The PTSD Checklist—Civilian Version: Reliability, Validity, and Factor Structure in a Nonclinical Sample Daniel Conybeare,1 Evelyn Behar,1 Ari Solomon,2 Michelle G. Newman,3 and T. D. Borkovec3 1

University of Illinois at Chicago Williams College 3 The Pennsylvania State University 2

Objectives: We examined the reliability, validity, and factor structure of the posttraumatic stress diorder (PTSD) Checklist-Civilian Version (PCL-C; Blanchard, Jones-Alexander, Buckley, & Forneris, 1996) Participants were 471 undergraduate among unselected undergraduate students. Participants: students at a large university in the Eastern United States and were not preselected based on trauma history or symptom severity. Results: The PCL-C demonstrated good internal consistency and retest reliability. Compared with alternative measures of PTSD, the PCL-C showed favorable patterns of convergent and discriminant validity. In contrast to previous research using samples with known trauma exposure, we found support for both 1-factor and 2-factor models of PTSD symptoms. Conclusions: Overall, the PCL-C appears to be a valid and reliable measure of PTSD symptoms, even among nonclinical samples, and is superior to some alternative measures of PTSD. The factor structure among nonclinical samples may not reflect each of the PTSD symptom “clusters” (i.e., reC 2012 Wiley Periodicals, Inc. J Clin Psychol experiencing, avoidance/numbing, and hyperarousal).  00:1–15, 2012. Keywords: PTSD; Assessment; Trauma

Clinicians and researchers commonly use self-report questionnaires to measure symptoms of psychopathology, including posttraumatic stress disorder (PTSD). In clinical settings, self-report measures are used to rapidly screen individuals for symptoms of PTSD (e.g., Blanchard et al., 1996; Dobie et al., 2002; Foa, Cashman, Jaycox, & Perry, 1997). In research settings, self-report measures may be used as the primary indicator of PTSD symptoms (e.g., Jakupcak et al., 2007; Ruscio, Ruscio, & Keane, 2002). In both of these settings, PTSD measures may be administered to unselected samples; not all studies using such measures restrict their samples to individuals with known PTSD symptoms or trauma exposure that meets Diagnostic and Statistical Manual, Fourth Edition, Text Revision (DSM-IV-TR; American Psychiatric Association, 2000) criteria for a traumatic event. Indeed, mechanisms studies of PTSD have included nonclinical samples (e.g., Merwin, Rosenthal, & Coffey, 2009; Rubin, Boals, & Bernsten, 2008; Verwoerd, Wessel, & de Jong, 2009), as have psychometric studies of PTSD self-report measures (e.g., Eid et al., 2009; Ruggiero, Del Ben, Scotti, & Rabalais, 2003). When administering self-report measures of PTSD to nonclinical samples, clinicians and researchers alike must exercise caution in several regards. First, measures of psychopathology among nonclinical samples may serve as indicators of nonspecific negative affect, rather than as markers of specific clinical syndromes (Feldman, 1993). This suggests that a questionnaire might measure different constructs depending on the type of sample being measured. As a result, the psychometric properties of PTSD self-report measures might vary based on the degree of PTSD symptoms in the measured sample. ∗ Preparation

of this manuscript was supported in part by National Institute of Mental Health (NIMH) predoctoral National Research Service Award 1 F31 MH068167-01 to Evelyn Behar, and by NIMH Research Grant RO1 MH58593 to T.D. Borkovec Please address correspondence to: Daniel Conybeare, Department of Psychology, 1007 W Harrison St, M/C 285, University of Illinois at Chicago, Chicago, IL 60607. E-mail: [email protected]

JOURNAL OF CLINICAL PSYCHOLOGY, Vol. 00(0), 1–15 (2012) Published online in Wiley Online Library (wileyonlinelibrary.com/journal/jclp).

 C 2012 Wiley Periodicals, Inc. DOI: 10.1002/jclp.21845

2

Journal of Clinical Psychology, June 2012

A recent review found that the diagnostic accuracy of one commonly used measure of PTSD— the PTSD Checklist (PCL; Blanchard et al., 1996, Weathers, Litz, Herman, Huska, & Keane, 1993)—varied across both clinical and nonclinical samples, and this variation was partly attributed to spectrum bias (McDonald & Calhoun, 2010). Spectrum factors include symptom severity (Ransohoff & Feinstein, 1978), and spectrum bias in part refers to the increasing difficulty of discriminating between cases and noncases of a condition as symptom severity becomes more evenly distributed across the spectrum (Goehring, Perrier, & Morabia, 2004). In other words, discriminating nonclinical from clinical cases is easiest when comparing purely nonsymptomatic cases with severe clinical cases. Nonetheless, the purpose of a screening instrument is to select clinical cases from among a pool of unselected cases whose diagnostic status is unknown. Thus, examining discriminant validity among an unselected sample is important for validating self-report measures that are used as screening instruments (despite increased difficulty in demonstrating discriminant validity), and researchers have applied this methodology to other measures of psychopathology (e.g., Newman et al., 2002; Newman, Kachin, Zuellig, Constantino, & Cashman-McGrath, 2003). Importantly, spectrum factors may also influence psychometric properties within nonclinical samples; the reliability and construct validity of PTSD questionnaires and measures of other psychopathology may vary depending on symptom severity. Taken together, these findings suggest that the psychometric properties of self-report measures of PTSD may differ among nonclinical, unselected (e.g., undergraduate) samples compared with other types of samples that are more likely to have experienced severe trauma (e.g., combat veterans, motor vehicle accident victims). Another possibility is that the PCL retains its psychometric properties and factor structure across samples. In a statistical review of a variety of personality and psychopathology measures, O’Connor (2002) found that factor structures were similar across normal and abnormal (i.e., clinical) samples. However, this study did not examine measures of PTSD, and the consistency across samples may not apply to that disorder. By definition, PTSD is dependent on the occurrence of a traumatic event—Criterion A—whereas other disorders (e.g., panic disorder, generalized anxiety disorder) do not presuppose a precipitating event. Thus, the differences between nonclinical and clinical samples based on PTSD symptoms may be greater compared with the differences between nonclinical and clinical samples based on other types of symptoms, at least in terms of the presence of traumatic events. These differences may in turn lead to variability in the psychometric properties across samples. Although several investigations have examined the psychometric properties of the PCL in trauma samples, very few studies have examined the measure among undergraduate students. In a study of trauma-exposed college students, the PCL-Civilian Version (PCL-C) demonstrated good retest reliability and internal consistency, as well as adequate convergent and discriminant validities (Adkins, Weather, McDevitt-Murphy, & Daniels, 2008). Importantly, the PCL-C showed greater construct validity compared with the Civilian Mississippi Scale (CMS; Keane, Caddell, & Taylor, 1988), another commonly used measure of trauma symptoms. Although this study seems to suggest superior validity of the PCL-C compared with the CMS, those findings may not translate to nonclinical samples. In a study of unselected undergraduate students, the PCL-C demonstrated excellent internal consistency (α = .94) and good convergent and discriminant validities (Ruggiero et al., 2003). The PCL-C was more highly correlated with the CMS (r = .82) than with the Center for Epidemiological Studies-Depressed Mood Scale (r = .67) and the Symptom Checklist 90-Revised (r = .70), and the differences between these correlations were statistically significant. Notably, the PCL-C’s retest reliability varied substantially depending on the amount of time between test administrations; the correlation coefficient was .92 for immediate retests, .88 for retests that were administered within 1 week, and .68 for retests that were administered within 2 weeks. This study demonstrated positive psychometric properties of the PCL-C; however, it did not directly compare the psychometric properties of the PCL-C with those of alternative measures of PTSD such as the CMS and the Trauma Symptom Checklist-40 (TSC; Elliott & Biere, 1992). In sum, initial research on the psychometric properties of the PCL-C among unselected samples has yielded promising estimates of internal consistency, retest reliability, convergent validity, and discriminant validity. Nonetheless, previous research among unselected samples has not

