journal of Traumatic Stress, 25, 348-352. doi:
10.1002/jts.21694
Slewa-Younan, S., Mond, J. M., Bussion, E., Melkonian, M., Mohammad, Y., Dover, H., …
Jorm, A. F. (2015). Psychological trauma and help seeking behaviour amongst resettled
205
Iraqi refugees in attending English tuition classes in Australia. International Journal of
Mental Health Systems, 9. doi: 10.1186/1752-4458-9-5.
Sondergaard, H. P., Ekblad, S., & Theorell, T. (2001). Self-reported life event patterns and their
relation to health among recently resettled Iraqi and Kurdish refugees in Sweden. Journal
of Nervous and Mental Disease, 189(12), 838-845.
Southivong, B., Ichikawa, M., Nakahara, S., & Southivong, C. (2013). A cross-sectional
community study of posttraumatic stress disorder and social support in Lao People's
Democratic Republic. Bulletin of the World Health Organization, 91, 765-772. doi:
10.2471/BLT.12.115311
Soysa, C. K. (2013). War and tsunami PTSD responses in Sri Lankan children: Primacy of
re-experiencing and arousal compared to avoidance-numbing. Journal of Aggression,
Maltreatment and Trauma, 22, 896-915. doi: 10.1080/10926771.2013.824056
Suarez, E. B. (2011). Surviving the "Sasachacuy Tiempu "[difficult times]: The resilience of
Quechua women in the aftermath of the Peruvian armed conflict (Unpublished doctoral
dissertation). University of Toronto, Toronto, ON.
Veer, I. M., Oei, N. Y. L., van Buchem, M. A., Spinhoven, P., Elzinga, B. M., & Rombouts, S.
A. R. B. (2015). Evidence for smaller right amygdala volumes in posttraumatic stress
disorder following childhood trauma. Psychiatry Research: Neuroimaging, 233, 436-442.
doi: 10.1016/j.pscychresns.2015.07.016
Vervliet, M., Meyer Demott, M. A., Jakobsen, M., Broekaert, E., Heir, T., & Derluyn, I. (2014).
The mental health of unaccompanied refugee minors on arrival in the host country.
Scandinavian Journal of Psychology, 33, 33-37. doi: 10.1111/sjop.12094
206
Vloeberghs, E., van der Kwaak, A., Knipscheer, J., & van den Muijsenbergh (2012). Coping and
chronic psychosocial consequences of female genital mutilation in the Netherlands.
Ethnicity and Health, 17, 677-695. doi: 10.1080/13557858.2013.771148
Vromans, L., Schweitzer, R. D., & Brough, M. (2012). The Multidimensional Loss Scale:
Validating a cross-cultural instrument for measuring loss. Journal of Nervous and Mental
Disease, 200, 349-357. doi: 10.1097/NMD.0b013e31824cc458
Wagner, J., Burke, G., Kuoch, T., Scully, M., Armeli, S., & Rajan, T. V. (2013). Trauma,
healthcare access, and health outcomes among Southeast Asian refugees in Conneticut.
Journal of Immigrant and Minority Health, 15, 1065-1072. doi: 10.1007/s10903-0129715-2
Ward, C. L., Flisher, A. J., Chrisostomos, Z., Muller, M., & Lombard, C. J. (2004). Measuring
adolescents’ exposure to violence and related PTSD symptoms: Reliability of an
adaptation of the Harvard Trauma Questionnaire. Journal of Child and Adolescent
Mental Health, 16, 31-37. doi; 10.2989/17280580409486561
Welsh, E. A. (2012). Predictors of mental health and treatment engagement in torture survivors
(Unpublished doctoral dissertation). University of Maryland, College Park, MD.
Yeomans, P. D. (2008). The effect of PTSD psychoeducation on the nature and severity of
traumatic stress symptoms in a Burundian sample (Unpublished doctoral dissertation).
Drexel University, Philadelphia, PA.
207
Appendix D
Studies Included in the Validity Generalization of the HTQ (Study 2)
208
Studies Included in the Validity Generalization of the HTQ (Study 2)
Ahmad, A., Sofi, M.A., Sundelin-Wahlsten, V. & von Knorring, A.-L. (2000). Posttraumatic
stress disorder in children after the military operation “Anfal” in Iraqi Kurdistan.
