Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Nov 30.
Published in final edited form as: Psychol Trauma. 2015 Nov;7(6):591–599. doi: 10.1037/tra0000045

Trauma Coping Self-Efficacy: A Context Specific Self-Efficacy Measure for Traumatic Stress

Charles C Benight 1, Kotaro Shoji 2, Lori E James 3, Edward E Waldrep 4, Douglas L Delahanty 5, Roman Cieslak 6
PMCID: PMC4664052  NIHMSID: NIHMS723409  PMID: 26524542

Abstract

The psychometric properties of a Trauma Coping Self-Efficacy (CSE-T) scale that assesses general trauma-related coping self-efficacy perceptions were assessed. Measurement equivalence was assessed using several different samples: hospitalized trauma patients (n1 = 74, n2 = 69, n3 = 60), three samples of disaster survivors (n1 = 273, n2 = 227, n3 = 138), and trauma exposed college students (N = 242). This is the first multi-sample evaluation of the psychometric properties for a general trauma-related CSE measure. Results showed that a brief and parsimonious 9-item version of the CSE performed well across the samples with a robust factor structure; factor structure and factor loadings were similar across study samples. The 9-item scale CSE-T demonstrated measurement equivalence across samples indicating that the underlying concept of general post-traumatic CSE is organized in a similar manner in the different trauma-exposed groups. These results offer strong support for cross-event construct validity of the CSE-T scale. Associations of the CSE-T with important expected covariates showed significant evidence for convergent validity. Finally, discriminant validity was also supported. Replication of the factor structure, internal reliability, and other evidence for construct validity is a critical next step for future research.

Keywords: coping self-efficacy, traumatic stress, validity, reliability, psychometric

Trauma adaptation consists of dynamic coping processes involving both the management of the original traumatic experience and the challenge of post-traumatic environmental demands (Park & Ai, 2006; Van der Kolk, McFarlane, & Weisaeth, 1996). One way of understanding this dynamic post-traumatic coping process is through self-regulation outlined in social cognitive theory (SCT; Benight & Bandura, 2004). According to SCT, humans utilize self-evaluation as a key mechanism for self-regulation in order to meet novel environmental challenges (Bandura, 1997). Coping with trauma creates an intense need for self-management in order to regain a sense of equilibrium. One of the most important components of self-evaluation is one's perceived ability to manage critical recovery demands (i.e., self-efficacy perceptions; Benight & Bandura, 2004). Coping self-efficacy (CSE) within a traumatic stress context, a key self-evaluative variable, refers the perceived capability for managing the internal and external post-traumatic recovery demands (Benight & Bandura, 2004). Benight and Bandura (2004) posited that positive self-efficacy is central to effective adaptation because it provides a sense of control facilitating adaptive coping.

CSE as a Predictor of Post-Traumatic Outcomes

Recent reviews support the predictive capacity of CSE in understanding the post-traumatic recovery process (Benight & Bandura, 2004; Luszczynska, Benight, & Cieslak, 2009). Indeed, the effect sizes (r range −.55 to −.62) for CSE as a predictor in longitudinal studies on mass trauma is significantly more powerful than other predictors evaluated in post-traumatic recovery studies (e.g., dissociation, social support, previous psychopathology: r range ± .17 to .35) (Ozer, Best, Lipsey, & Weiss, 2008).

Coping self-efficacy predicted post-trauma recovery for survivors of a myriad of different traumatic experiences including: childhood sexual abuse (Cieslak, Benight, & Lehman, 2008), domestic violence (Benight, Harding-Taylor, Midboe, & Durham, 2004), combat (Solomon, Benbenishty, & Mikulincer, 1991), hurricanes (Benight, Ironson, & Durham, 1999; Hirschel & Schulenberg, 2009), and terrorist attacks (Benight et al., 2000). Further, longitudinal research on disaster survivors (Benight & Harper, 2002), victims of acute physical injuries (Flatten, Walte, & Perlitz, 2008), and survivors of motor vehicle accidents (Benight, Cieslak, Molton, & Johnson, 2008) supported CSE as a key mediating mechanism by which initial distress influence subsequent development of post-traumatic symptoms. Collectively, these studies support CSE as a pivotal factor in trauma recovery.

CSE as an Intervention Target

Beyond the evidence supporting the predictive capacity of CSE in trauma adaptation, CSE beliefs provide an important target for intervention. Bandura (1997) reviewed substantial empirical evidence for increasing CSE perceptions by promoting mastery experiences, opportunities for vicarious success modeling, positive verbal persuasion, and reductions in physiological arousal. Furthermore, all the current evidence based trauma treatments (e.g., CBT-PE, CPT, EMDR) include a component focused on improving self-referent beliefs. A trauma specific measure of CSE provides a useful outcome measurement for these interventions.