PCL-C in a Nonclinical Sample

3

examined the validity of the PCL-C relative to other measures of PTSD (although this has been done in an undergraduate sample preselected for trauma exposure; Adkins et al., 2008). Thus, examination of the reliability and validity of the PCL-C within this type of sample is warranted, particularly research that directly compares the psychometric properties of the PCL-C with those of alternative measures of PTSD. Moreover, subthreshold levels of PTSD are associated with increased psychiatric comorbidity, functional impairment, and suicidality (Marshall et al., 2001), suggesting that screening for PTSD symptoms may help detect clinical cases. Currently, we are not aware of the relative characteristics of the PCL-C among unselected samples; examining discriminant validity of various PTSD measures among unselected undergraduates may enhance screening efforts in a variety of settings. Our investigation had two primary goals. First, we examined the psychometric properties of the PCL-C in a nonclinical student sample and compared the PCL-C with other measures of trauma symptoms. For the PCL-C, we examined internal consistency, retest reliability, and convergent validity with two other commonly used measures of PTSD, the CMS and the TSC. We also examined discriminant validity of the PCL-C compared with self-report measures of other syndromes known to co-occur with PTSD (social anxiety disorder, generalized anxiety disorder, panic disorder, obsessive compulsive disorder, and depression). Unlike the CMS and TSC, each question on the PCL-C assesses each of the 17 DSM-IV-TR symptoms of PTSD (APA, 2000), suggesting that the face validity of the PCL-C is higher compared with either measure. Consequently, we hypothesized that the PCL-C would demonstrate greater construct validity compared with the CMS and TSC, such that the PCL-C would correlate less strongly with other measures of psychopathology. The second purpose of this investigation was to explore the factor structure of the PCL-C in a nonclinical student sample. Numerous factor analytic studies have examined the structure of the PCL-C among clinical samples. Most commonly, researchers find support for fourfactor models of PTSD, comprising reexperiencing, avoidance, numbing, and hyperarousal factors (Asmundson, Stapleton, & Taylor, 2004). More recently, four-factor solutions have been found among disaster workers at the World Trade Center Ground Zero site (Palmieri, Weather, Difede, & King, 2007), sexually harassed women (Palmieri & Fitzgerald, 2005), and motor vehicle accident victims (Elklit & Shevlin, 2007). However, several investigations have also found support for two-factor and three-factor solutions. Researchers have found twofactor solutions (reexperiencing/avoidance and hyperarousal/numbing) among motor vehicle accident victims (Buckley, Blanchard, & Hickling, 1998; Taylor, Kuch, Koch, Crockett, & Passey, 1998), and three-factor solutions among cancer survivors (DuHamel et al., 2004). Among a college sample selected for trauma exposure, one investigation found support for both threefactor and four-factors solutions (Elhai, Gray, Docherty, Kashdan, & Kose, 2007). Thus far, no investigations have examined the PCL-C’s factor structure in a nonclinical student sample. Given the consistency with which researchers find structures that represent the PTSD symptom clusters in one form or another, we expected to find multiple-factor solutions based on symptom clusters.

Method Participants We recruited 500 participants from introductory undergraduate psychology classes at a large university. The majority of participants (n = 471) completed all 17 items on the PCL-C. The sample was predominantly female (67%) and Caucasian (84.7%), and reported an average age of 19.33 years (standard deviation [SD] = 2.54). A minority of participants reported their ethnicity as Asian or Pacific Islander (6.4%), Hispanic (4.5%), or African-American (4.0%). Students received course credit in exchange for participation.

Procedure Participants completed a battery of self-report questionnaires (in counterbalanced order) in a large group setting. The PCL-C was readministered 2 weeks after the initial administration

4

Journal of Clinical Psychology, June 2012

(the CMS and TSC were not readministered). Of the original 471 participants, 316 (67%) returned for this follow-up visit and completed the second administration of the PCL-C.