European Child & Adolescent Psychiatry, 9, 235-243. doi: 10.1007/s007870070026
Aragona, M., Pucci, D., Mazzetti, M., Maisano, B., & Geraci, S. (2013). Traumatic events, postmigration living difficulties and post-traumatic symptoms in first generation immigrants:
A primary care study. Annali dell'Istituto Superiore di Sanità, 49, 169-175. doi:
10.4415/ANN_13_02_08
Bentley, J. A. (2011). Cross-cultural assessment of psychological symptoms among Somali
refugees (Unpublished doctoral dissertation). Seattle Pacific University, Seattle, WA.
Buhmann, C. B. (2014). Traumatized refugees: Morbidity, treatment and predictors of outcome.
Danish Medical Journal, 61(8), B4871.
Carswell, K., Blackburn, P., & Barker, C. (2011). Relationship between trauma, post-migration
problems and the psychological well-being of refugees and asylum seekers. International
Journal of Social Psychiatry, 57, 107-119. doi: 10.1177/0020764008105699
Chu, T., Keller, A. S., & Rasmussen, A. (2013). Effects of post-migration factors on PTSD
outcomes among immigrant survivors of political violence. Journal of Immigrant and
Minority Health, 15, 890-897. doi: 10.1007/s10903-012-9696-1
Culhane, M. A. (2009). Suffering strangers: The social construction of the torture survivor in the
American healing community (Unpublished doctoral dissertation). Boston University,
Boston, MA.
209
Dewaraja, R., & Kawamura, N. (2006). Trauma intensity and posttraumatic stress: Implications
of the tsunami experience in Sri Lanka for the management of future disasters.
International Congress Series, 1287, 69-73. doi: 10.1016/j.ics.2005.11.098
Domanskaité-Gota, V., Elklit, A., & Christiansen, D. M. (2009). Victimization and PTSD in a
Lithuanian national youth probability sample. Nordic Psychology, 61, 66-81. doi:
10.1027/1901-2276.61.3.66
Doran, J. M., Kalayjian, A., Toussaint, L., & DeMucci, J. (2012). The relationship between
trauma and forgiveness in post-conflict Sierra Leone. Psychological Trauma-Theory
Research Practice and Policy, 4, 614-623. doi: 10.1037/a0025470
Dworkin, J., Prescott, M., Jamal, R., Hardawan, S. A., Abdullah, A., & Galea, S. (2008). The
long-term psychosocial impact of a surprise chemical weapons attack on civilians in
Halabja, Iraqi Kurdistan. Journal of Nervous and Mental Disease, 196, 772-775. doi:
10.1097/NMD.0b013e3181878b69
Ekblad, S., Prochazka, H. & Roth, G. (2002). Psychological impact of torture: A 3 month
follow-up of mass-evacuated Kosovan adults in Sweden: Lessons learnt for prevention.
Acta Psychiatrica Scandinavica, 106, 30-36. doi: 10.1034/j.1600- 0447.106.s412.7.x
Elklit, A., Christiansen, D. M., Palic, S., Karsberg, S., & Eriksen, S. B. (2014). Impact of
traumatic events on posttraumatic stress disorder among Danish survivors of sexual abuse
in childhood. Journal of Child Sexual Abuse, 23, 918-934. doi:
10.1080/10538712.2014.964440
Elklit, A., & Kurdahl, S. (2013). The psychological reactions after witnessing a killing in public
in a Danish high school. European Journal of Psychotraumatology, 4. doi:
10.3402/ejpt.v4i0.19826
210
Elklit, A., Nielsen, L. H., Lasgaard, M., & Duch, C. (2013). A cartoon-based measure of PTSD
symptomatology in children exposed to a disaster. Journal of Loss & Trauma, 18, 54-63.
