Psychological Trauma

your own opinion for each article and 2 questions you ask for each article

Journal of Traumatic Stress
October 2019, 32,

799

–805

B R I E F R E P O R T

An Empirical Crosswalk for the PTSD Checklist: Translating
DSM-IV to DSM-5 Using a Veteran Sample

Samantha J. Moshier,1,2 Daniel J. Lee,2,3 Michelle J. Bovin,2,3 Gabrielle Gauthier,1 Alexandra Zax,1

Raymond C. Rosen,4 Terence M. Keane,2,3 and Brian P. Marx2,3
1Veterans Affairs Boston Healthcare System, Boston, Massachusetts, USA

2The National Center for PTSD at Veterans Affairs Boston Healthcare System, Boston, Massachusetts, USA
3Department of Psychiatry, Boston University School of Medicine, Boston, Massachusetts, USA

4Healthcore/New England Research Institutes, Watertown, Massachusetts, USA

The fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) introduced numerous revisions to the fourth
edition’s (DSM-IV) criteria for posttraumatic stress disorder (PTSD), posing a challenge to clinicians and researchers who wish to assess
PTSD symptoms continuously over time. The aim of this study was to develop a crosswalk between the DSM-IV and DSM-5 versions of the
PTSD Checklist (PCL), a widely used self-rated measure of PTSD symptom severity. Participants were 1,003 U.S. veterans (58.7% with
PTSD) who completed the PCL for DSM-IV (the PCL-C) and DSM-5 (the PCL-5) during their participation in an ongoing longitudinal
registry study. In a randomly selected training sample (n = 800), we used equipercentile equating with loglinear smoothing to compute
a “crosswalk” between PCL-C and PCL-5 scores. We evaluated the correspondence between the crosswalk-determined predicted scores
and observed PCL-5 scores in the remaining validation sample (n = 203). The results showed strong correspondence between crosswalk-
predicted PCL-5 scores and observed PCL-5 scores in the validation sample, ICC = .96. Predicted PCL-5 scores performed comparably
to observed PCL-5 scores when examining their agreement with PTSD diagnosis ascertained by clinical interview: predicted PCL-5,
κ = 0.57; observed PCL-5, κ = 0.59. Subsample comparisons indicated that the crosswalk’s accuracy did not differ across characteristics
including gender, age, racial minority status, and PTSD status. The results support the validity of this newly developed PCL-C to PCL-5
crosswalk in a veteran sample, providing a tool with which to interpret and translate scores across the two measures.

The publication of the fifth edition of the Diagnostic and
Statistical Manual of Mental Disorders (DSM-5;American Psy-
chiatric Association [APA], 2013) introduced numerous revi-
sions to the diagnostic criteria for posttraumatic stress disorder
(PTSD), including the addition of new symptoms; the modi-
fication of several existing symptoms; and the introduction of
four, rather than three, symptom clusters. These changes to the
diagnostic criteria pose a challenge to clinicians and researchers

Samantha Moshier is now at Emmanuel College (Boston, MA, USA).

This research was funded by the U.S. Department of Defense, Congres-
sionally Directed Medical Research Programs (designations W81XWH08-
2-0100/W81XWH-08-2-0102 and W81XWH-12-2-0117/W81XWH12-2-
0121). Dr. Lee is supported by the National Institute of Mental Health
(5T32MH019836-16). Any opinions, findings, and conclusions or recommen-
dations expressed in this material are those of the authors and do not necessarily
reflect the view of the U.S. government.

Correspondence concerning this article should be addressed to Brian Marx,
Ph.D., 150 South Huntington Ave (116B-4), Boston, MA 02130, E-mail:
Brian.Marx@VA.gov

C© 2019 International Society for Traumatic Stress Studies. View this article
online at wileyonlinelibrary.com
DOI: 10.1002/jts.22438

who previously collected symptom data using measures reflect-
ing the PTSD diagnostic criteria in the prior version of the DSM
(i.e., the fourth edition, text revision; DSM-IV-TR; APA, 2000)
but who wish to follow the course of PTSD symptoms over time,
including after the revisions to the criteria were published. This
shift may be especially challenging to longitudinal investiga-
tions of PTSD, in which continuity of symptom measurement
over time is critical for many statistical analyses.

Clinicians and researchers with these continuity concerns
must choose among using symptom severity measures that cor-
respond with outdated PTSD diagnostic criteria; using mea-
sures that correspond with the updated DSM-5 PTSD diagnos-
tic criteria; or creating idiosyncratic, unvalidated measures that
simultaneously collect information about both sets of diagnos-
tic criteria. None of these choices is ideal. Instead, researchers
and clinicians would benefit from a guide that translates re-
sults of DSM-IV congruent measures to estimated results on
DSM-5 congruent measures, and vice versa. Recent research
has suggested that DSM-IV congruent symptom ratings can be
used to approximate a diagnosis of DSM-5 PTSD (Rosellini
et al., 2015). However, there is currently no tool available
to enable linking of continuous total or cluster-specific PTSD

799

http://crossmark.crossref.org/dialog/?doi=10.1002%2Fjts.22438&domain=pdf&date_stamp=2019-10-18

800 Moshier et al.

symptom severity scores derived from DSM-IV and DSM-5 con-
gruent measures. Therefore, the aim of the present study was to
establish a translational crosswalk between symptom severity
scores on the PTSD Checklist–Civilian Version for DSM-IV-
TR (PCL-C) and the PCL for DSM-5 (PCL-5; Weathers, Litz,
Herman, Huska, & Keane, 1993; Weathers et al., 2013), as the
PCL is the most commonly used self-rated measure of PTSD
symptom severity. To do so, we conducted test-equating proce-
dures using data from both versions of the measure collected
concurrently in a sample of U.S. military veterans.

Method

Participants

Participants were 1,003 United States Army or Marine vet-
erans enrolled in the Veterans After-Discharge Longitudinal
Registry (Project VALOR). Project VALOR is a registry of Vet-
erans’ Affairs (VA) mental health care users with and without
PTSD who were deployed in support of recent military oper-
ations in Afghanistan and Iraq. To be included in the cohort,
veterans must have undergone a mental health evaluation at a
VA facility. The cohort oversampled for veterans with proba-
ble PTSD according to VA medical records (i.e., at least two
instances of a PTSD diagnosis by a mental health professional
associated with two separate visits) at a 3:1 ratio. Female veter-
ans were oversampled at a rate of 1:1 (female to male). A sample
of 1,649 (60.8%) veterans completed the baseline assessment
for Project VALOR. For the current analysis, we focused on
a subsample of this group that consisted of 1,003 participants
who reported experiencing a DSM-5 Criterion A traumatic event
during a clinical interview and had complete data (required for
the test-equating analyses) on both the PCL-C and PCL-5 dur-
ing the fourth wave of study assessments (Time 4 [T4]). There
were no significant differences in sex, racial minority status, or
PTSD diagnostic status or symptom severity at the first wave of
data collection (Time 1 [T1]) between the 1,003 participants in-
cluded in this analysis and the remaining cohort members, ps =
.262–.891. However, participants included in this analysis were
older (M age = 38 years) compared with the remaining cohort
members (M age = 36 years), t(1,647) = −3.56, p = .000, and
had a higher level of educational attainment (i.e., 38% of the
analytic sample had a bachelor’s degree vs. 30% of remaining
cohort members), χ2(6, N = 1,642) = 15.74, p = .015.

Procedure

At T4 of Project VALOR, participants provided informed
consent verbally over the telephone in accordance with the re-
search protocol approved by the VA Boston Healthcare System
institutional review boards and the Human Research Protec-
tion Office of the U.S. Army Medical Research and Mate-
rial Command. Participants then completed a self-administered
questionnaire (SAQ) online and, following this, completed a
telephone-based diagnostic clinical interview. The SAQ con-
sisted of a large battery of questionnaires that, in total, included

over 740 questions pertaining to physical health, functional im-
pairment, psychiatric symptoms, deployment experiences, and
lifetime trauma exposure.

Measures

Demographic information. Participant age and sex were
extracted from a U.S. Department of Defense database. Race,
ethnicity, and education were collected via self-report in the T4
SAQ.

PTSD symptom severity. The PCL-C is a self-rated mea-
sure of PTSD symptom severity designed to correspond to the
17 core DSM-IV PTSD symptoms (Weathers et al., 1993). Re-
spondents use a scale ranging from 1 (not at all) to 5 (extremely)
to rate how much each symptom has bothered them in the past
month. Although a military version of the PCL (the PCL-M) is
available, we used the civilian version because it corresponded
with the study’s clinical interview procedures, which did not re-
strict potential index traumatic events solely to military-related
events. The PCL-C is one of the most commonly used self-
rated measures of DSM-IV PTSD symptom severity, and it
has demonstrated excellent psychometric properties across a
range of samples and settings (for review, see Norris & Ham-
blen, 2004). In the current sample, internal reliability of PCL-C
scores was excellent, Cronbach’s α = .96.

The PCL-5 (Weathers et al., 2013) is a self-rated measure
of PTSD symptom severity designed to correspond to the 20
core DSM-5 PTSD symptoms. Respondents use a scale ranging
from 0 (not at all) to 4 (extremely) to rate how much each symp-
tom has bothered them in the past month. Like its predecessor,
the PCL-5 is frequently used across a range of settings for
a variety of purposes, including monitoring symptom change
as well as screening for and providing a provisional diagno-
sis of PTSD. Data from the PCL-5 have demonstrated good
test–retest reliability, r = .84, and convergent and discriminant
validity (Blevins, Weathers, Davis, Witte, & Domino, 2015;
Bovin et al., 2015; Keane et al., 2014; Wortmann et al., 2016).
Internal reliability of PCL-5 scores was excellent in the current
sample, Cronbach’s α = .96.