Measures PCL-C (Weathers et al., 1994). The PCL-C is a 17-item self-report measure of the DSM-IV symptoms of PTSD and was derived from the PCL-Military Version (PCL-M; Weathers et al., 1993). Responses range from 1 to 5, and the total score is computed by summing all items. The civilian version is identical to the military version, except that it inquires about a “stressful experience from the past” as opposed to military trauma. Among Vietnam veterans, the PCL-M has excellent retest reliability after 2-3 days (r = .96), excellent internal consistency (α = .97), and good convergent validity as reflected by high correlations with the CMS for combat-related PTSD (r = .93), the PTSD subscale of the MMPI-2 (r = .77), and the Impact of Events Scale (r = .90). A PCL total score of greater than or equal to 50 predicts a clinical diagnosis of PTSD (as assessed by the Structured Clinical Interview for DSM-IV; First, Spitzer, Gibbon, & Williams, 1997), with an efficient balance of sensitivity and specificity (equal to .82 and .83, respectively). Since those initial findings (Weathers et al., 1993), the PCL has become one of the most widely studied self-report measures of PTSD. The psychometric properties are also good among undergraduates selected for preexisting trauma (Adkins et al., 2008; Elhai et al., 2007), as well as unselected undergraduates (Ruggiero et al., 2003). CMS (Keane et al., 1988). The CMS is a 35-item measure of the three symptom clusters of PTSD (reexperiencing, avoidance and numbing, hyperarousal). Responses range from 0 to 4, and the total score is computed by reverse scoring items 2, 6, 11, 17, 19, 22, 24, 27, 30, and 34, and then summing all items. When administered to a sample of treatment-seeking Vietnam veterans, the CMS was shown to have excellent internal consistency (α = .94) and 1-week retest reliability (r = .97). When administered to a sample of undergraduate students, the CMS demonstrated good internal consistency (α = .89) but showed less favorable patterns of convergent and discriminant validity; the CMS correlated more highly with the Beck Depression Inventory (r = .71) and the Stait-Trait Anxiety Invenotry-Trait subscale (r = .70) than with the Purdue PTSD QuestionnaireRevised (r = .52) and the Impact of Events Scale (r = .36; Lauterbach, Vrana, King, & King, 1997). Trauma Symptom Checklist-40 (TSC-40; Elliott & Briere, 1992). The TSC is a 40item measure of the psychological consequences of childhood sexual abuse, including anxiety, depression, dissociation, sexual problems, and sleep disturbance. Responses range from 0 to 3, and the total score is computed by summing all items. Among a sample of professional, adult women, the TSC demonstrated good internal consistency (α = .90) and discriminated women who reported sexual abuse in childhood from those who denied it (Elliot & Briere, 1992). Among a sample of females in a psychiatric inpatient setting, the dissociation, anxiety, and depression subscales of the TSC demonstrated convergent and discriminant validity with alternative measures of those constructs (Zlotnick et al., 1996). Social Phobia Diagnostic Questionnaire (SPDQ; Newman et al., 2003). The SPDQ is a 25-item measure of the DSM-IV criteria for social anxiety disorder, as well as fear and avoidance associated with various social situations. Among treatment-seeking undergraduates, the SPDQ predicts clinical diagnoses of social anxiety disorder (using the Anxiety Disorders Interview Schedule for DSM-IV) with a sensitivity of .82 and a specificity of .85. Among unselected undergraduates, the SPDQ demonstrates good retest reliability after 2 weeks (κ = .63) and excellent internal consistency (α = .92), as well as convergent validity with the Social Interaction Anxiety Scale (r = .64) and discriminant validity with the PCL-C (r = .29), CMS (r = .34), and other measures of psychopathology. In the current study, the SPDQ was scored according to the algorithm described by Newman et al. (2003).

PCL-C in a Nonclinical Sample

5

Social Phobia and Anxiety Inventory (SPAI; Turner, Beidel, Dancu, & Stanley, 1989). The SPAI is a 45-item measure of the cognitive, somatic, and behavioral symptoms of social phobia. Responses range from 1 to 7, and the measure is scored according to the algorithm described by Turner et al. (1989). Among unselected undergraduates, the SPAI has demonstrated good internal consistency (α = .96) and retest reliability after a 2-week interval (r = .86). Among individuals with clinical diagnoses of social phobia, panic disorder with agoraphobia, or obsessive-compulsive disorder, the SPAI correctly identified 76% of those with social phobia (κ = .47). Scores on the SPAI are sensitive to treatment change (Beidel, Turner, & Cooley, 1993; Cox, Ross, Swinson, & Direnfeld, 1998; Ries et al., 1998), convergent with alternative measures of social anxiety (including the Social Interaction Anxiety Scale [SIAS]; Cox et al., 1998; Ries et al., 1998), and predict the degree of distress and avoidance during daily social interactions among individuals with social phobia (Beidel, Borden, Turner, & Jacob, 1989).

SIAS (Mattick & Clark, 1998). The SIAS is a 19-item scale that assesses symptoms of social anxiety disorder, generalized type; the SIAS measures fears of general social interactions rather than fears of being evaluated during routine activities (e.g., writing). Responses range from 0 to 4, and the measure is scored by reverse scoring items 8 and 10, and then summing all items. Among unselected undergraduates, the SIAS demonstrates high internal consistency (α = .88), as well as excellent retest reliability after 4-week and 8-week intervals (rs = .92). Among individuals with social anxiety disorder, the SIAS demonstrates sensitivity to treatment change, convergent validity with alternative measures of social anxiety disorder (including the SPAI; Cox et al., 1998; Mattick & Clark, 1998; Ries et al., 1998), and discriminant validity with measures of depression and state and trait anxiety (Mattick & Clark, 1998). Finally, the SIAS may be used to discriminate individuals with social phobia from individuals without social phobia, as well as from individuals with panic disorder and simple phobia (Mattick & Clark, 1998; Peters, 2000). Generalized Anxiety Disorder Questionnaire-IV (GAD-Q-IV; Newman et al., 2002). The GAD-Q-IV is a 9-item self-report measure of the symptoms of generalized anxiety disorder (GAD) as outlined in the DSM-IV, including the excessiveness, frequency, and uncontrollability of worry, physical symptoms of GAD, and interference and distress associated with both worry and physical symptoms. The measure is scored according to the algorithm described by Newman et al. (2002). A dimensional scoring scheme with a cutoff score of 5.7 yields high sensitivity (83%) and specificity (89%). Moreover, the GAD-Q-IV has demonstrated stability over a 2-week period, and kappa agreement of .67 with the Anxiety Disorders Interview Schedule for DSM-IV. The GAD-Q-IV has also demonstrated convergent and discriminant validity; it is more highly correlated with the PSWQ than with measures of posttraumatic stress disorder and social anxiety (Newman et al., 2002).

Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec, 1990). The PSWQ is a 16-item self-report measure of the frequency and intensity of worry. Responses range from 1 to 5, and the total score is calculated by reverse scoring items 1, 3, 8, 10, and 11, and then summing all items. Among unselected undergraduates, the PSWQ has demonstrated high internal consistency (αs ranging from .91 to .95) and good retest reliability after 4-week intervals (rs ranging from .74 to 93) and after 8-week to 10-week intervals (r = .92). A cutoff score of 62 identifies GAD cases with a sensitivity of .75 and specificity of .86 (Behar, Alcaine, Zuellig, & Borkovec, 2003). Among participants with DSM-III-diagnosed GAD, the PSWQ was uncorrelated with other measures of anxiety and depression, indicating that trait worry is a distinct construct (Meyer et al., 1990). Similarly, the PSWQ may be used distinguish GAD from other anxiety disorders (Brown, Antony, & Barlow, 1992).

Agoraphobic Cognitions Questionnaire (ACQ; Chambless, Caputo, Bright, & Gallagher, 1984). The ACQ is a 15-item measure for assessing the extent to which an individual experiences panic-related cognitions, including thoughts related to loss of control (e.g., “I am going to go crazy”) and physical concerns (e.g., “I am going to pass out”). Responses range from

6

Journal of Clinical Psychology, June 2012

1 to 5, and the total score is calculated by summing all items. Among a mixed sample of agoraphobics and nonanxious control participants, the ACQ has demonstrated adequate internal consistency (α = .80) and good retest reliability following a 1-month interval (r = .86) and was able to discriminate individuals with and without agoraphobia. The ACQ has also demonstrated convergent validity with the Body Sensations Questionnaire (r = .67) and discriminant validity with the Beck Depression Inventory (r = .38) and the State-Trait Anxiety Inventory-rait subscale (r = .35).