doi: 10.1080/15325024.2012.679125
Eng, S., Mulsow, M., Cleveland, H., & Hart, S. L. (2009). Academic achievement among
adolescents in Cambodia: Does caregiver trauma matter? Journal of Community
Psychology, 37, 754-768. doi: 10.1002/jcop.20329
Farhood, L. F., & Dimassi, H. (2012). Prevalence and predictors for post-traumatic stress
disorder, depression and general health in a population from six villages in South
Lebanon. Social Psychiatry and Psychiatric Epidemiology, 47, 639-649. doi:
10.1007/s00127-011-0368-6
Farhood, L., Dimassi, H., & Lehtinen, T. (2006). Exposure to war-related traumatic events,
prevalence of PTSD, and general psychiatric morbidity in a civilian population from
southern Lebanon. Journal of Transcultural Nursing, 17, 333-340. doi:
10.1177/1043659606291549
Fox, S. H., & Willis, M. S. (2009). Initiatory mental health assessments for Dinka and Nuer
refugees from Sudan. Journal of Immigrant and Refugee Studies, 7, 159-179. doi:
10.1080/15562940902935621
Galea, S., Rockers, P. C., Saydee, G., Macauley, R., Varpilah, S. T., & Kruk, M. E. (2010).
Persistent psychopathology in the wake of civil war: Long-term posttraumatic stress
disorder in Nimba County, Liberia. American Journal of Public Health, 100, 1745-1751.
doi: 10.2105/ajph.2009.179697
Geltman, P. L., Grant-Knight, W., Ellis, H., & Landgraf, J. M. (2008). The "Lost Boys" of
Sudan: Use of health services and functional health outcomes of unaccompanied refugee
211
minors resettled in the U. S. Journal of Immigrant and Minority Health, 10, 389-396. doi:
10.1007/s10903-007-9110-6
Gordon, J. S., Staples, J. K., He, D. Y., Atti, J. A. A. (2016). Mind–body skills groups for
posttraumatic stress disorder in Palestinian adults in Gaza. Traumatology, 22, 155-164.
doi: 10.1037/trm0000081
Grant-Knight, W., Geltman, P. L., & Ellis, H. B. (2009). Physical and mental health functioning
in Sudanese unaccompanied minors. In W. Grant-Knight, P. L. Geltman & H. B. Ellis
(Eds.), Treating traumatized children: Risk, resilience and recovery (pp. 102-116). New
York: Routledge.
Gupta, J., Falb, K. L., Carliner, H., Hossain, M., Kpebo, D., & Annan, J. (2014). Associations
between exposure to intimate partner violence, armed conflict, and probable PTSD
among women in rural Côte d'Ivoire. PLOS One, 9, e96300. doi:
10.1371/journal.pone.0096300
Hadley, C., Tegegn, A., Tessema, F., Cowan, J. A., Asefa, M., & Galea, S. (2008). Food
insecurity, stressful life events and symptoms of anxiety and depression in east Africa:
Evidence from the Gilgel Gibe growth and development study. Journal of Epidemiology
and Community Health, 62, 980-986. doi: 10.1136/jech.2007.068460
Halvorsen, J. O., & Kagee, A. (2010). Predictors of psychological sequelae of torture among
South African former political prisoners. Journal of Interpersonal Violence, 25, 9891005. doi: 10.1177/0886260509340547
Hooberman, J. B., Rosenfeld, B., Lhewa, D., Rasmussen, A., & Keller, A. (2007). Classifying
the torture experiences of refugees living in the United States. Journal of Interpersonal
Violence, 22, 108-123. doi: 10.1177/0886260506294999
212
Hossain, M., Zimmerman, C., Abas, M., Light, M., & Watts, C. (2010). The relationship of
trauma to mental disorders among trafficked and sexually exploited girls and women.
American Journal of Public Health, 100, 2442-2449. doi: 10.2105/AJPH.2009.173229
Huijts, I., Kleijn, W. C., van Emmerik, A. A. P., Noordhof, A., & Smith, A. J. M. (2012).
Dealing with man-made trauma: The relationship between coping style, posttraumatic
stress, and quality of life in resettled, traumatized refugees in the Netherlands. Journal of
Traumatic Stress, 25, 71-78. doi: 10.1002/jts.21665
Husain, F., Anderson, M., Cardozo, B. L., Becknell, K., Blanton, C., Araki, D., & Vithana, E. K.