Aim of the Present Study

Bandura (1997) suggested self-efficacy measurement should be contextualized to improve predictive power. Previously, we created specific measures for hurricane recovery (Benight et al., 1999), domestic violence trauma (Benight et al., 2004), and motor vehicle accident recovery (Benight et al., 2008). However, developing new CSE measures for each possible trauma is untenable. The present study describes the development and validation of a trauma-focused CSE measure (CSE-T) that assesses these perceptions across trauma experiences.

Evaluating a new measure is predicated on the idea that no other measure currently exists that captures what we want to assess. Chesney, Neilands, Chambers, Taylor, and Folkman (2006) published a scale titled the “Coping Self-Efficacy Scale”. Specifically, it is noted that this measure was designed to evaluate the perceived ability to utilize different general coping strategies (e.g., problem focused coping), not the situational demands of trauma recovery. Traumatic stressors are far more extreme than day-to-day stressors, and as such challenge coping resources beyond anything previously experienced (Benight & Bandura, 2004). The CSE-T attempts to tap these trauma-related challenges as well as posttraumatic symptoms. Indeed, the CSE-T addresses typical issues faced by survivors of trauma (e.g., coping with reminders of the trauma), while being general enough that it can be used to evaluate and compare CSE among groups exposed to different traumas.

A general self-efficacy measure also exists and is utilized extensively in health psychology (Schwarzer & Jerusalem, 1995) and limitedly in trauma studies. This scale is not context specific and is therefore less useful in measuring CSE related to trauma recovery. The CSE-T provides a critical self-regulatory trauma specific measure that is currently unavailable.

The present analysis utilized several samples with different types of traumatic stress exposure. An acutely traumatized hospital-based sample served as the anchor sample to determine initial scale structure. We then confirmed the scale structure in a sample comprised of disaster survivors from three different disasters (a hurricane and two wildfire samples). Finally, we utilized a trauma exposed undergraduate sample for our final comparison group. Factor structure, item-refinement, cross-sample confirmation, reliability/validity, and factor/intercept/variance invariance were all evaluated. We hypothesized that the CSE-T would correlate negatively with measures of post-traumatic distress (e.g., post-traumatic stress symptoms, worry, depression) and negative post-traumatic cognitions. We anticipated positive correlations between the CSE-T and post-traumatic growth and psychological well-being.

Method

Sample 1: Hospitalized trauma patients

Participants were recruited through the Summa Health System Community Oriented Medical Professionals and Surgeons (COMPAS) Clinic while attending a follow-up appointment between two days and two weeks after being admitted as trauma in-patients. Seventy-four participants completed the questionnaires at the baseline, 69 participants at 6 weeks; and 60 at 3 months. Inclusion criteria were being between the ages of 18 and 65, proficiency in English, and Glasgow Coma score of 14 or above (Teasdale & Jennett, 1974). Injury Severity Scores (Baker, O'Neill, Haddon Jr, & Long, 1974) ranged from 1 to 22 (M = 6.08, SD = 4.68). The mean age of the sample was 35.69 years old (SD = 12.28) at Time 1. The sample consisted of 73.0% Caucasian and 27.0% African American. Mean post-traumatic stress symptom scores on the PTSD Checklist-Civilian Version (PCL-C; Blanchard et al., 1996) at Time 2 (M = 48.37) and Time 3 (M = 41.28) suggest a significantly distressed sample (cutoff score for likely Posttraumatic Stress Disorder is 50). The breakdown for trauma exposure was: motor vehicle accident (n = 30), other accident (n = 25), non-sexual assault (n = 18, 9 of whom were shot), sexual assault (n = 1). Table 1 displays participants’ demographic and primary outcome variable descriptive information.

Table 1.

Demographic Characteristics, Means, and Standard Deviations for Study Variables in the Hospital, Disaster Survivor, and Undergraduate Student Samples