Padua Inventory-Washington State University Revision (PIWSR; Burns, Keortge, Formea, & Sternberger, 1996). The PIWSR is a 39-item self-report measure of obsessions and compulsions, revised from the original 60-item Padua Inventory (Sanavio, 1988) to reduce measurement overlap between obsessions and worries. The PIWSR measures five domains of OCD symptoms: obsessional thoughts about harming the self or others; obsessional impulses to harm the self or others; contamination obsessions and washing compulsions; checking compulsions; and dressing and grooming compulsions. Responses range from 0 to 4, and the total score is calculated by summing all items. Among unselected undergraduate participants, the PIWSR has demonstrated high internal consistency (α = .92) and good retest reliability following 6-month to 7-month intervals (r = .76) and can discriminate between individuals with and without OCD. The PIWSR has also demonstrated convergent validity with the Maudsley Obsessive-Compulsive Inventory (r = .61) and discriminant validity with the PSWQ (r = .37) among a European sample of ´ ´ unselected undergraduates (Jonsd ottir & Sm´ari, 2000).

Zung Self-Rating Depression Scale (Zung, 1965). The Zung is a 20-item scale that measures pervasive depressed affect, as well as physiological and psychological concomitants of depression. Responses range from 1 to 4, and the total score is calculated by reverse scoring items 2, 5, 6, 11, 12, 14, 16, 17, 18, and 20, and then summing all items. Among an unselected community sample, the Zung has demonstrated satisfactory internal consistency (α = .79; Knight, Waal-Manning, & Spears, 1983). Among individuals with depression, the Zung has demonstrated convergent validity with the Hamilton Rating Scale for Depression (r = .80) and was able to discriminate between severity levels of depression, as determined by clinical interview (Biggs, Wylie, & Ziegler, 1978). Data Transformation Prior to conducting further analyses, we examined the skewness and kurtosis of the total score for each measure to ensure that the distributions were normal. The distributions for all measures were positively skewed. We applied a combination of log and square root transformations to attain normal distributions for each measure prior to calculating correlations between measures. The purpose of these transformations was to conform the data to the assumptions of normality that are required for multivariate statistical techniques. This procedure is recommended when it does not hinder interpretability of the data (Tabachnick & Fidell, 2006). All correlations presented below were calculated using transformed variables.

Statistical Approach for Comparing Correlation Coefficients To examine the relationships between measures, we first calculated Pearson correlation coefficients between all measures. Cases with incomplete data were not included in the analyses; the n ranged from a low of 338 for the correlation between the GAD-Q-IV and the TSC to a high of 462 for the correlation between the PCL-C and the SIAS. To compare correlation coefficients, we adopted procedures described by Meng, Rosenthal, & Rubin (1992) for comparing correlated (overlapping) correlation coefficients between variables x and y and variables x and z. The procedure includes a Fisher z transformation of each correlation coefficient, followed by a difference calculation. The resulting statistical output includes a z-score

PCL-C in a Nonclinical Sample

7

and p-value for the significance of the difference between the two correlation coefficients that are being compared. To adjust for multiple comparisons, we applied Holm’s technique in which actual alpha values for each hypothesis test are calculated, ordered from smallest to largest, and then compared with significance level cutoffs that are calculated stepwise using a modified Bonferroni correction (α divided by n number of tests), in which n is equal to the number of remaining tests after the rejection of each null hypothesis (Holm, 1979). This technique affords increased power compared with the standard Bonferroni correction (Holland & Copenhaven, 1988). We calculated 40 comparisons in total; thus, our initial p-value was equal to α divided by 40 (.05/40 = .00125). This cutoff was compared with the smallest obtained alpha and was adjusted stepwise after each subsequent rejection of a null hypothesis.

Factor Analytic Procedure Prior to analysis, we calculated the Kaiser-Meyer-Olkin statistic and Bartlett’s test for sphericity and found that the data were suitable for factor analysis. To examine the factor structure of the PCL-C in our sample, we conducted an exploratory factor analysis using principal axis factoring (standard factor analysis) with varimax rotation. Exploratory factor analysis is recommended in the absence of specific predictions regarding the number of factors and the relationships between factors (Tabachnik & Fidell, 2006; Thompson, 2004). Although prior research has indicated specific two-factor to four-factor solutions among clinical samples, no previous research has indicated specific solutions among nonclinical samples, thus justifying the use of exploratory factor analysis in our investigation. Principal axis factoring is recommended when the primary goal is to understand the relationships between underlying factors and observed variables (Floyd & Widaman, 1995; Tabachnik & Fidell, 2006). To determine the number of factors, we examined the scree plot, applied the “eigenvalues-greater-than-one” rule (Kaiser, 1960), and used parallel analysis. For the parallel analysis, we used 100 random data sets and retained factors if the eigenvalue from the original data set was higher than the mean eigenvalue from the random data set (O’Connor, 2000). We used varimax rotation to allow factors to correlate, given that PTSD symptom clusters would be expected to correlate in nature.

Results Descriptive Statistics Table 1 presents the means and standard deviations for each measure, as well as the number of participants who completed each measure.

Internal Consistency of the PCL-C The PCL-C was highly internally consistent during the initial administration (α = .94, n = 471) and the retest administration (α = .92, n = 316). For both administrations, the deletion of any one item from the PCL-C would not have increased alpha significantly.

Retest Reliability of the PCL-C The PCL-C demonstrated good retest reliability (r = .66; n = 316) after a 2-week interval after the initial administration.

Convergent and Discriminant Validity of the PCL-C To examine the convergent and discriminant validity of the PCL-C (see Table 2), we first compared the PCL-CMS correlation (r = .60) with the correlations between the PCL-C and all other measures (rs ranging from .28 to .59). The PCL-C was more strongly correlated with the CMS than with the SPDQ (z = 5.25, p < .001), SPAI (z = 7.33, p < .001), SIAS (z = 5.93,

8

Journal of Clinical Psychology, June 2012

Table 1 Means, Standard Deviations, and Number of Completers for Each Measure

PCL-C PCL-C (reliability) CMS TSC SPDQ SPAI SIAS GAD-Q-IV PSWQ ACQ PIWSR Zung

M

SD

N

29.12 27.67 69.89 21.53 6.66 39.87 18.82 2.81 43.84 1.41 19.57 36.67

12.31 10.24 15.11 15.88 4.73 29.80 11.78 3.28 14.59 .41 18.90 8.66

471 316 430 423 413 403 462 354 430 439 453 424

Note. M = mean; SD = standard deviation; PCL-C = PTSD Checklist; CMS = Civilian Mississippi Scale; TSC = Trauma Symptoms Checklist; SPDQ = Social Phobia Diagnostic Questionnaire; SPAI = Social Phobia and Anxiety Inventory; SIAS = Social Interaction Anxiety Scale; GAD-Q-IV = Generalized Anxiety Disorder Questionnaire for DSM-IV; PSWQ = Penn State Worry Questionnaire; ACQ = Agoraphobic Cognitions Questionnaire; PIWSR = Padua Inventory – Washing State University Revision; Zung = Zung Self-Rating Scale for Depression.