Major depression and PTSD diagnosis. The PTSD and
Major Depressive Episode (MDE) modules of the Structured
Clinical Interview for DSM-5 (SCID-5; First, Williams, Karg,
& Spitzer, 2015) were used to assess exposure to a Criterion
A event and to assess current PTSD diagnostic status and pres-
ence or absence of a current MDE. Interrater agreement was
evaluated for a random sample of 100 cases and was excellent
for both current PTSD, κ = 0.85, and current MDE, κ = .98.
Data Analysis

To link PCL-C and PCL-5 scores, we used equipercentile
equating, a test-equating procedure that is commonly used
in educational measurement fields to determine comparable
scores on different versions of the same exam (for a review,
see Dorans, Moses, & Eigner, 2010). Equipercentile equating

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

A Crosswalk for the PTSD Checklist 801

Table 1
Demographic Characteristics of the Total, Test, and Validation Samples

Total Sample Test Sample Validation Sample
(n = 1,003) (n = 800) (n = 203)

Variable % M SD % M SD % M SD

Sex
Female 51.1 50.8 52.7
Male 48.9 49.2 47.3

Age (years) 43.2 9.8 43.2 9.9 43.2 9.5
Racial minority status

Non-White 23.2 22.6 25.2
White 76.8 77.4 74.8

Highest education level
High school or GED 6.6 6.9 5.4
Some college 38.9 39.7 37.1
Bachelor’s degree or higher 50.8 50.1 53.7

Current PTSD 58.7 58.9 58.1
Lifetime PTSD 87.5 87.4 88.2
Current MDE 34.5 34.4 35.0
PCL-C score 49.7 17.3 50.0 17.3 50.1 17.6
PCL-5 score 36.2 20.6 36.1 20.7 36.7 20.6

Note. PCL-C = Posttraumatic Stress Disorder Checklist–Civilian Version (for DSM-IV); PCL-5 = Posttraumatic Stress Disorder Checklist for DSM-5; PTSD =
posttraumatic stress disorder; MDE = major depressive episode; GED = general education development.

considers scores on two measures to be equivalent to one an-
other if their percentile ranks in a given group are equal. This
approach has a number of benefits relative to mean or linear
equating methods; for example, it results in all imputed scores
falling within the actual range of the scale and does not rely on
the assumption of a normal distribution of test scores. Equiper-
centile equating methods have been used to develop crosswalks
for a number of neurocognitive and psychiatric rating scales
(e.g., Choi, Schalet, Cook, & Cella, 2014; Monsell et al., 2016).

Figure 1. Histograms of total Posttraumatic Stress Disorder Checklist–Civilian
Version (for DSM-IV; PCL-C) and PCL for DSM-5 (PCL-5) scores in the
training sample (N = 800). DSM = Diagnostic and Statistical Manual of
Mental Disorders (DSM-IV = fourth edition; DSM-5 = fifth edition).

Prior to performing the equating procedure, we randomly
split the sample into a training sample (n = 800) and a vali-
dation sample (n = 203; a split which allows for a large sam-
ple size to be retained for the equating procedure, consistent
with recommendations by Dorans et al., 2010). In the training
dataset, equipercentile equating with loglinear smoothing was
performed using the R package Equate (Albano, 2016). Stan-
dard errors and 95% confidence intervals of the crosswalk esti-
mates were calculated using 10,000 bootstrapped samples. Af-
ter completing the equating procedure in the training dataset, we
used the resulting crosswalk to impute predicted PCL-5 scores
from PCL-C scores for all participants in the validation data set.

To evaluate the accuracy of the crosswalk in the valida-
tion sample, we examined the intraclass correlation coefficient
(ICC) between predicted and observed PCL-5 scores and cal-
culated the average difference between predicted and observed
PCL-5 scores. We calculated sensitivity, specificity, efficiency
(correct classification rate), quality of efficiency (i.e., Cohen’s
kappa), and area under the curve (AUC) for use of crosswalk-
predicted PCL-5 cut scores, using the cutoff of PCL-5 score
of 33 or greater (Bovin et al., 2015) in identifying PTSD
diagnosis as determined by the SCID interview. Finally, in
order to evaluate whether the crosswalk demonstrated accu-
racy across relevant subgroups of individuals, we compared
these same markers of accuracy when the sample was di-
vided into subgroups based on education level, age, gender,
racial minority status, and presence or absence of PTSD and
MDE.

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

802 Moshier et al.

Figure 2. Crosswalk of corresponding Posttraumatic Stress Disorder Checklist–
Civilian Version (for DSM-IV; PCL-C) and PCL for DSM-5 (PCL-5) total
scores with 95% confidence intervals from 10,000 bootstrapped samples.
DSM = Diagnostic and Statistical Manual of Mental Disorders (DSM-
IV = fourth edition; DSM-5 = fifth edition).

We used the same test-equating procedures to create cross-
walks from PCL-C subscale scores to PCL-5 subscale scores,
representing each of the DSM-5 PTSD symptom clusters (Clus-
ter B = intrusion symptoms, Cluster C = avoidance symptoms,
Cluster D = negative alterations in cognitions and mood, Clus-
ter E = alterations in arousal and reactivity). These symptom
clusters were approximated in the PCL-C data by summing
Items 1–5 (Cluster B), Items 6 and 7 (Cluster C), Items 8–12
(Cluster D), and Items 13–17 (Cluster E). Missing data were
minimal (one missing case each for variables of age, race, and
education status; and three cases missing the MDE module of
the SCID) and were therefore handled using pairwise deletion.

Results

The characteristics of the sample and subsample are pre-
sented in Table 1. In all, 58.7% percent of participants met cri-
teria for current (i.e., past month) PTSD and 34.5% met criteria
for current MDE. Group comparison tests revealed no signifi-
cant differences among the training and validation samples on
sex, race, ethnicity, education level, PCL-C or PCL-5 score,
or proportion of sample with current PTSD or MDE, ps =
.363–.878. The PCL-C and PCL-5 were highly correlated in

both the training and validation samples, rs = .95 and .96,
respectively. These correlations were well over thresholds rec-
ommended for equating procedures (i.e., .75–.86; Choi et al.,
2014). A histogram of total score frequencies in the training
sample is presented in Figure 1.

The crosswalk for converting PCL-C to PCL-5 scores based
on equipercentile equating results is presented in Figure 2. The
PCL-C scores were equated to lower PCL-5 scores, which is
not surprising given the difference in scaling ranges between
the two measures (PCL-C scores range from 17 to 85 and
PCL-5 scores range from 0 to 80). For example, a score of 50
on the PCL-C was equated with a score of 36 on the PCL-5.
In the validation sample, the ICC among the observed and
predicted PCL-5 scores was .96. The mean difference between
observed and predicted PCL-5 scores was 0.20 (SD = 6.30).
Using the cutoff score of 33 or higher, the predicted PCL-5
score had similar diagnostic utility to the observed PCL-5 score
in predicting PTSD diagnosis determined by clinical interview:
Cohen’s κ = .55, sensitivity = .81, specificity = .74, AUC
= .77, correct classification of 78% of cases for the predicted
PCL-5; Cohen’s κ = .58, sensitivity = .84, specificity = .74,
AUC = .79, correct classification of 80% of cases for the
observed PCL-5.

The accuracy of the crosswalk was highly consistent across
subgroups based on sex, age, racial minority status, education
level, PTSD diagnostic status as determined by clinical inter-
view, and presence or absence of current MDE (see Table 2).
The ICCs between predicted and observed PCL-5 scores were
very high for all subgroups, ICCs = .92–.96. There were no sig-
nificant differences in the mean differences between observed
and predicted PCL-5 score between any of these demographic
subgroups. The kappa values between observed and predicted
probable DSM-5 PTSD diagnosis were good for all subgroups
examined, and the proportion of correctly classified cases did
not differ significantly by subgroup.

The items comprising Clusters B and C are highly simi-
lar between the PCL-C and PCL-5, with only minor wording
changes (e.g., the addition of “unwanted” to Item 1 or the ad-
dition of “strong” to Item 5 on the PCL-5). Not surprisingly
then, the equipercentile-equated crosswalk for the Cluster C
subscale was identical to a linear transformation of subtracting
2 points from PCL-C scores to reflect the change in scaling be-
tween the two measures. Similarly, the equipercentile-equated
crosswalk for Cluster B subscale scores was nearly identical
to a linear transformation involving subtracting 5 points from
PCL-C scores. The ICC between equated and observed scores
using these two methods was equal to .997. Additionally, the
equipercentile-equated crosswalk for Cluster B did not outper-
form the linear transformation method in the accuracy analyses
conducted in the validation sample, which suggests that the lin-
ear transformation can be used for simplicity when converting
Cluster B subscale scores between the PCL-C and PCL-5. How-
ever, such a linear transformation would not be appropriate for
Clusters D and E given that both clusters include new symptoms
in DSM-5 relative to DSM-IV-TR. The crosswalks for Cluster D

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

A Crosswalk for the PTSD Checklist 803

Table 2
Posttraumatic Stress Disorder Checklist (PCL) Crosswalk Accuracy in Clinical and Demographic Subgroups Within the Validation
Sample

Crosswalk-
Predicted and

Observed PCL-5
Scores

Difference Between
Crosswalk-Predicted and Observed

PCL-5 Scoresa

Crosswalk-
Predicted and

Observed
Probable PTSDb

Variable n ICC M SD κ

Sex
Male 96 .95 −0.29 6.72 0.91
Female 107 .96 −0.08 5.84 0.85

Age (years)
< 40 116 .95 0.30 6.55 0.83 � 40 86 .96 −0.84 5.85 0.95

Racial minority status
Non-White 51 .96 0.75 6.23 0.88
White 151 .95 −0.49 6.27 0.88

Education level
High school or some college 93 .94 −0.65 6.71 0.90
Bachelor’s degree or higher 109 .96 0.22 5.83 0.85

PTSD diagnosis
Present 118 .92 −0.28 6.33 0.85
Absent 85 .93 −0.16 6.31 0.78

Current MDE
Present 70 .93 −0.90 6.29 0.83
Absent 130 .95 0.25 6.21 0.86

Note. n = 203. ICC = intraclass correlation coefficient; PCL-C = Posttraumatic Stress Disorder Checklist–Civilian Version (for DSM-IV); PCL-5 = Posttraumatic
Stress Disorder Checklist for DSM-5; PTSD = posttraumatic stress disorder; MDE = major depressive episode.
aIn t tests between all subgroups, ps = .190–.812. bProbable PTSD defined as a PCL-5 score � 33.