(2011). Prevalence of war-related mental health conditions and association with
displacement status in postwar Jaffna District, Sri Lanka. Journal of the American
Medical Association, 306, 522-531. doi: 10.1001/jama.2011.1052
Hussain, D., & Bhushan, B. (2009). Development and validation of the Refugee Trauma
Experience Inventory. Psychological Trauma: Theory, Research, Practice and Policy, 1,
107-117. doi: 10.1037/a0016120
James, L. E., Noel, J. R., & Pierre, Y. M. (2014). A mixed-methods assessment of the
experiences of lay mental health workers in post-earthquake Haiti. American Journal of
Orthopsychiatry, 84, 152-163. doi: 10.1037/h0099387
Jones, L., & Kafetsios, K. (2002). Assessing adolescent mental health in war-affected societies:
The significance of symptoms. Child Abuse & Neglect, 26, 1059-1080. doi:
10.1016/S0145-2134(02)00381-2
Joska, J. A., Fincham, D. S., Stein, D. J., Paul, R. H., & Seedat, S. (2010). Clinical correlates of
HIV-associated neurocognitive disorders in South Africa. AIDS and Behavior, 14, 371378. doi: 10.1007/s10461-009-9538-x
213
Keller, A., Lhewa, D., Rosenfeld, B., Sachs, E., Aladjem, A., Cohen, I., . . . Portetfield, K.
(2006). Traumatic experiences and psychological distress in an urban refugee population
seeking treatment services. Journal of Nervous and Mental Disease, 194, 188-194. doi:
10.1097/01.nmd.0000202494.75723.83
Kumar, M. S., Murhekar, M. V., Hutin, Y., Subramanian, T., Ramachandran, V., & Gupte, M. D.
(2007). Prevalence of posttraumatic stress disorder in a coastal fishing village in Tamil
Nadu, India, after the December 2004 tsunami. American Journal of Public Health, 97,
99-101. doi: 10.2105/ajjph.2005.071167
Kun, P., Han, S., Chen, X., & Yao, L. (2009). Prevalence and risk factors for posttraumatic stress
disorder: A cross-sectional study among survivors of the Wenchuan 2008 earthquake in
China. Depression and Anxiety, 26, 1134-1140. doi: 10.1002/da.20612
Leaman, S. C. (2009). Risk factors for psychological distress and uses of religious coping
(Unpublished doctoral dissertation). The George Washington University, Washington,
DC.
Leaman, S. C., & Gee, C. B. (2012). Religious coping and risk factors for psychological distress
among African torture survivors. Psychological Trauma: Theory, Research, Practice,
and Policy, 4, 457-465. doi: 10.1037/a0026622
Lie, B. (2002). A 3-year follow-up study of psychosocial functioning and general symptoms in
settled refugees. Acta Psychiatrica Scandinavica, 106, 415-425. doi: 10.1034/j.16000447.2002.01436.x
Mai, L. G. (2010). Exploration of the intergenerational transmission of historical trauma in
Vietnamese Americans (Unpublished doctoral dissertation). California Institute of
Integral Studies, San Francisco, CA.
214
Matthey, S., Silove, D., Barnett, B., Fitzgerald, M.H. & Mitchell, P. (1999). Correlates of
depression and PTSD in Cambodian women with young children: A pilot study. Stress
Medicine, 15, 103-107. doi: 10.1002/(SICI)1099-1700(199904)15:2<103::AIDSMI791>3.0.CO;2-T
Meffert, S. M., Abdo, A. O., Abd Alla, O. A., Elmakki, Y. O. M., Omer, A. A., Yousif, S., …
Marmar, C. R. (2014). A pilot randomized controlled trial of interpersonal psychotherapy
for Sudanese refugees in Cairo, Egypt. Psychological Trauma: Theory, Research,
Practice, and Policy, 6, 240-249. doi: 10.1037/a0023540
Mghir, R. & Raskin, A. (1999). The psychological effects of the war in Afghanistan on young
Afghan refugees from different ethnic backgrounds. International Journal of Social
Psychiatry, 45, 29-36. doi: 10.1177/002076409904500104
Möhlen, H., Parzer, P., Resch, F., & Brunner, R. (2005). Psychosocial support for wartraumatized child and adolescent refugees: Evaluation of a short-term treatment program.