Table 2 Pearson Correlation Coefficients Between the PCL-C, CMS, TSC, and Measures of Other Types of Psychopathology

PCL-C CMS TSC SPDQ SPAI SIAS GAD-Q-IV PSWQ ACQ PIWSR Zung

PCL-C

CMS

TSC

– .60 .61 .39 .28a .39b .45 .43 .50c .56 .59d,e

– – .75 .50 .46a .56b .46 .46 .55 .54 .76d

– – – .49 .38 .49 .55 .51 .61c .55 .72e

Note. PCL-C = PTSD Checklist; CMS = Civilian Mississippi Scale; TSC = Trauma Symptoms Checklist; SPDQ = Social Phobia Diagnostic Questionnaire; SPAI = Social Phobia and Anxiety Inventory; SIAS = Social Interaction Anxiety Scale; GAD-Q-IV = Generalized Anxiety Disorder Questionnaire for DSM-IV; PSWQ = Penn State Worry Questionnaire; ACQ = Agoraphobic Cognitions Questionnaire; PIWSR = Padua Inventory – Washing State University Revision; Zung = Zung Self-Rating Scale for Depression. Superscripts indicate significant differences between pairs of correlation coefficients. For example, the correlation between the PCL-C and SPAI (r = .28) is significantly different compared to the correlation between the CMS and the SPAI (r = .46).

p < .001), GAD-Q-IV (z = 3.41, p = .001), and PSWQ (z = 4.26, p < .001). The PCL-C was not more strongly correlated with the CMS than with the ACQ, PIWSR, or Zung. We then compared the PCL-TSC correlation (r = .61) with the correlations between the PCL-C and the other measures. The PCL-C was more strongly correlated with the TSC than with the SPDQ (z = 5.41, p < .001), SPAI (z = 7.01, p < .001), SIAS (z = 5.76, p < .001),

PCL-C in a Nonclinical Sample

9

GAD-Q-IV (z = 3.89, p = .001), PSWQ (z = 4.70, p < .001), and ACQ (z = 3.41, p = .001). The PCL-C was not more strongly correlated with the TSC than with the PIWSR or Zung.

Validity Comparisons Between the PCL-C, CMS, and TSC To further examine the validity of the PCL-C, we first compared the correlations between the PCL-C and the nontrauma measures with the correlations between the CMS and the nontrauma measures. In other words, we compared the PCL-SPDQ correlation with the CMS-SPDQ correlation, the PCL-SPAI correlation with the CMS-SPAI correlation, and so on. Compared with the CMS, the PCL-C was less strongly correlated with the SPAI (z = 4.51, p < .001), SIAS (z = 4.89, p < .001), and Zung (z = 5.83, p < .001). We then compared the correlations between the PCL-C and the nontrauma measures with the correlations between the TSC and the nontrauma measures. Compared with the TSC, the PCL-C was less strongly correlated with the ACQ (z = 3.49, p < .001) and the Zung (z = 4.34, p < .001). Finally, we compared the correlations between the CMS and the nontrauma measures with the correlations between the TSC and the nontrauma measures. There were no significant differences between CMS and TSC in terms of their relationships with the nontrauma measures.

Factor Structure of the PCL-C Principal axis factoring with varimax rotation yielded two factors with eigenvalues greater than one, suggesting a two-factor solution. However, the eigenvalue for the first factor was substantially higher than that for the second factor, whereas the eigenvalue for the second factor was only marginally higher than that for proximal subsequent factors (eigenvalues for the first five items were 8.26, 1.26, .95, .86, and .74). Nonetheless, an examination of the scree plot provided support for a two-factor solution. Results of the parallel analysis provided equivocal support for a one-factor solution. The mean eigenvalues of the first two factors of the random data set were 1.34 and 1.27, meaning that the eigenvalue for the first factor from the raw data set was well above the cutoff from the random data set, whereas the eigenvalue for the second factor from the raw data set was just below the cutoff from the random data set. Overall, these findings provide support for both one-factor and two-factor models. Because we found some support for a two-factor solution, we examined it for interpretability (see Table 3 for rotated individual factor loadings). The first factor accounted for 44.65% of the total variance, and included PCL-C items 7 through 17. The second factor accounted for 7.50% of the total variance, and included items 1 through 6. All of the PCL-C reexperiencing items (as well as item 6, “avoiding thinking or talking about a stressful event”) were included in the second component. All other symptom clusters (avoidance, numbing, and hyperarousal) were included in the first component.

Discussion Although the PCL-C has demonstrated excellent psychometric properties, including convergent and discriminant validity, among clinical samples with interview-based diagnoses of PTSD (Weathers et al., 1993), only one previous study examined the PCL-C among an unselected student sample. Given that researchers and clinicians alike may utilize the PCL-C in such an unselected sample, we sought to further elucidate the psychometric properties of the PCL-C among unselected undergraduate participants. Accordingly, we examined the reliability, validity, and factor structure of the PCL-C in a sample of unselected undergraduate research participants, as well as the relative validity of the PCL-C compared with two alternative measures of trauma symptoms (the CMS and the TSC). First, we examined the reliability and validity of the PCL-C. We found similarly favorable levels of internal consistency and retest reliability as previously reported for a sample of unselected undergraduate students (Ruggiero et al., 2003). We also found support for the convergent validity of the PCL-C, although these correlations were smaller than in previous investigations. For example, the PCL-C and CMS correlated .60 in our study as compared with .82 in previous

10

Journal of Clinical Psychology, June 2012

Table 3 Rotated Factor Loadings for the 2-Factor Model PCL-C Item (10) Feeling distant or cut off from others (A/N) (9) Loss of interest in activities (A/N) (11) Feeling emotionally numb (A/N) (15) Difficulty concentrating (H) (12) Feeling as if future cut short (A/N) (14) Irritability/angry outbursts (H) (17) Jumpy or easily started (H) (16) Super alert or watchful (H) (7) Avoid activities/situations (A/N) (13) Trouble sleeping (H) (8) Trouble remember details (A/N) (4) Feeling upset at reminders (R) (1) Memories/thoughts/images (R) (2) Repeated, disturbing dreams (R) (3) Acting/feeling as if happening again (R) (5) Physical reactions at reminders (R) (6) Avoid thinking/talking/feelings (A/N)