Figure 3. Crosswalk of corresponding Posttraumatic Stress Disorder Checklist–
Civilian Version (for DSM-IV; PCL-C) and PCL for DSM-5 (PCL-5) Clusters D
and E scores with 95% confidence intervals from 10,000 bootstrapped samples.
Approximated PCL-C scores for Clusters D and E were computed by summing
Items 8–12 (Cluster D) and Items 13–17 (Cluster E) of the PCL-C. DSM =
Diagnostic and Statistical Manual of Mental Disorders (DSM-IV = fourth
edition; DSM-5 = fifth edition).

and E subscores based on equipercentile equating with loglin-
ear presmoothing are presented in Figure 3. Predicted cluster
subscores were very strongly correlated with observed cluster
subscores in the validation sample for all four clusters; the ICC
values between observed and predicted subscale scores were
.94 for Cluster B, .88 for Cluster C, .89 for Cluster D, and .91
for Cluster E.

Discussion

This is the first known study that attempted to equate scores
between two versions of a frequently used PTSD symptom
severity measure: the DSM-IV-based PCL-C and the DSM-
5-based PCL-5. The resulting crosswalk enables researchers
and clinicians to interpret and translate scores across the two
measures, an important consideration in longitudinal obser-
vational and clinical treatment studies that cross iterations of
the DSM. A particular strength of this study was the use of
both training and validation samples, which allowed us to
evaluate the accuracy of the crosswalk. Supporting the valid-
ity of the crosswalk, results demonstrated a strong degree of

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

804 Moshier et al.

concordance between observed and predicted PCL-5 scores
(both total and cluster subscale scores) in the validation sam-
ple. Additionally, predicted PCL-5 scores performed compara-
bly to observed PCL-5 scores when examining their agreement
with PTSD diagnosis ascertained by clinical interview. Finally,
the results suggest a similar degree of concordance between
crosswalk–predicted and observed subscale scores and indicate
that the metrics of crosswalk accuracy did not differ across
subgroups.

We anticipate that the PCL crosswalk may be particularly
useful for longitudinal research or for interpretation of clinical
data that have been collected over a time period spanning the use
of both the DSM-IV and DSM-5. It may also allow for the com-
bining of data sets from studies using different versions of the
PCL, facilitating research that requires large sample sizes, such
as gene association studies. Moreover, the availability of cross-
walks for computing DSM-5 symptom cluster subscale scores
will allow for further study of the association between specific
domains of symptoms (e.g., avoidance, arousal) and risk factors
or outcomes of interest. However, it should be noted that the
evolution of the diagnostic criteria from DSM-IV to DSM-5 has
led to some substantive differences in how the PTSD construct
is defined in each version. The strong correlation among PCL-C
and PCL-5 scores (r = .95) suggests that it was statistically ap-
propriate to use test-equating procedures to link the scales. This
strong association has been demonstrated in prior studies of the
PCL-5 (e.g.,Wortmann et al., 2016) and is consistent with other
research suggesting a strong degree of overlap between the
two DSM criteria sets (e.g., Kilpatrick et al., 2013). However,
it should also be acknowledged that the resulting crosswalk
cannot provide specific information about the elements of the
PTSD construct that are new to DSM-5 and were not assessed
in DSM-IV (i.e., distorted blame, reckless behavior), and it also
does not address differences in the definition of a Criterion A
traumatic event.

This study has a number of strengths for a test-equating de-
sign. We used a single-group design in which all participants
completed both versions of the PCL, thus producing more reli-
able linking across measures. The sample was large and gender-
balanced, and participants showed a wide degree of variation in
PTSD symptom severity. However, the sample consisted solely
of veterans serving in recent-era (i.e., after the September 11,
2001, terrorist attacks) combat operations in Afghanistan and
Iraq. Although the crosswalk showed invariance to several de-
mographic characteristics within the sample, it is not clear to
what extent the results would generalize to civilian samples.
We suggest caution in applying the crosswalk to these sam-
ples and encourage continued study of these results in other
trauma-exposed samples. Additionally, it should be noted that
the PCL-C and PCL-5 were administered in the same order for
every participant, with the PCL-C administered first. Therefore,
order effects may have influenced our results, and future re-
search should examine this possibility, using a counter-balanced
design.

In this study, we present a crosswalk that will allow for con-
version between PCL-C and PCL-5 symptom severity scores.
The results provide support for the validity of the crosswalk
within a veteran sample. This tool will allow researchers and
clinicians to make use of archival PCL-C data in longitudinal
research, clinical settings, and beyond.

References
Albano, A. D. (2016). Equate: An R package for observed-score

linking and equating. Journal of Statistical Software, 74, 1–36.
https://doi.org/10.18637/jss.v074.i08

American Psychiatric Association. (2000). Diagnostic and statistical manual
of mental disorders (4th ed., text revision). Washington, DC: Author.

American Psychiatric Association. (2013). Diagnostic and statistical manual of
mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing.

Blevins, C. A., Weathers, F. W., Davis, M. T., Witte, T. K., & Domino,
J. L. (2015). The Posttraumatic Stress Disorder Checklist for DSM-5
(PCL-5): Development and initial psychometric evaluation. Journal of Trau-
matic Stress, 28, 489–498. https://doi.org/10.1002/jts.22059

Bovin, M. J., Marx, B. P., Weathers, F. W., Gallagher, M. W., Rodriguez, P.,
Schnurr, P. P., & Keane, T. M. (2015). 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.
https://doi.org/10.1037/pas0000254

Choi, S. W., Schalet, B., Cook, K. F., & Cella, D. (2014). Establishing a
common metric for depressive symptoms: Linking the BDI-II, CES-D, and
PHQ-9 to PROMIS depression. Psychological Assessment, 26, 513–527.
https://doi.org/10.1037/a0035768

Dorans, N. J., Moses, T., & Eignor, D. E. (2010). Principles and practices of
test score equating (ETS Research Report No. RR-10-29). Princeton, NJ:
Educational Testing Service.

First, M. B., Williams, J. W., Karg, R. S., & Spitzer, R. L. (2015). Structured
Clinical Interview for DSM-5–Research Version. Arlington, VA: American
Psychiatric Association.

Keane, T. M., Rubin, A., Lachowicz, M., Brief, D. J., Enggasser, J., Roy, M., . . .
Rosenbloom, D. (2014). Temporal Stability of DSM-5 posttraumatic stress
disorder criteria in a problem drinking sample. Psychological Assessment,
26, 1138–1145. https://doi.org/10.1037/a0037133

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. https://doi.org/10.1002/da.22364

Monsell, S. E., Dodge, H. H., Zhou, X. H., Bu, Y., Besser, L. M., Mock,
C., . . . Weintraub, S. (2016). Results from the NACC uniform data set
neuropsychological battery crosswalk. Alzheimer Disease and Associated
Disorders, 30, 134–139. https://doi.org/10.1097/WAD.0000000000000111

Norris, F. H., & Hamblen, J. L. (2004). Standardized self-report measures of
civilian trauma and PTSD. In J. P. Wilson, T. M. Keane, & T. Martin (Eds.),
Assessing psychological trauma and PTSD (pp. 63–102). New York, NY:
Guilford Press.

Rosellini, A. J., Stein, M. B., Colpe, L. J., Heeringa, S. G., Petukhova, M.
V., Sampson, N. A., . . . & Army STARRS Collaborators. (2015). Approxi-
mating a DSM-5 diagnosis of PTSD using DSM-IV criteria. Depression and
Anxiety, 32, 493–501. https://doi.org/10.1002/da.22364

Weathers, F., Litz, B., Herman, D., Huska, J., & Keane, T. (1993, October).
The PTSD Checklist (PCL): Reliability, Validity, and Diagnostic Utility.

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

https://doi.org/10.18637/jss.v074.i08

https://doi.org/10.1002/jts.22059

https://doi.org/10.1037/pas0000254

https://doi.org/10.1037/a0035768

https://doi.org/10.1037/a0037133

https://doi.org/10.1002/da.22364

https://doi.org/10.1097/WAD.0000000000000111

https://doi.org/10.1002/da.22364

A Crosswalk for the PTSD Checklist 805

Paper presented at the Annual Convention of the International Society for
Traumatic Stress Studies, San Antonio, TX.