Australian and New Zealand Journal of Psychiatry, 39, 81-87. doi: 10.1080/j.14401614.2005.01513.x
Mollica, R. F., McInnes, K., Pham, T., Smith, M.C., Murphy, E. & Lin, L. (1998). The doseeffect relationships between torture and psychiatric symptoms in Vietnamese ex-political
detainees and a comparison group. Journal of Nervous and Mental Disease, 186, 543553. doi: 10.1097/00005053-199809000-00005
Mollica, R. F., McInnes, K., Poole, C. & Tor, S. (1998). Dose-effect relationships of trauma to
symptoms of depression and post-traumatic stress disorder among Cambodian survivors
of mass violence. British Journal of Psychiatry, 173, 482-488. doi:
10.1192/bjp.173.6.482
215
Morina, N., & Ford, J. D. (2008). Complex sequelae of psychological trauma among Kosovar
civilian war victims. International Journal of Social Psychiatry, 54, 425-436. doi:
10.1177/0020764008090505
Nicholson, B.L. (1997). The influence of pre-emigration and post-migration stressors on mental
health: A study of Southeast Asian refugees. Social Work Research, 21(1), 19-31.
O'Connor, M. (2010). PTSD in older bereaved people. Aging and Mental Health, 14, 670-678.
doi: 10.1080/13607860903311725
O'Connor, M. (2010). PTSD in the older bereaved people. Aging and Mental Health, 14, 310318. doi: 10.1080/13607860903228770
O'Connor, M., Nickerson, A., Aderka, I. M., & Bryant, R. A. (2015). The temporal relationship
between change in symptoms of prolonged grief and posttraumatic stress following old
age spousal bereavement. Depression and Anxiety, 32, 335-340. doi: 10.1002/da.22349
Oruc, L., Kapetanovic, A., Pojskic, N., Miley, K., Forstbauer, S., Mollica, R. F., & Henderson,
D. C. (2008). Screening for PTSD and depression in Bosnia and Herzegovina: Validating
the Harvard Trauma Questionnaire and the Hopkins Symptom Checklist. International
Journal of Culture and Mental Health, 1, 105-116. doi: 10.1080/17542860802456620
Ovuga, E., Oyok, T. O., & Moro, E. B. (2008). Posttraumatic stress disorder among former child
soldiers attending a rehabilitative service and primary school education in northern
Uganda. African Health Sciences, 8(3), 136-141.
Pham, K. T. (2013). Trauma transmission between first-generation and second-generation
Vietnamese refugees and immigrants in the United States: A study of mother and
daughter (Unpublished doctoral dissertation). Palo Alto University, Palo Alto, CA.
216
Phipps, R. M. (2014). Transgenerational transmission of trauma in second generation latino
children and adolescents (Unpublished doctoral dissertation). University of Mississippi,
Oxford, MS.
Prot, K. (2010). Late effects of trauma: PTSD in Holocaust survivors. Journal of Loss and
Trauma, 15, 28-42. doi: 10.1080/15325020902925506
Roth, G., Ekbad, S., & Agren, H. (2006). A longitudinal study of PTSD in a sample of adult
mass-evacuated Kosovars, some of whom returned to their home country. European
Psychiatry, 21, 152-159. doi: 10.1016/j.eurpsy.2005.11.004
Sabin, M., Sabin, K., Kim, H. Y., Vergara, M., & Varese, L. (2006). The mental health status of
Mayan refugees after repatriation to Guatemala. Pan American Journal of Public Health,
19, 163-171. doi: 10.1590/s1020-49892006000300004
Sachs, E. (2011). Assessing the impact of psychological distress on the daily functioning of
refugees: A high-risk symptom profile for disability among non-Western trauma survivors
(Unpublished doctoral dissertation). Fordham University, New York.
Sachs, E., Rosenfeld, B., Lhewa, D., Rasmussen, A., & Keller, A. (2008). Entering exile:
Trauma, mental health, and coping among Tibetan refugees arriving in Dharamsala,
India. Journal of Traumatic Stress, 21, 199-208. doi: 10.1002/jts.20324
Schweitzer, R. D., Brough, M., Vromans, L., Asic-Kobe, M. (2011). Mental health of newly
arrived Burmese refugees in Australia: Contributions of pre-migration and post-migration
experience. Australian and New Zealand Journal of Psychiatry, 45, 299-307. doi:
10.3109/00048674.2010.543412
Semb, O., Henningsson, M., Fransson, P., & Sundbom, E. (2009). Trauma-related symptoms
after violent crime: The role of risk factors before, during and eight months after
217
victimization. The Open Psychology Journal, 2, 77-88. doi:
10.2174/1874350100902010077
Sheikh, T. L., Abdulaziz, M., Agunbiade, S., Joseph, I., Ebiti, B., & Adekeye, O. (2015).