Factor 1

Factor 2

.77 .76 .74 .69 .65 .64 .60 .60 .52 .50 .39 .31 .27 .26 .37 .39 .39

.28 .32 .26 .39 .29 .40 .35 .35 .46 .44 .25 .77 .75 .63 .62 .61 .52

Note. Boldface indicate assigned factor. Symptoms are labeled based on DSM-IV-TR symptom clusters (APA, 2000): A/N = avoidance/numbing; H = hyperarousal; R = reexperiencing.

research (Ruggiero et al., 2003). With respect to discriminant validity, we found that the PCL-C was more highly correlated with alternative measures of trauma symptoms than with measures of social anxiety and generalized anxiety disorder, including trait worry. The PCL-C was also more highly correlated with the TSC (but not with the CMS) than with a measure of panic symptoms. Second, we extended previous findings by comparing the discriminant validity of the PCL-C with that of the CMS and TSC. Compared with the CMS, the PCL-C demonstrated lower correlations with two measures of social anxiety and a measure of depression. Compared with the TSC, the PCL-C demonstrated lower correlations with measures of panic symptoms and depression. These findings indicate that in an unselected sample, the PCL-C may result in improved discriminant validity compared with the CMS and TSC when differentiating between trauma and other types of symptoms (e.g., depression). Our results indicate that the PCL-C may be superior compared with the CMS and TSC in terms of discriminating between trauma symptoms and symptoms of social anxiety disorder, panic disorder, OCD, and depression. The PCL-C did not show higher correlations with the CMS and TSC than with measures of panic, OCD, and depression, possibly indicating that the PCL-C does not optimally discriminate trauma symptoms from some other types of symptoms. However, a rival hypothesis is that this finding reflects poor discriminant validity of the CMS and TSC. This possibility seems likely, given that convergent validity of the PCL-C is based on the relationship between the PCL-C and those measures, and that our findings overall support the validity of the PCL-C. It is possible that the correlation between the PCL-C and an alternative measure of trauma would be higher compared with the correlation between the PCL-C and CMS or the PCL-C and TSC. This possibility seems especially likely given the high convergence between the CMS, TSC, and Zung (all rs > .70), which suggests that those questionnaires may measure similar constructs. Moreover, the measures of panic, OCD, and depression used in the current study were not intended to assess for DSM criteria, and this may adversely affect discriminant validity with diagnostic measures (e.g., the difference between the PCL-CMS correlation and the PCL-ACQ). To summarize, compared with the PCL-C, the CMS and TSC may be less specific to PTSD symptoms in an unselected sample.

PCL-C in a Nonclinical Sample

11

In terms of factor structure, we found support for both one-factor and two-factor solutions, and our findings do not support the existence of the two-, three-, or four-factor models found in various clinical samples. Previous two-factor solutions were defined by reexperiencing/avoidance and hyperarousal/numbing factors (Buckley et al., 1998; Talyor et al., 1998). In contrast, our two-factor solution indicated that nearly half of the total variance (44.65%) was accounted for by 11 of 17 PCL-C items, while the remaining six items (including all five reexperiencing symptoms) accounted for only 7.50% of the variance. Moreover, we also found support for a onefactor solution. This highlights the potential difficulties in detecting PTSD symptom clusters in nonclinical samples with the PCL-C, despite promising psychometric properties. Neither of our solutions was consistent with factor structures found in clinical samples (Asmundson et al., 2004). These findings provide further support for the notion that self-report measures validated among clinical samples may measure different structures of the same construct—or different constructs altogether—among nonclinical samples. Indeed, in the two-factor solution, the component accounted for by 11 of 17 items includes symptoms from three of four symptom clusters, suggesting that avoidance, numbing, and hyperarousal symptoms may be indistinguishable in an unselected sample. Nonetheless, reexperiencing symptoms were contained within one factor, which indicates that the PCL-C may be able to detect those specific symptoms and distinguish them from other symptoms of PTSD, even in samples without trauma exposure. This finding may also reflect a difference in the nature of “trauma” symptoms in an unselected sample such as ours. Because this version of the PCL-C asks respondents to answer the items with respect to a past stressful event (as opposed to a traumatic event), the reexperiencing symptoms that emerge as a unique factor may instead reflect symptoms of rumination. The factor that contained the other 11 symptoms, on the other hand, may represent general reactions to stress. Last, there was some support for a one-factor solution. This highlights the difficulty in detecting specific symptom clusters in a largely nonsymptomatic sample. Although we found some support for the validity of the PCL-C among unselected undergraduates, our results underscore potential difficulties in discriminating between different types of symptoms among largely nonsymptomatic samples. For example, although the CMS and TSC were more strongly related to the Zung than was the PCL-C, the correlation between the PCL-C and the Zung was nonetheless moderately high (r = .59), possibly due to overlap between measures—overlap that may be attenuated in a clinical sample, given the greater ease with which different types of symptoms can be discriminated from one another among samples with a broader range of symptom severity (Goehring et al., 2004). Alternatively, the overlap between the PCL-C and the Zung may instead reflect high rates of comorbidity between PTSD and depression rather than overlap between items and/or constructs, in which case the correlations may not be attenuated in a clinical sample. Moreover, the convergent validity found between the PCL-C and the other measures of trauma symptoms was weaker compared with that found in research using clinical samples (Weathers et al., 1993). Although this may reflect decreased convergent validity of the PCL-C among nonclinical samples compared with clinical samples, it more likely reflects decreased convergent validity of the CMS and TSC among nonclinical samples, given that our findings overall support the validity of the PCL-C. Finally, the factor analysis indicated that although the reexperiencing symptoms may be accounted for by a single factor, a large proportion of the variance was accounted for by all of the other symptom clusters; despite some specificity, the majority of the PCL-C items were accounted for by a single factor, indicating an overall lack of discrimination between symptom clusters. This latter finding is consistent with previous arguments that measures of psychopathology, when used among nonclinical samples, actually assess general distress as opposed to specific syndromes (Feldman, 1993). Given that the majority of the variance in the PCL-C total score was accounted for by a single factor, the hypothesis that the PCL-C primarily measures negative affect among nonclinical samples is still viable. Thus, future investigations of the PCL-C in unselected samples should examine nonspecific negative affect (as well as related constructs, such as neuroticism), and examine the discriminant validity of the PCL-C in relation to those constructs. Such studies may determine which constructs (i.e., symptoms of

12

Journal of Clinical Psychology, June 2012

PTSD or nonspecific negative affect) most accurately represent the constructs measured by the PCL-C.