Weathers, F. W., Litz, B. T., Keane, T. M., Palmieri, P. A., Marx, B. P., &
Schnurr, P. P. (2013). The PTSD Checklist for DSM-5 (PCL-5). Scale avail-
able from the National Center for PTSD at www.ptsd.va.gov

Wortmann, J. H., Jordan, A. H., Weathers, F. W., Resick, P. A., Don-
danville, K. A., Hall-Clark, B., . . . Litz, B. T. (2016). Psychomet-
ric analysis of the PTSD Checklist-5 (PCL-5) among treatment-seeking
military service members. Psychological Assessment, 28, 1392–1403.
https://doi.org/10.1037/pas0000260

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

http://www.ptsd.va.gov

https://doi.org/10.1037/pas0000260

Journal of Traumatic Stress
October 2015, 28,

480

–483

B R I E F R E P O R T

Comparison of the PTSD Checklist (PCL) Administered via
a Mobile Device Relative to a Paper Form

Matthew Price,1 Eric Kuhn,2 Julia E. Hoffman,2,3 Josef Ruzek,2 and Ron Acierno4,5
1Department of Psychological Science, University of Vermont, Burlington, Vermont, USA

2National Center for PTSD, Dissemination and Training Division, Department of Veterans Affairs Palo Alto Health Care System,
Palo Alto, California, USA

3Center for Healthcare Evaluation, Department of Veterans Affairs Palo Alto Healthcare System, Palo Alto, California, USA
4Ralph H. Johnson Veterans Affairs Medical Center, Charleston, South Carolina

5Medical University of South Carolina, Charleston, South Carolina, USA

Mobile devices are increasingly used to administer self-report measures of mental health symptoms. There are significant differences,
however, in the way that information is presented on mobile devices compared to the traditional paper forms that were used to administer
such measures. Such differences may systematically alter responses. The present study evaluated if and how responses differed for a
self-report measure, the PTSD Checklist (PCL), administered via mobile device relative to paper and pencil. Participants were 153 trauma-
exposed individuals who completed counterbalanced administrations of the PCL on a mobile device and on paper. PCL total scores (d =
0.07) and item responses did not meaningfully or significantly differ across administrations. Power was sufficient to detect a difference in
total score between administrations determined by prior work of 3.46 with a d = 0.23. The magnitude of differences between administration
formats was unrelated to prior use of mobile devices or participant age. These findings suggest that responses to self-report measures
administered via mobile device are equivalent to those obtained via paper and they can be used with experienced as well as naı̈ve users of
mobile devices.

Mobile devices can advance traumatic stress research and
treatment (Luxton et al., 2011; Price et al., 2014) through the
collection of ecologically valid data (Shiffman, Stone, & Huf-
ford, 2008). Use of mobile devices requires that responses to
mobile-administered measures are equivalent to responses from
paper measures. This assumption is open to empirical investiga-
tion and should be evaluated to ensure mobile devices provide
valid and reliable measurements.

Mobile devices systematically change the administration of
self-report measures. When delivered via paper, items are dis-
played in an array that allows all responses to be viewed simul-
taneously such that initial responses may influence subsequent
answers (Richman, Kiesler, Weisband, & Drasgow, 1999). Al-
ternatively, mobile devices typically display a single item per
screen. Administration of individual items may focus atten-
tion towards item content resulting in systemically different
responses.

The present study examined if responses to a self-report mea-
sure, the PTSD Checklist (PCL; Weathers et al., 2013), admin-

Copyright C© 2015 Wiley Periodicals, Inc., A Wiley Company. View this article
online at wileyonlinelibrary.com
DOI: 10.1002/jts.22037

istrated via mobile device differed from paper administration.
The PCL has been extensively validated as a measure of PTSD
symptoms across diverse samples (Ruggiero, Ben, Scotti, & Ra-
balais, 2003). A standardized paper version of the PCL is avail-
able via request from the National Center for PTSD (NCPTSD).
The PCL is available in a standardized format for mobile de-
vices as part of the PE Coach mobile application (Reger et al.,
2013). It was hypothesized that responses between PCL total
score and item responses across mobile and paper administra-
tions would be comparable due to prior evidence that suggested
minimal differences between standardized tests administered
via paper and computer (Bush et al., 2013; Campbell et al.,
1999; Finger & Ones, 1999).

Method

Participants

Participants, aged M = 32.34 years (SD = 14.42), were 153
individuals recruited from a Level 1 trauma center (n = 22,
14.3%), a Veteran’s Affairs medical center outpatient mental
health service (VAMC; n = 38, 24.7%), an outpatient clinic

480

Mobile Comparison of PCL 481

Table 1
Descriptive Information Unadjusted for Time Between PCL
Administrations

Variable n %

Location
Veteran Affairs Medical Center 38 24.7
Female 8 21.1
Community 87 57.1
Female 65 73.9
Outpatient clinic 6 3.9
Female 6 100.0
Trauma center 22 14.3
Female 6 27.3

PTSD diagnosis 62 40.3
Own smartphone 118 76.6

Use e-mail on phone 117 76.0
Use apps on phone 113 73.4
Use games on phone 97 63.0
Use Internet on phone 122 79.2

Note. N = 153. PCL = Posttraumatic Stress Checklist.

for trauma victims (n = 6, 3.9%), and the community (n = 87,
57.1%). Descriptive information is presented in Table 1.

Measures and Procedure

The Posttraumatic Checklist-Civilian Version (PCL-C: Weath-
ers, Litz, Huska, & Keane, 1994) is a 17-item self-report mea-
sure that assesses PTSD symptom severity. Symptoms are rated
on a 5-point Likert-type scale, ranging from 1 = not at all to
5 = extremely, for the past month. Internal consistency for
the current study was excellent with α = .95 for both ad-
ministrations. The measure was administered twice, once via
the paper form available from the NCPTSD (Weathers, Litz,
Huska, & Keane, 2003) and once via PE Coach. The Life
Events Checklist (LEC; Weathers et al., 2013) is a 17-item
self-report measure assessing trauma exposure. Use of Inter-
net and mobile devices was assessed with questions adapted
from a survey from the Pew Internet and American Life Project
(2012). Questions assessed if various tasks were completed reg-
ularly completed on smartphones and mobile devices using a
yes/no format (e.g., “Do you regularly check e-mail on your
smartphone?”).

Medical records were used to confirm trauma exposure for
Level-1 trauma, VAMC, and outpatient clinic participants. A
diagnosis of PTSD was the indicator of trauma exposure for
VAMC and outpatient clinic participants whereas the presenting
trauma was used for Level-1 trauma center participants. Com-
munity participants were screened with the LEC to determine
if they experienced or witnessed a traumatic event. Follow-
up questions confirmed the validity of the Criterion A event.
The community sample was administered the PTSD module of
the Structured Clinical Interview for the DSM-IV by trained

research staff for the most stressful event identified by the LEC
(SCID; First, Spitzer, Gibbon, & Williams, 2002). No other
modules of the SCID were administered.

Participants completed the PCL on an iPod Touch (4th gen-
eration, 3.5′′ screen) and on paper with a 35-minute (Med =
35, interquartile range: 25) interval between administrations.
After the second administration, participants completed the
use of Internet and mobile devices survey, and demographics
questionnaire. Participants from the community were also given
the PTSD module from the SCID and 27% met criteria for
PTSD. Interviews were administered by trained research assis-
tants and audio recorded. Interviews were double coded from
the recording by a clinical psychologist with 100% diagnos-
tic agreement. The order in which mobile and paper versions
were administered was counterbalanced using a randomization
sequence. Randomization occurred in blocks of 10 and each
data collection site was allocated 10 blocks. Institutional re-
view boards of the agencies where this research was conducted
approved all procedures and all participants consented to the
study.

Data Analysis

Using the guidelines of Bland and Altman (1986), a clini-
cally meaningful margin of error between the two methods
of measurement of 3.46 was established (see Supplemental
Table 1) from nine prior studies where the PCL was adminis-
tered repeatedly. A difference score between the total scores
for both administrations was obtained by subtracting mobile
device scores from paper scores. Comparisons were made with
repeated-measures analysis of covariance in which length be-
tween administrations was used as a covariate. The mean of
the distribution of difference scores was calculated with the
95% confidence interval (CI). If the 95% CI of the difference
scores was within the clinically meaningful margin of error
then the two methods were considered interchangeable. A mar-
gin of error of 1.00 was used for differences between indi-
vidual items. Bivariate correlations between both measure ad-
ministrations and intraclass correlation coefficients (ICC) were
also computed. One participant declined to answer questions
about use of a mobile devices after reporting they did not
own a smartphone. There were no missing data on the PCL
administrations.

Results

Adjusted for time between administrations, the mean difference
between paper (M = 40.24, SD = 16.69) and mobile device (M
= 39.08, SD = 15.97) administration was 1.17 points with 95%
CI [1.13, 1.21] (Table 2). The upper limit of the 95% CI for the
mean difference was within the margin of error. The effect size
for the difference was d = 0.07. Test-retest reliability was r =
.93. The ICC was .96, 95% CI [.95, .97]. Mean differences at the
item level ranged from 0.001 to 0.22. The highest upper limit
for the 95% CI at the item level was 0.37 for Item 8. Therefore,

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

482 Price et al.

Table 2
Mean Difference and 95% CI for PCL Items and Total Score

PCL M Diff 95% CI

Item
1. Intrusive thoughts 0.02 [−0.11, 0.15]
2. Nightmares 0.05 [−0.07, 0.18]
3. Reliving 0.13 [−0.01, 0.27]
4. Emotional cue reactivity 0.12 [−0.04, 0.28]
5. Physiological cue reactivity 0.14 [0.00, 0.27]
6. Avoidance of thoughts 0.04 [−0.14, 0.22]
7. Avoidance of reminders 0.13 [−0.04, 0.29]
8. Trauma-related amnesia 0.22 [0.08, 0.37]
9. Loss of interest 0.05 [−0.08, 0.17]
10. Feeling detached −0.09 [−0.22, 0.05]
11. Lack of positive emotion 0.09 [−0.03, 0.22]
12. Foreshortened future 0.03 [−0.11, 0.17]
13. Sleep problems −0.01 [−0.13, 0.12]
14. Irritability or anger 0.07 [−0.05, 0.20]
15. Difficulty of concentrating −0.04 [−0.18, 0.10]
16. Overly alert 0.04 [−0.08, 0.16]
17. Easily startled 0.14 [0.02, 0.26]
Total 1.17 [1.13, 1.21]

Note. Sample size = 153. Margin of error for Total scale = 3.46. Margin of error
for items = 1.00. Difference score calculated as paper minus mobile. PCL =
Posttraumatic Stress Checklist.

all of the items were within the margin of error (1.00). Test-
retest reliability at the item level ranged from r = .66 to .88 and
ICC = .75 to .93.