Correlates of depression among internally displaced persons after post-election violence
in Kaduna, North Western Nigeria. Journal of Affective Disorders, 170, 46-51. doi:
10.1016/j.jad.2014.08.050
Silove, D., Brooks, R., Steel Bateman, C., Steel, Z., Fonseca, Z., Amaral, C.,… Soosay, I.
(2010). Social and trauma-related pathways leading to psychological distress and
functional limitations four years after the humanitarian emergency in Timor Leste.
Journal of Traumatic Stress, 23, 151-160. doi: 10.1002/jts.20499
Silove, D., Steel, Z., McGorry, P., Miles, V. & Drobny, J. (2002). The impact of torture on posttraumatic stress symptoms in war-affected Tamil refugees and immigrants.
Comprehensive Psychiatry, 43, 49-55. doi: 10.1053/comp.2002.29843
Slewa-Younan, S., Chippendale, K., Heriseanu, A., Lujic, S., Atto, J., & Raphael, B. (2012).
Measures of psychophysiological arousal among resettled traumatized iraqi refugees
seeking psychological treatment. Journal of Traumatic Stress, 25, 348-352. doi:
10.1002/jts.21694
Steel, Z., & Silove, D. (2000). The psychosocial cost of seeking and granting asylum. In Z. Steel
& D. Silove (Eds.), International handbook of human response to trauma (pp. 421-438).
New York: Kluwer Academic/Plenum Publishers.
Suarez, E. B. (2011). Surviving the "Sasachacuy Tiempu "[difficult times]: The resilience of
Quechua women in the aftermath of the Peruvian armed conflict (Unpublished doctoral
dissertation). University of Toronto, Toronto, ON.
218
Suarez, E. B. (2013). Two decades later: The resilience and posttraumatic responses of
Indigenous Quechua girls and adolescents in the aftermath of the Peruvian armed
conflict. Child Abuse and Neglect, 37, 200-210. doi: 10.1016/j.chiabu.2012.09.011
Tay, A. K., Rees, S., Chen, J., Kareth, M., & Silove, D. (2015). The structure of post-traumatic
stress disorder and complex post-traumatic stress disorder amongst West Papuan
refugees. BMC Psychiatry, 15, 111. doi: 10.1186/s12888-015-0480-3
Toar, M., O'Brien, K. K., & Fahey, T. (2009). Comparison of self-reported health and healthcare
utilisation between asylum seekers and refugees: An observational study. BioMed
Central Public Health, 9. doi: 10.1186/1471-2458-9-214
Tomazinis, M. L. (2009). Sub-Saharan African refugees in Hong Kong: Pre- and post-migration
experiences, depression, anxiety and PTSD (Unpublished doctoral dissertation). Alliant
International University, Los Angeles, CA.
van Griensven, F., Chakkraband, M. L. S., Thienkrua, W., Pengjuntr, W., Cardozo, B. L.,
Tantipiwatanaskul, P., . . . Tappero, J. W. (2006). Mental health problems among adults
in tsunami-affected areas in southern Thailand. Journal of the American Medical
Association, 296, 537-548. doi: 10.1001/jama.296.5.537
van Wyk, A., Schweitzer, R. D., Brough, M., Vromans, L., & Murray, K. (2012). A longitudinal
study of mental health in refugees from Burma: The impact of therapeutic interventions.
Australian and New Zealand Journal of Psychiatry, 46, 995. doi:
10.1177/0004867412443059
Vromans, L., Schweitzer, R. D., & Brough, M. (2012). The Multidimensional Loss Scale:
Validating a cross-cultural instrument for measuring loss. Journal of Nervous and Mental
Disease, 200, 349-357. doi: 10.1097/NMD.0b013e31824cc458
219
Welsh, E. A. (2012). Predictors of mental health and treatment engagement in torture survivors
(Unpublished doctoral dissertation). University of Maryland, College Park, MD.
Wyshak, G. (1994). The relation between change in reports of traumatic events and symptoms of
psychiatric distress. General Hospital Psychiatry, 16, 290-297. doi: 10.1016/01638343(94)90009-4

Similar documents

×

Report this document