Conclusion Compared with the TSC and CMS, the PCL-C may result in improved validity for measuring trauma symptoms in unselected undergraduate populations. Indeed, the PCL-C demonstrated the strongest discriminant validity in terms of the relationships between trauma measures and measures of other syndromes. However, the PCL-C was substantially correlated with various measures of other types of psychopathology, suggesting that discriminating symptoms of PTSD with those of other disorders may be difficult, at least with the self-report questionnaires used in the current study. These findings underscore the importance of (a) using caution when administering to unselected samples questionnaires that were validated among clinical samples and (b) cross-validation of diagnostic status (or severity level) using clinical interviews. Nonetheless, the overall pattern of findings supports the validity of the PCL-C among unselected undergraduate samples. Moreover, instances in which the PCL-C was unable to discriminate from (or converge with) other measures may be because of a lack of validity of those other measures rather than problems with the PCL-C. In conclusion, the PCL-C appears to be a common measure of choice for PTSD symptoms and may be used with some confidence among nonclinical, student samples. For researchers who wish to use the PCL-C among nonclinical samples, we suggest adding measures of general negative affect (e.g., Positive and Negative Affect Schedule; Clark & Watson, 1990) and/or neuroticism (e.g., General Temperament Survey; Carver & White, 1994). These measures may elucidate whether participants scoring highly on the PCL-C score highly as a result of factors other than trauma symptoms.

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, 1393–1402. doi:10.1016/j.janxdis.2008 .02.002 American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author. doi:10.1176/appi.books.9780890423349 Asmundson, G.J.G., Stapleton, J.A., & Taylor, S. (2004). Are avoidance and numbing distinct PTSD symptom clusters? Journal of Traumatic Stress, 17, 467–475. doi:10.1007/s10960-004-5795-7 Behar, E., Alcaine, O., Zuellig, A.R., & Borkovec, T.D. (2003). Screening for generalized anxiety disorder using the Penn State Worry Questionnaire: A receiver operating characteristic analysis. Journal of Behavior Therapy and Experimental Psychiatry, 34, 25–43. doi:10.1016/S0005-7916(03)00004-1 Beidel, D.C., Turner, S.M., Cooley, M.R., Woody, S.R., & Messer, S.C. (1994). A multicomponent behavioral treatment for social phobia: Social effectiveness therapy. Behaviour Research and Therapy, 32, 381–390. doi:10.1016/0005-7967(94)90001-9 Beidel, D.C., Turner, S.M., Jacob, R.G., & Cooley, M.R. (1989). Assessment of social phobia: Reliability of an impromptu speech task. Journal of Anxiety Disorders, 3, 149–158. doi:10.1016/0887-6185(89)90009-1 Biggs, J.T., Wylie, L.T., & Ziegler, V.E. (1978). Validity of the Zung self-rating depression scale. British Journal of Psychiatry, 132, 381–385. doi:10.1192/bjp.132.4.381 Blanchard, E.B., Jones-Alexander, J., Buckley, T.C., & Forneris, C.A. (1996). Psychometric properties of the PTSD Checklist (PCL). Behaviour Research and Therapy, 34, 669–673. doi:10.1016/00057967(96)00033-2 Brown, T.A., Antony, M.M., & Barlow, D.H. (1992). Psychometric properties of the Penn State Worry Questionnaire in a clinical anxiety disorders sample. Behaviour Research and Therapy, 30, 33–37. doi:10.1016/0005-7967(92)90093-V Burns, G.L., Keortge, S.G., Formea, G.M., & Sternberger, L.G. (1996). Revision of the Padua Inventory of obsessive compulsive disorder symptoms: Distinctions between worry, obsessions and compulsions. Behaviour Research and Therapy, 34, 163–173. doi:10.1016/0005-7967(95)00035-6

PCL-C in a Nonclinical Sample

13

Carver, C.L., & White, T.L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The BIS/BAS scales. Journal of Personality and Social Psychology, 67, 319–333. doi:10.1037/0022-3514.67.2.319 Chambless, D.L., Caputo, G.C., Bright, P., & Gallagher, R. (1984). Assessment of ‘fear of fear’ in agoraphobics: The Body Sensations Questionnaire and the Agoraphobic Cognitions Questionnaire. Journal of Consulting and Clinical Psychology, 52, 1090–1097. doi:10.1037/0022-006X.52.6. 1090 Clark, L.A., & Watson, D. (1990). General Temperament Survey (GTS; Unpublished manuscript). Southern Methodist University, Dallas, TX. Cox, B.J., Ross, L., Swinson, R.P., & Direnfeld, D.M. (1998). A comparison of social phobia outcome measures in cognitive-behavioral group therapy. Behavior Modification, 22, 285–297. doi: 10.1177/01454455980223004 Dobie, D.J., Kivlahan, D.R., Maynard, C., Bush, K.R., McFall, M., Epler, A.J., & Bradley, K.A. (2002). Screening for post-traumatic stress disorder in female Veteran’s Affairs patients: Validation of the PTSD checklist. Psychiatry and Primary Care, 24, 367–374. doi:10.1016/S0163-8343(02)00207-4 Eid, J., Larsson, G., Johnsen, B.H., Laberg, J.C., Bartone, P.T., & Carlstedt, B. (2009). Psychometric properties of the Norwegian Impact of Event Scale-Revised in a nonclinical sample. Nordic Journal of Psychiatry, 63, 426–432. doi:10.1080/08039480903118190 Elhai, J.D., Gray, M.J., Docherty, A.R., Kashdan, T.B., & Kose, S. (2007). Structural validity of the posttraumatic stress disorder checklist among college students with a trauma history. Journal of Interpersonal Violence, 22, 1471–1478. doi:10.1177/0886260507305569 Elliott, D.M., & Briere, J. (1992). Sexual abuse trauma among professional women: Validating the Trauma Symptom Checklist-40 (TSC-40). Child Abuse and Neglect, 16, 391–398. doi:10.1016/01452134(92)90048-V Fechner-Bates, S., Coyne, J.C., & Schwenk, T.L. (1994). The relationship of self-reported distress to depressive disorders and other psychopathology. Journal of Consulting and Clinical Psychology, 62, 550–559. doi:10.1037/0022-006X.62.3.550 Feldman, L.A. (1993). Distinguishing depression and anxiety in self-report: Evidence from confirmatory factor analysis on nonclinical and clinical samples. Journal of Consulting and Clinical Psychology, 61, 631–638. doi:10.1037/0022-006X.61.4.631 First, M.B., Spitzer, R.L., Gibbon, M., & Williams, J.B.W. (1997). Structured clinical interview for DSM-IV Axis I disorders: Clinician version (SCID-CV). Washington, DC: American Psychiatric Press. Foa, E.B., 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 Goehring, C., Perrier, A., & Morabia, A. (2004). Spectrum bias: A quantitative and graphical analysis of the variability of medical diagnostic test performance. Statistical Medicine, 23, 125–133. doi:10.1002/sim.1591 Haaga, D.A.F., & Solomon, A. (1987) Impact of Kendall, Hollon, Beck, Hammen, and Ingram on treatment of the continuity issue in “depression” research. Cognitive Therapy and Research, 17, 313–324. doi:10.1007/BF01177657 Holland, B.S., & Copenhaver, M.D. (1988). Improved Bonferroni-type multiple testing procedures. Psychological Bulletin, 104, 145–149. doi:10.1037/00332909.104.1.145 Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6, 65–70. Jakupcak, M., Conybeare, D., Phelps, L., Hunt, S., Holmes, H.A., Felker, B., . . . McFall, M.E. (2007). Anger, hostility, and aggression among Iraq and Afghanistan War veterans reporting PTSD and subthreshold PTSD. Journal of Traumatic Stress, 20, 945–954. doi:10.1002/jts.20258 Jonsdottir, S.D., & Smari, J. (2000). Measuring obsessions without worry: Convergent and discriminant validity of the revised Padua Inventory in an Icelandic student population. Scandinavian Journal of Behaviour Therapy, 29, 49–56. doi:10.1080/028457100750066397 Kaiser, H.F. (1960). The application of electronic computers to factor analysis. Education and Psychological Measurement, 20, 141–151. doi:10.1177/001316446002000116 Keane, T.M., Caddell, J.M., & Taylor, K.L. (1988). Mississippi Scale for combat-related posttraumatic stress disorder: Three studies in reliability and validity. Journal of Consulting and Clinical Psychology, 56, 85–90. doi:10.1037/0022-006X.56.1.85