There were no differences in administrations across the dif-
ferent locations, F(3, 149) = 1.05, p = .373. Results were
consistent across the combined sample in that the upper limit
of the 95% CI for the sample obtained from the trauma cen-
ter, M = 0.45, 95% CI [0.45, 0.45]; VAMC, M = 2.72, 95%
CI [2.60, 2.85]; and community sample, M = 0.65, 95% CI
[0.58, 0.72] were within the margin of error for the total scale.
Test-retest reliability within each group was consistent with the
total sample: trauma center, r = .89, ICC = .94, 95% CI [.86,
.98]; VAMC, r = .89, ICC = .94, 95% CI [.89, .97]; com-
munity sample, r = .91, ICC = .95, 95% CI [.93, .97]. Mean
differences at the item level ranged from 0.00 to 0.36 for the
trauma center, from 0.00 to 0.37 for the VAMC, and from 0.01
to 0.20 for the community sample. The highest upper limit for
the 95% CI for each item was within the margin of error for the
trauma center (0.65), VAMC (0.65), and the community sample
(0.40).

The relation between use of smartphone functions and dif-
ference in total PCL scores across the administrations was as-
sessed with one-way analyses of variance. Differences in total
scores were not related to smartphone ownership, F(1, 149)
= 1.51, p = .221; use of e-mail via smartphone, F(1, 148) =
0.60, p = .439); use of apps, F(1, 147) = 0.78, p = .378);
use of games, F(1, 148) = 0.78, p = .379; and use of the In-

ternet on a smartphone, F(1, 148) = 0.78, p = .379. Finally,
differences in total PCL scores were unrelated to age (r = .04,
p = .598).

Discussion

The present study suggested that there were minimal dif-
ferences between a self-report measure of PTSD symptoms
administered via mobile device or paper in a heterogeneous
sample of trauma-exposed adults. The lack of a relation be-
tween prior experiences using a mobile device, age, and differ-
ences in total score indicates that mobile devices are a viable
strategy for those who have minimal training or experience with
this technology. Prior work demonstrated that among patients,
demographic characteristics and prior experience is largely un-
related to willingness to use technology for healthcare (Price
et al., 2013). There is evidence, however, to suggest that prior
use is relevant for clinicians (Kuhn et al., 2014). Clinicians
with experience using mobile devices or who own a personal
mobile device were more receptive to use such technologies
in treatment. Ensuring that clinicians are capable and comfort-
able with such devices will be necessary for proper measure
administration as patients are likely to turn to their therapist for
technical assistance or tutorials with these technologies (Price &
Gros, 2014).

The present study had several limitations. The mobile ad-
ministration was not conducted in a naturalistic environment
where such measures administered via mobile device are most
likely to be completed insofar as this was a research study
with informed consent processes. The effect of environmental
influences on responses is unknown. Although it is unlikely
that the environment would systematically influence mobile re-
sponses relative to paper response, measures completed on a
mobile device are more likely to be completed in a variety of
contexts in which other factors could influence responses. Re-
searchers are advised to collect data on the context in which
measures are completed to assess potential sources of bias. The
study evaluated a single self-report measure of PTSD without a
lengthy assessment battery. Thus, the current study was unable
to examine effects related to fatigue across the administration
of multiple measures via a mobile device. The current study
supported the null hypothesis that there were no differences
between scores across paper and mobile versions of the PCL,
which is conceptually and pragmatically challenging (Piaggo,
Elbourne, Pocock, & Evans, 2006). Although the current study
had sufficient power to detect an effect as small as 0.23, con-
siderably more power would be needed to detect an effect at
the obtained effect size of 0.07 (n = 1,604). Continued studies
that demonstrate the clinical equivalence of measurements ob-
tained via mobile device relative to paper should be conducted
to further validate these findings. Finally, PTSD diagnoses were
obtained with different methods across the subsamples, and the
accuracy of diagnoses in medical records has been questioned
(Holowka et al., 2014).

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

Mobile Comparison of PCL 483

The current study provides empirical support regarding the
lack of differences for measures administered via mobile de-
vice. Given the high rates of smartphone ownership, the results
from the present study suggest that mobile devices are an appro-
priate method for population screens of PTSD. Such a method
would assist in the efficient allocation of resources in events of
mass trauma such as a natural disaster.

References
Bland, M. J., & Altman, D. G. (1986). Statistical methods for assessing agree-

ment between two methods of clinical assessment. The Lancet, 327, 307–
310. doi:10.1016/S0140-6736(86)90837-8

Bush, N. E., Skopp, N., Smolenski, D., Crumpton, R., & Fairall, J. (2013).
Behavioral screening measures delivered with a smartphone app: psycho-
metric properties and user preference. The Journal of Nervous and Mental
Disease, 201, 991–995. doi:10.1097/NMD.0000000000000039

Campbell, K. A., Rohlman, D. S., Storzbach, D., Binder, L. M., Anger, W.
K., Kovera, C. A., . . . Grossmann, S. J. (1999). Test-retest reliability of
psychological and neurobehavioral tests self-administered by computer. As-
sessment, 6, 21–32. doi:10.1177/107319119900600103

Finger, M. S., & Ones, D. S. (1999). Psychometric equivalence of the computer
and booklet forms of the MMPI: A meta-analysis. Psychological Assessment,
11, 58–66. doi:10.1037/1040-3590.11.1.58

First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (2002). Struc-
tured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version,
Patient Edition. N ew York, NY: Biometrics Research, New York State
Psychiatric Institute.

Holowka, D. W., Marx, B. P., Gates, M. A., Litman, H. J., Ranganathan, G.,
Rosen, R. C., & Keane, T. M. (2014). PTSD diagnostic validity in Vet-
erans Affairs electronic records of Iraq and Afghanistan veterans. Jour-
nal of Consulting and Clinical Psychology, 82, 569–579. doi:10.1037/
a0036347

Kuhn, E., Eftekhari, A., Hoffman, J. E., Crowley, J. J., Ramsey, K. M., Reger,
G. M., & Ruzek, J. I. (2014). Clinician perceptions of using a smartphone
app with prolonged exposure therapy. Administration and Policy in Mental
Health and Mental Health Services Research, 1–8. doi:10.1007/s10488-013-
0532-2

Luxton, D. D., McCann, R. A., Bush, N. E., Mishkind, M. C., & Reger, G.
M. (2011). mHealth for mental health: Integrating smartphone technology
in behavioral healthcare. Professional Psychology: Research and Practice,
42, 505–512. doi:10.1037/a0024485

Pew Internet and American Life Project. (2012, September). Explore Survey
Questions. Retrieved from http://www.pewinternet.org/Static-Pages/Data-

Tools/Explore-Survey-Questions/Roper-Center.aspx?item={0368CEFB-
1706-4995-B395-925639C0B22F}

Piaggo, G., Elbourne, D. R., Pocock, S. J., & Evans, S. J. W. (2006). Reporting
of noninferority and equivalence randomized trials: An extension of the
CONSORT statement. Journal of the American Medical Association, 295,
1152–1161. doi:10.1001/jama.295.10.1152

Price, M., & Gros, D. F. (2014). Examination of prior experience with telehealth
and comfort with telehealth technology as a moderator of treatment response
for PTSD and depression with veterans. International Journal of Psychiatry
in Medicine, 48, 57–67. doi:10.2190/PM.48.1.e

Price, M., Williamson, D., McCandless, R., Mueller, M., Gregoski, M.,
Brunner-Jackson, B., . . . Treiber, F. (2013). Hispanic migrant farm work-
ers’ attitudes toward mobile phone-based telehealth for management of
chronic health conditions. Journal of Medical Internet Research, 15, e76.
doi:10.2196/jmir.2500

Price, M., Yuen, E. K., Goetter, E. M., Herbert, J. D., Forman, E. M., Acierno,
R., & Ruggiero, K. J. (2014). mHealth: A mechanism to deliver more acces-
sible, more effective mental health care. Clinical Psychology & Psychother-
apy, 21, 427–436. doi:10.1002/cpp.1855

Reger, G. M., Hoffman, J., Riggs, D., Rothbaum, B. O., Ruzek, J., Holloway,
K. M., & Kuhn, E. (2013). The “PE coach” smartphone application: An
innovative approach to improving implementation, fidelity, and homework
adherence during prolonged exposure. Psychological Services, 10, 342–349.
doi:10.1037/a0032774

Richman, W. L., Kiesler, S., Weisband, S., & Drasgow, F. (1999). A meta-
analytic study of social desirability distortion in computer-administered
questionnaires, traditional questionnaires, and interviews. Journal of Ap-
plied Psychology, 84, 754–775. doi:10.1037/0021-9010.84.5.754

Ruggiero, K. J., Ben, K. D., Scotti, J. R., & Rabalais, A. E. (2003). Psy-
chometric properties of the PTSD Checklist—Civilian version. Journal of
Traumatic Stress, 16, 495–502. doi:10.1023/A:1025714729117

Shiffman, S., Stone, A. A., & Hufford, M. R. (2008). Ecological mo-
mentary assessment. Annual Review of Clinical Psychology, 4, 1–32.
doi:10.1146/annurev.clinpsy.3.022806.091415

Weathers, F., Litz, B., Huska, J., & Keane, T. (1994). Post-Traumatic Stress
Disorder Checklist (PCL-C) for DSM-IV. Boston, MA: National Center for
PTSD.