14

Journal of Clinical Psychology, June 2012

Knight, R.G., Waal-Manning, H.J., & Spears, G.F. (1983). Some norms and reliability data for the State-Trait Anxiety Inventory and the Zung Self-Rating Depression Scale. British Journal of Clinical Psychology, 22, 245–249. doi:10.1111/j.2044-8260.1983.tb00610.x 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, 499–513. doi:10.1002/jts.2490100313 Mattick, R.P., & Clarke, J.C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behaviour Research and Therapy, 36, 455–470. doi:10.1016/S00057967(97)10031-6 McDonald, S.D., & Calhoun, P.S. (2010). The diagnostic accuracy of the PTSD checklist: A critical review. Clinical Psychology Review, 30, 976–987. doi:10.1016/j.cpr.2010.06.012 Meng, X., Rosenthal, R., & Rubin, D.B. (1992). Comparing correlated correlation coefficients. Psychological Bulletin, 111, 172–175. doi:10.1037/0033-2909.111.1.172 Merwin, R.M., Rosenthal, M.Z., & Coffey, K.A. (2009). Experiential avoidance mediates the relationship between sexual victimization and psychological symptoms: Replicating findings with an ethnically diverse sample. Cognitive Therapy and Research, 33, 537–542. doi:10.1007/s10608-008-9225-7 Meyer, T.J., Miller, M.L., Metzger, R.L., & Borkovec, T.D. (1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28, 487–495. doi:10.1016/00057967(90)90135-6 Newman, M.G., Kachin, K.E., Zuellig, A.R., Constantino, M.J., & Cashman, L. (2003). The Social Phobia Diagnostic Questionnaire: Preliminary validation of a new self-report diagnostic measure of social phobia. Psychological Medicine, 33, 623–635. doi:10.1017/S0033291703007669 Newman, M.G., Zuellig, A.R., Kachin, K.E., Constantino, M.J., Przeworski, A., Erickson, T., & CashmanMcGrath, L. (2002). Preliminary reliability and validity of the GAD-Q-IV: A revised self-report diagnostic measure of generalized anxiety disorder. Behavior Therapy, 33, 215–233. doi:10.1016/S00057894(02)80026-0 Peters, L. (2000). Discriminant validity of the Social Phobia and Anxiety Inventory (SPAI), the Social Phobia Scale (SPS) and the Social Interaction Anxiety Scale (SIAS). Behaviour Research and Therapy, 38, 943–950. doi:10.1016/S0005-7967(99)00131-X Ransohoff, D.F., & Feinstein, A.R. (1978). Problems of spectrum and bias in evaluating the efficacy of diagnostic tests. New England Journal of Medicine, 26, 926–930. doi:10.1056/NEJM197810262991705 Ries, B.J., McNeil, D.W., Boone, M.L., Turk, C.L., Carter, L.E., & Heimberg, R.G. (1998). Assessment of contemporary social phobia verbal report instruments. Behaviour Research and Therapy, 36, 983–994. doi:10.1016/S0005-7967(98)00078-3 Rubin, D.C., Boals, A., & Bernsten, D. (2008). Memory in posttraumatic stress disorder: Properties of voluntary and involuntary, traumatic and nontraumatic autobiographical memories in people with and without posttraumatic stress disorder symptoms. Journal of Experimental Psychology: General, 137, 591–614. doi:10.1037/a0013165 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 Ruscio, A.M., Ruscio, J., & Keane, T.M. (2002). The latent structure of posttraumatic stress disorder: A taxometric investigation of reactions to extreme stress. Journal of Abnormal Psychology, 111, 290–301. doi:10.1037/0021-843X.111.2.290 Sanavio, E. (1988). Obsessions and compulsions: The Padua Inventory. Behaviour Research and Therapy, 26, 169–177. doi:10.1016/0005-7967(88)90116-7 Tabachnick, B.G., & Fidell, L.S. (2006). Using multivariate statistics (5th ed.). Needham Heights: Allyn & Bacon. Thompson, B. (2004). Exploratory and confirmatory factor analysis: Understanding concepts and applications. Washington, DC: American Psychological Association. doi:10.1037/10694-000 Turner, S.M., Beidel, D.C., Dancu, C.V., & Stanley, M.A. (1989). An empirically derived inventory to measure social fears and anxiety: The Social Phobia and Anxiety Inventory. Psychological Assessment, 1, 35–40. doi:10.1037/1040-3590.1.1.35 Verwoerd, J., Wessel, I., & de Jong, P.J. (2009). Individual differences in experiencing intrusive memories: The role of the ability to resist proactive interference. Journal of Behavior Therapy and Experimental Psychiatry, 40, 189–201. doi:10.1016/j.jbtep.2008.08.002

PCL-C in a Nonclinical Sample

15

Weathers, F.W., Litz, B.T., Herman, D.S., Huska, J.A., & Keane, T.M. (1993). The PTSD checklist: Reliability, validity, and diagnostic utility. Paper presented at the Annual Meeting of International Society for Traumatic Stress Studies, San Antonio, TX, October. Weathers, F.W., Litz, B.T., Huska, J.A., & Keane, T.M. (1994). PTSD Checklist—Civilian version. Boston: National Center for PTSD, Behavioral Science Division. Zlotnick, C., Shea, M.T., Begin, A., Pearlstein, T., Simpson, E., & Costello, E. (1996). The validation of the Trauma Symptom Checklist – 40 (TSC-40) in a sample of inpatients. Child Abuse and Neglect, 20, 503–510. doi:10.1016/0145-2134(96)00032-4 Zung, W. (1965). A self-rating depression scale. Archives of General Psychiatry, 12, 63–70.