Weathers, F., Litz, B., Huska, J., & Keane, T. (2003). PTSD Check-
list – Civilian Version. Retrieved from http://www.mirecc.va.gov/docs/
visn6/3_PTSD_CheckList_and_Scoring

Weathers, F. W., Blake, D. D., Schnurr, P. P., Kaloupek, D. G., Marx, B. P.,
& Keane, T. M. (2013). The Life Events Checklist for DSM-5 (LEC-5).
Unpublished instrument. Retrieved from http://www.ptsd.va.gov

Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.

Journal of Traumatic Stress, Vol. 13, No. 2, 2000

Comparison of the PTSD Symptom Scale-Interview
Version and the Clinician- Administered PTSD Scale

Edna B. and David F. Tolin’

The Clinician-Administered PTSD Scale (CAPS) is one of the most frequently
used measures of posttraumatic stress disorder (PTSD). It has been shown to be
a reliable and valid measure, although its psychometric properties in nonveteran
populations are not well known. One problem with the CAPS is its long assess-
ment time. The PTSD Symptom Scale-Interview Version (PSS-I) is an alternative
measure of PTSD severity, requiring less assessment time than the CAPS. Pre-
liminary studies indicate that the PSS-I is reliable and valid in civilian trauma
survivors. In the present study we compared the psychometric properties of the
CAPS and the PSS-I in a sample of 64 civilian trauma survivors with and without
PTSD. Participants were administered the CAPS, the PSS-I, and the Structured
Clinical Interview f o r DSM-IV (SCID) by separate interviewers, and their re-
sponses were videotaped and rated by independent clinicians. Results indicated
that the CAPS and the PSS-I showed high internal consistency, with no direr-
ences between the two measures. Interrater reliability was also high f o r both
measures, with the PSS-I yielding a slightly higher coeficient. The CAPS and
the PSS-I correlated strongly with each other and with the SCID. Although the
CAPS had slightly higher specijcity and the PSS-I had slightly higher sensitivity
to PTSD, overall the CAPS and the PSS-I peqormed about equally well. These
results suggest that the PSS-I can be used instead of the CAPS in the assess-
ment of PTSD, thus decreasing assessment time without sacrijcing reliability or
validity.
KEY WORDS: posttraumatic stress disorder; CAPS; PSS-I; SCID.

I Center for Treatment and Study of Anxiety, Department of Psychiatry, University of Pennsylvania,
3535 Market Street, 6th Floor, Philadelphia, Pennsylvania 19104.

*To whom correspondence should be addressed.

181

0894-98h7/00/0400-0181$18.00/1 c 2000 International Soclefy for Traumatic Streaa Sludie\

182

Foa and Tolin

One of the most widely used measures of posttraumatic stress disorder (PTSD)
is the Clinician-Administered PTSD Scale (CAPS; Blake et al., 1990), often re-
ferred to as the “gold-standard” measure for PTSD. The CAPS is a semistructured
interview that measures the 17 symptoms of PTSD. Each symptom is assessed
using two questions (for a total of 34 items): one measuring frequency of the
symptom’s occurrence, and the other, its intensity (e.g., distress or functional im-
pairment). To ascertain validity of response, each question is followed by a number
of probe questions that aim at clarifying the frequency and intensity of the symp-
tom. CAPS responses are used not only for making a dichotomous PTSD diagnosis,
but also for quantifying the seventy of PTSD. The CAPS was originally devel-
oped for use with combat veterans and most studies of its psychometric properties
have used this population (e.g.. Blake et al., 1990). More recently, to our knowl-
edge only one study (Blanchard et al., 1995) has examined the reliability of the
CAPS in civilian populations, yielding high to very high reliability coefficients.
Hovens et al. (1994) found high reliability and moderate validity coefficients us-
ing a Dutch-language version of the CAPS. However, that sample contained both
civilians and combat veterans; therefore, it is difficult to determine whether the
same results would apply to a civilian sample.

Although the CAPS has excellent psychometric properties, as noted by
Newman, Kaloupek, and Keane (1996), its major drawback is the substantial
amount of time required for its administration due to its large number of items.
Depending on the interviewee’s symptom picture, administration of the CAPS can
take 40 to 60 min.

One potential alternative to the CAPS is the PTSD Symptom Scale-Interview
Version (PSS-I; Foa, Riggs, Dancu, & Rothbaum, 1993). The PSS-I is a semistruc-
tured interview that consists of 17 items, corresponding to the 17 symptoms of
PTSD. Unlike the CAPS, frequency and intensity of symptoms are combined on
the PSS-I into a single rater estimate of seventy. The reason for combining these
two dimensions is that some symptoms lend themselves more readily to frequency
estimates (e.g., nightmares) whereas others are more readily described in terms of
intensity (e.g., hypervigilance). Excellent reliability and validity have been found
for the PSS-I using female victims of rape and nonsexual assault (Foa et al., 1993).
Because the PSS-I consists of only 17 items (compared to the CAPS’S 34), its
administration time is relatively short, approximately 20 to 30 min.

The purpose of the present study was to compare the psychometric proper-
ties of the CAPS and the PSS-I using a sample of individuals with and without
PTSD who had experienced a variety of traumatic events. We administered the two
interviews and compared the resulting diagnostic status and symptom severity to
one another and to that yielded by the Structured Clinical Interview for DSM-IV
(SCID; First, Spitzer, Gibbon, & Williams, 1995). If the CAPS and the PSS-I
show similar reliability and validity to each other, then the PSS-I may be a useful
alternative to the CAPS when resources are limited.

PSS-I versus CAPS 183

Method

Participants

Participants were a convenience sample of 12 clinic patients and 52 non-
clinical adult volunteers (total = a), recruited from a relatively heterogeneous
community sample in the greater Philadelphia area. The clinic patients were re-
ceiving outpatient treatment for PTSD; the remainder responded to advertisements
and requests for volunteers at community presentations. All participants were re-
imbursed $30 for their participation.

Fifty-three percent of the participants were female, and 47% were male.
Mean age was 37 years (SD = 10). Fifty-two percent were Caucasian, 39% were
African American, 3% were Hispanic, 5% were Asian American, and 1% were
other ethnicity.

All participants reported experiencing a traumatic incident that met Crite-
rion A of the DSM-ZV (American Psychiatric Association, 1994) PTSD diagnosis.
The sample included a heterogeneous range of traumatic experiences, with per-
centages as follows: rape 18%, other sexual assault 8%, nonsexual assault 32%,
fire/explosion 11 %, accident 14%, and other trauma 17%. None of the participants
were combat veterans.

Measures

PSS-I (Foa et al., 1993). The PSS-I is a semistructured interview designed
to assess current symptoms of PTSD as defined by DSM-ZV (American Psychi-
atric Association, 1994) criteria. The PSS-I consists of 17 items corresponding
to the 17 symptoms of PTSD, and yields a total PTSD severity score as well as
reexperiencing, avoidance, and arousal subscores. Each item consists of one brief
question. The participant’s answer is rated by the interviewer from 0 (Not at all)
to 3 (5 o r more times p e r week/Very much). Total severity scores on the PSS-I are
based on sums of the raw items. Symptoms measured by the PSS-I are considered
present if they are rated as 1 (Once p e r week or less/A little) or greater.

Factor analysis of the PSS-I yielded three factors: avoidancehrousal, numb-
ing, and intrusion (Foa, Riggs, & Gershuny, 1995). Internal consistency coefficients
for the PSS-I subscales range from .65 to .71 in a sample of female sexual and
nonsexual assault victims. Test-retest reliabilities range from .66 to .77 over a
1-month period. Interrater reliabilities range from .93 to .95. The PSS-I shows
good concurrent validity, as indicated by significant correlations with measures of
PTSD symptoms, depression, and general anxiety (Foa et al., 1993).

CAPS (Blake et al., 1990). The CAPS is a semistructured interview designed
to measure symptoms of PTSD according to DSM-ZZZ-R (American Psychiatric

184 Foa and Tolin

Association, 1987) criteria. The CAPS has 34 symptom-oriented items, each rated
on a 5-point scale, which correspond to the 17 symptoms of PTSD. The

CAPS

yields two total scores, one for frequency and one for intensity, as well as two sub-
scores for each of the reexperiencing, avoidance, and arousal subscales. The anchor
points of the scales vary according to symptom, but higher numbers consistently
indicate either higher frequency or intensity of the symptom.

In addition to having separate ratings of frequency and intensity, the CAPS
differs from the PSS-I in that it includes questions to be used as prompts if the
assessor needs further clarification. The CAPS also can be used to assess both
lifetime and current PTSD symptomatology; however, for the purposes of the
present study only current symptoms were assessed.

Previous research indicates that the CAPS shows excellent interrater reliabil-
ity ( r = .92 to .99) for all three subscales in combat veterans. Internal consistency
coefficients range from .73 to 3 5 . The CAPS shows good concurrent validity, as
indicated by significant correlations with self-report measures of PTSD symptoms
(Blake et al., 1990). Thus, the CAPS appears to be a reliable and valid mea-
sure. Partly because of the complexity inherent in obtaining separate scores for
frequency and intensity, several scoring rules have been proposed for the CAPS
(Blanchard et al., 1995; Weathers, Ruscio, & Keane, 1999). With motor vehicle
accident victims, Blanchard et al. (1 995) used three scoring rules: a liberal rule
requiring a score of at least 2 as the sum of the frequency and intensity ratings
for a given item; a moderate rule requiring a score of 3, and a conservative rule
requiring a score of 4. As expected, rates of PTSD were highest using the liberal
rule, and lowest using the conservative rule.

With combat veterans, Weathers et al. (1999) examined nine different ra-
tionally and empirically derived scoring rules for the CAPS. Three scoring rules
were particularly recommended: the “F 1/12’’ rule (liberal rule) required a frequency
score of at least 1 and an intensity score of at least 2 for each item. This rule was
recommended for screening purposes to avoid false negatives. When false positives
and false negatives are equally undesirable (e.g., differential diagnoses), the “SCID
Symptom-Calibrated (SXCAL)” rule was recommended. The SXCAL rule uses
the optimally efficient severity-score cutoff for each item for predicting the pres-
ence or absence of the corresponding PTSD symptom on the SCID (Weathers et al.,
1999). When false positives needs to be minimized (e.g., confirming a diagnosis),
the conservative “Clinician-Rated 60” scoring was recommended. Accordingly, a
symptom is considered present if the combination of frequency and intensity for
that item was rated as present by at least 60% of a sample of 25 expert clinicians
(Weathers et al., 1999). This resulted in different cutoff scores for each CAPS item.

SCID. (First et al., 1995). The SCID is a structured interview measuring
DSM-ZV (American Psychiatric Association, 1994) symptoms of PTSD. The SCID
diagnosis of PTSD showed acceptable agreement with indexes obtained from
previously validated assessment instruments included in the National Vietnam
Veterans Readjustment Study (Kulka, Schlenger, Jordan, & Hough, 1988), and

PSS-I versus CAPS 185

was identified previously as an instrument of choice in the assessment of rape-
related PTSD (Resnick, Kilpatrick, & Lipovsky, 1991).

On the SCID, each symptom is assessed using one question, and the inter-
viewer rates each symptom on a 3-point scale: absent or false, subthreshold, and
threshold or true. Symptoms are considered present if they are assigned the latter
rating.

Procedure

Thirty-nine participants were interviewed by two clinicians. The first inter-
viewer queried the participant about trauma history and assisted the participant
in identifying a single traumatic even that would be the focus of the interview.
Participants reporting more than one traumatic event were instructed to select the
most bothersome incident for this interview. Participants were also instructed to
refer to the same traumatic event for all interviews, and reviews of videotapes indi-
cated that all participants complied with this instruction. One interviewer used the
CAPS and the other, the PSS-I. The order of administering the two instruments as
well as which instrument would be used by which clinician were each determined
randomly. Over the course of the study, 22 clinicians conducted the interviews.
Participants were instructed to refer to the same traumatic event in both interviews.
Clinicians were instructed not to discuss a participant’s interview with one another
until all interview data had been collected for that individual.

All interviews were videotaped. The videotapes were reviewed by at least two
raters who did not have access to the interviewers’ ratings. These raters scored the
CAPS and the PSS-I on the basis of the participant’s responses in the videotapes;
later, these ratings were compared to those of the interviewer.

To assess convergent validity with the SCID, an additional 25 participants
were administered the CAPS and the PSS-I as described above as well as the
PTSD module of the SCID; the latter was administered by a third clinician. The
order of the three interviews and the assignment of the clinician-interviewer were
determined randomly.

All interviewers and raters were doctoral or master’s level clinicians who were
trained in the use of both instruments by the instruments’ developers (Dr. Edna Foa
for the PSS-I and Dr. Frank Weathers for the CAPS). To ensure standard admin-
istration and scoring, interviewers and raters met weekly to review the interviews,
ascertain adherence to interview procedures, and resolve scoring discrepancies.

Results

Kolmogoroff-Smirnov tests of the distribution of scores on the PSS-I and
CAPS indicated that scores were not normally distributed. Therefore, nonpara-
metric statistics were used wherever possible.

186

Table 1. Cronbach’s Aluha Coefficients for the PSS-I and the CAPSa

Foa and Tolin

PSS-I CAPS

No. ofItems (Y No.ofItems (Y

Total score 17 3 6 34 .88
Reexperiencing subscale 5 .70 10 .70
Avoidance subscale 7 .I4 14 .76
Arousal subscale 5 .65 10 .7 I

“PSS-I = PTSO Symptom Scale-Interview Version; CAPS = Clinician-
Administered PTSD Scale.

Reliability of the PSS-I and the CAPS

Internal consisrency. Cronbach’s alpha was calculated on PSS-I and CAPS
total scores and subscale scores. Because the CAPS includes two items per symp-
tom (frequency and intensity) and the PSS-I includes only one item, we used a
dichotomous coding of each item to indicate its presence or absence. By doing so,
we controlled for the different number of items.

Alpha coefficients for the PSS-I and the CAPS are shown in Table 1 . Internal
consistency was good to very good for all scales and subscales of both the PSS-I
and the CAPS, with the alpha coefficient ranging from .70 to .88 for the CAPS
and from .65 to .86 for the PSS-I. Thus, the internal consistency of the PSS-I and
the CAPS were comparable.

To further examine internal consistency, we correlated each item’s raw score
with the total score. The average item-total correlation for the PSS-I was S 9 ,
with correlations ranging from .11 to .74. For the CAPS, the average item-total
correlation was .52 with arange of .21 to .68. On both interviews, the item reflecting
the symptom of “inability to recall an important aspect of the trauma” showed low
correlations with the total score (on the PSS-I, p(63) = . l I , p = .39; on the CAPS,
p(63) = .21, p = .09). Thus, on this index of internal consistency, the CAPS and
the PSS-I were again quite similar.

The correlations among the three symptom cluster and the total severity scores
for the CAPS and the PSS-I are presented in Table 2. The intercorrelations among
subscales for each instrument were moderate to high and the overall picture was
again quite similar.

Interviewer-rater reliability. Interviewer-rater reliability was calculated by
comparing the interviewer’s ratings to those of the videotape raters. Because there
were several raters and one interviewer for each instrument, reliability coefficients
were calculated as follows: First, each videotape rater was assigned a number (1-4).
Next, Spearman correlation coefficients were calculated between the interviewer
and rater 1, the interviewer and rater 2, and so on. The resulting coefficients were
translated into Fisher’s z scores (Rosenthal & Rosnow, 1984) and averaged. Then,
the average z score was translated back to p to yield a single interrater reliability

PSS-I versus CAPS 187

Table 2. Spearman Correlations Among the Subscales of the PSS-I
and the CAPS

Subscale Total Score Reexperiencing

Avoidance

PSS-I
Reexperiencing .82*
Avoidance .92* .63*
Arousal .88* .63* .71*

Reexperiencing .87*
Avoidance .90* .68*
Arousal .88* .67* .70*

CAPS

* p < ,001.

Table 3. Interviewer-Rater Reliability Coefficients and Percentage Agreement
for the PSS-I and the CAPS

~~ ~ ~

Pss-I CAPS
p % Agreement p %Agreement

Reexperiencing subscale .93* 99.2 .89* 92.5
Avoidance subscale .91* 97.5 .86* 88.5
Arousal subscale .92* 94.2 .8 1 * 93.4
Total score/PTSD diagnosis .93* 98.3 .95* 86.6

coefficient. Percentage of rater agreement for the presence or absence of each
symptom was calculated by averaging the agreement of each videotape rater with
that of the interviewer. Rater agreement for the CAPS was calculated using the
F1/I2 rule (Weathers et al., 1999), since this was the original scoring rule reported
by Blake et al. (1990). Using other scoring rules for the CAPS did not change
interrater reliability significantly.

Table 3 presents the reliability coefficients of the total scores and for each
subscale, as well as the percentage of rater agreement on the presence or absence
of each symptom cluster and PTSD diagnosis. As can be seen in Table 3, both the
CAPS and the PSS-I showed excellent interviewer-rater reliability. There were
no substantial differences between the two measures, although the PSS-I showed
consistently higher rates of agreement between raters for both the correlations and
percentage agreements.

Validity of the PSS-I and the CAPS

Frequency of PTSD diagnosis. Thirty (46%) of participants met diagnostic
criteria for PTSD according to the PSS-I. Rates of PTSD with the CAPS varied

188 Foa and Tolin

Table 4. Diagnostic Agreement Between the CAPS and the PSS-I

PSS-SR

CAPS Scoring Rule % Agreement Kappa

Liberal (Weathers) 83 .65
Moderate (Weathers) 78 .55
Conservative (Weathers) 70 .38
Liberal (Blanchard) 86 .I2
Moderate (Blanchard) 84 .68
Conservative (Blanchard) 80 .58

Note. Blanchard = Blanchard et al. (1995); Weathers = Weathers et al. (1999).

Table 5. Correlations Between the Subscales of the CAPS and the PSS-I

CAPS

Reexperiencing Avoidance

Arousal

PSS-I Total Score Subscale Subscale Subscale

Total score .87* .76* .74* .76*

Avoidance subscale .75* .55* .75* .64*
Arousal subscale .17* .64* .63* .78*

Reexperiencing subscale .76* .79* .57* .64*

Note. Correlation coefficients between scales measuring the same symptoms on both
interviews are italicized.
* p < .001.

according to the scoring rule used. Using the Blanchard et al. (1995) diagnostic
rules, 33 (5 1 %) were diagnosed with PTSD with the liberal rule, 2 8 (43%) with the
moderate rule, and 2 1 (32%) with the conservative rule. Rates of PTSD diagnosis
on the CAPS also vaned across the different scoring rules described by Weathers
et al. (1999). Using the liberal rule, 23 (35%) were diagnosed with PTSD; 20
(31%) with the moderate rule, and 11 (17%) with the conservative rule. Thus,
PTSD rates yielded by the PSS-I were similar to those obtained with the Blanchard
et al. moderate scoring rule. Both the Blanchard et al. and the PSS-I rates were
somewhat higher than those emerging from the Weathers et al. rules.

Concurrent vuiidity. A high correlation of p = .87 (p < .001) was found be- tween the CAPS and the PSS-I for the total score. Agreement across the two measures on PTSD diagnosis varied according to the CAPS scoring rule used (see Table 4). Table 5 displays the Spearman correlations between the interview scales.

Convergent validity. To assess convergent validity, CAPS and PSS-I scores
were compared to the PTSD section of the SCID. Spearman correlation coefficients
indicated that the SCID total score correlated strongly with the CAPS total score
p ( 2 3 ) = 3 3 , p < .001; and PSS-I total score, p ( 2 3 ) = .73, p < .001. To examine whether the correlation between SCID and CAPS total scores was greater than the correlation between SCID and PSS-I total scores, a Hotelling’s t test was performed. Results were not significant: t ( 2 4 ) = 1.68, p > .05.

PSS-I versus CAPS 189

Table 6. Agreement Between the SCID and the CAPS and the PSS-I

CAPS

Liberal Moderate Conservative PSS-I
Scoring Rule Scoring Rule Scoring Rule

Standard
SCID Subscale Blanchard Weathers Blanchard Weathers Blanchard Weathers Scoring Rule

Total Score
%Agreement 80 84 80 88 88 84 80
Sensitivity 0.86 0.71 0.71 0.71 0.71 0.43 0.86
Specificity 0.78 0.89 0.83 0.94 0.94 1 .oo 0.78
Kappa .56 .60 .52 .69 .69 .5 1 .56

%Agreement 84 80 84 84 80 56 92
, Sensitivity 0.85 0.80 0.85 0.85 0.80 0.45 0.90

K”PP” .57 .49 .57 .57 .49 2.5 .78

Reexperiencing

Specificity 0.80 0.80 0.80 0.80 0.80 1.00 1 .oo

% Agreement 80 84 80 84 88 88 80
Sensitivity 0.88 0.75 0.75 0.62 0.75 0.62 0.88

Avoidance

Specificity 0.76 0.88 0.82 0.94 0.94 1.00 0.76
Kappa .58 .63 .56 .61 .7 1 .69 .58

%Agreement 64 84 68 72 80 72 76
Sensitivity 1 .oo 1 .oo 1.00 1.00 1.00 0.50 I .oo
Specificity 0.31 0.69 0.39 0.46 0.62 0.92 0.54
Kappa .30 .68 .38 .45 .6 1 .43 .53

Arousal

Notes. Blanchard = scoring rule from Blanchard et al. (1995); Weathers = scoring rule from Weathers
et al. (1999). Percent agreements are calculated to reflect whether participants met or exceeded the
symptom count for the DSM-IV diagnosis.

When data were analyzed according to the presence or absence of symptoms
rather than a continuous score, the results varied according to the scoring rule
used. As shown in Table 6, both the PSS-I and the CAPS showed moderate to
strong agreement with the SCID. The PSS-I showed somewhat higher sensitivity,
whereas the CAPS showed somewhat higher specificity, especially using more
conservative scoring rules. On both the CAPS and the PSS-I, the arousal subscales
showed high sensitivity but relatively low specificity with the SCID. Given the
strong agreement between the PSS-I and CAPS on the arousal subscale (r = .78),
the low specificity may reflect a psychometric weakness of the SCID rather than
of the two instruments in question. Overall, however, the CAPS and the PSS-I
performed quite similarly in relation to the SCID.

Interview duration. Precise interview times were available for 42 sets of in-
terviews. Mean time to complete the PSS-I was 21.96 min (SD = 1 1 S l ) , and mean
time to complete the CAPS was 32.75 min (SD = 15.94). The CAPS was found
to take significantly longer than the PSS-I to administer, t(41) = 5.93, p c .001,
Cohen’s d = 0.78. When we sampled only those patients with PTSD (as indi-
cated by the PSS-I; n = 16), the CAPS still took significantly longer (M = 42.76,

190 Foa and Tolin

SD = 10.74) than did the PSS-I (M = 28.69, SD = 9.92), t ( 15) = 4.64, p < .001, and the effect was greater than before (Cohen’s d = 1.36). Thus, the PSS-I ap- pears to be a briefer instrument than the CAPS, and this is particularly true for interviewees reporting significant PTSD symptoms.

Discussion

Results of the present study suggest that the PSS-I compares favorably to the
CAPS, as evidenced by internal consistency, item-total correlations, intersubscale
correlations, and interviewer-rater agreement. In terms of validity, the total score
and subscale scores of the PSS-I correlate strongly with the corresponding scores
on the CAPS. When the PSS-I and the CAPS are used to predict PTSD diagnosis
according to the SCID, both the PSS-I and the CAPS show moderately strong
agreement with the SCID. Results for the CAPS vary according to the scoring
rule used; however, in general, it appears that the PSS-I may have slightly higher
sensitivity, whereas the CAPS may have slightly higher specificity. Thus, the PSS-
I may have a small advantage in detecting actual PTSD, whereas the CAPS’S
advantage may be in ruling out false positives. However, it should be emphasized
that differences between the CAPS and the PSS-I were relatively small compared
to their similarities.

Limitations of the present study include a relatively small sample size, com-
pared to the large numbers of participants to whom the CAPS has been administered
(e.g., Weathers et al., 1999). The present study examined only civilian trauma vic-
tims, and thus the obtained results may not generalize to combat veterans. We did
not collect data on the test-retest stability of either the CAPS or the PSS-I; such
data would shed more light on the comparability of the two interviews. Finally,
although interviewers were trained in both the CAPS and the PSS-I, because of the
institution where the study was conducted (MCP Hahnemann University), most
of the interviewers were more familiar with the PSS-I. Additional studies using
interviewers who are equally familiar with the CAPS and the PSS-I would help to
clarify this issue.

Because the two instruments show such similar internal consistency, inter-
viewer-rater reliability, and validity, the PSS-I may be a useful alternative to the
CAPS. In this study, the PSS-I took significantly less time to administer, with
no appreciable loss of psychometric strength. Thus, when time and/or financial
resources are limited, the PSS-I may be the interview method of choice for the
assessment of PTSD.

References

American Psychiatric Association ( 1 987). Diagnostic and statistical manual of mental dis0rder.s (3rd
ed.-rev.). Washington, DC: Author.

PSS-I versus CAPS 191

American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th
ed.). Washington, DC: Author.

Blake, D. D., Weathers, F. W., 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-I.
Behavior Therapist, 13, 187-1 88.

Blanchard, E. B., Hickling, E. J., Taylor, A. E., Forneris, C. A., Loos, W., & Jaccard, J. (1995). Effects
of varying scoring rules of the Clinician-Administered PTSD Scale (CAPS) for the diagnosis of
posttraumatic stress disorder in motor vehicle accident victims. Behaviour Research and Therupy,
33,471-475.

First, M . B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1995). Structured clinical interview for
DSM-IVuxis I disorders-Patient edition (SCID I / r ) version 2.0). New York: Biometrics Research
Department.

Foa, E. B., Riggs, D. S., Dancu, C. V., & Rothbaum, B. 0. (1993). Reliability and validity of a brief
instrument for assessing posttraumatic stress disorder. Journal of Traumatic Stress, 6,459-473.

Foa, E. B., Riggs, D. S., & Gershuny, B. S. (1995). Arousal, numbing, and intrusion: Symptom structure
of PTSD following assault. American Journal ofPsychiarg 152, 1 16-120.

Hovens, J. E., van der Ploeg, H. M., Klaarenbeek, M. T. A., Bramsen, I., Schreuder, J. N., & Rivero,
V. V. (1994). The assessment of posttraumatic stress disorder with the Clinician Administered
PTSD Scale: Dutch results. Journal of Clinical Psychology, 50,325-340.

Kulka, R. A., Schlenger, W. E., Jordan, B. K., & Hough, R. L. (1988, October). Preliminan, survey
jndings of the National Vietnam Veterans’ Readjustment Study. Symposium presented at the 4th
annual meeting of the International Society for Traumatic Stress Studies, Dallas.

Newman, E., Kaloupek, D. G., & Keane, T. M. (1996). Assessment of posttraumatic stress disorder
in clinical and research settings. In B. A. van der Kolk, A. C. McFarlane, & L. Weisaeth (Eds.),
Traumatic stress: The effects of overwhelming experience on mind, body, and society (pp. 242-
273). New York Guilford Press.

Resnick, H. S., Kilpatrick, D. G., & Lipovsky, J. A. (1991). Assessment of rape-related posttraumatic
stress disorder: Stressor and symptom dimensions. fsychologicul Assessment, 3,561-572.

Rosenthal, R., & Rosnow, R. L. (1984). Essentials of behavioral research: Methods and data analysis
(2nd ed.). New York: McGraw-Hill.

Weathers, F. W., Ruscio, A. M., & Keane, T. M. (1999). Psychometric properties of nine scoring rules
for the Clinian-Administered PTSD Scale (CAPS). Psychological Assessment, 11, 124-1 33.