Variable Sample 1 T1 Sample 1 T 2 Sample 1 T 3 Sample 2 T1 Sample 2 T2 Sample 2 T3 Sample 3
Mean age in years (SD) 35.69 (12.28) 35.88 (12.16) 36.40 (12.35) 45.76 (14.08) 46.54 (13.82) 47.48 (14.00) 22.24 (5.64)
Gender (n)
    Male 59.5% (44) 62.3% (43) 65.0% (39) 27.1% (74) 28.6% (65) 31.9% (44) 17.8% (43)
    Female 40.5% (30) 37.7% (26) 35.0% (21) 72.5% (198) 70.9% (161) 67.4% (93) 82.2% (199)
Ethnicity (n)
    Caucasian 73.0% (54) 72.5% (50) 71.7% (43) 92.35 (252) 92.1% (209) 92.0% (127) 78.9% (191)
    African American 27.0% (20) 27.5% (19) 28.3% (17) 1.47% (4) 1.3% (3) 1.4% (2) 4.1% (10)
    Hispanic 2.56% (7) 2.6% (6) 2.2% (3) 9.9% (24)
    Asian 0.37% (1) 0.4% (1) 1.7% (4)
    Native American 0.37% (1) 0.4% (1) 2.1% (5)
    Other 2.93% (8) 3.08% (7) 4.3% (6) 3.3% (8)
Marital status (n)
    Single 66.2% (49) 66.7% (46) 65.0% (39) 16.5% (56) 17.2% (39) 15.9% (22) 70.7% (171)
    Married 21.6% (16) 20.3% (14) 23.3% (14) 66.7% (182) 66.1% (150) 69.6% (96) 23.1% (56)
    Separated/divorced/widowed 12.2% (9) 13.0% (9) 11.6% (7) 16.5% (45) 16.3% (37) 15.2% (21) 6.2% (15)
Highest education (n)
    1st – 11th Grades 9.5% (7) 10.1% (7) 11.7% (7) 2.93% (8) 2.6% (6) 2.9% (4)
    High school/GED 40.5% (30) 39.1% (27) 38.4% (23) 5.49% (15) 4.0% (9) 1.4% (2) 3.7% (9)
    Trade school 1.4% (1) 1.4% (1) 1.7% (1) 1.47% (4) 1.8% (4)
    Associate degree/some college 39.2% (29) 39.1% (27) 36.7% (22) 22.0% (60) 22.9% (52) 23.2% (32) 90.1% (218)
    Bachelor's degree 8.1% (6) 8.7% (6) 10.0% (6) 31.9% (87) 32.2% (73) 37.0% (51) 6.2% (15)
    Advanced college degree 1.4% (1) 1.4% (1) 1.7% (1) 36.3% (99) 36.6% (83) 35.5% (49)
Annual income
    Less than $10,000 32.4% (24) 30.4% (21) 30.0% (18) 3.5% (7) 3.6% (6) 4.0% (5) 34.3% (83)
    $10,000 – 20,000 28.4% (21) 30.4% (21) 30.0% (18) 1.5% (3) 1.8% (3) 2.4% (3) 17.4% (42)
    $20,000 – 30,000 14.9% (11) 14.5% (10) 16.7% (10) 5.5% (11) 3.0% (5) 3.2% (4) 14.9% (36)
    $30.000 – 40,000 5.4% (4) 5.8% (4) 3.3% (2) 4.5% (9) 5.4% (9) 6.4% (8) 6.6% (16)
    $40,000 – 50,000 8.1% (6) 8.7% (6) 8.3% (5) 8.5% (17) 9.0% (15) 10.4% (13) 7.% (19)
    More than $50,000 10.9% (8) 11.1% (7) 11.7% (7) 67.0% (134) 68.4% (114) 68.8% (86) 19.0% (46)
Study Variables (SD)
    Self-efficacy 5.17 (1.22) 5.07 (1.28) 5.41 (1.30) 5.18 (1.26) 5.44 (1.25) 5.67 (1.17) 5.41 (1.03)
    Posttraumatic stress symptoms 48.37 (18.68) 41.28 (17.47) 18.88 (7.60) 17.40 (7.10) 17.08 (7.04) 38.05 (13.20)
    Posttraumatic cognition 2.65 (1.19) 2.69 (1.28) 2.47 (1.13)
    Peritraumatic dissociation 2.47 (0.92)
    Worry 2.74 (0.76) 2.71 (0.72) 2.57 (0.54)
    Depression 22.23 (9.83) 19.73 (9.67) 17.89 (8.55)
    Psychological well-being 4.55 (0.67)
    Posttraumatic growth 2.77 (1.19)
Open in a new tab

Note. Sample 1 = hospital patients; Sample 2 = disaster survivor sample; Sample 3 = student sample. Some percentages did not add up to 100% because of missing data. n for Sample 1 T1 = 74; n for Sample 1 T2 = 69; n for Sample 1 T3 = 60; n for Sample 2 T1 = 273; n for Sample 2 T2 = 227; n for Sample 2 T3 = 138; n for Sample 3 = 242.. SD = standard deviation; n = the number of participants. Scores of posttraumatic stress symptoms for the disaster survivor sample were based on 10 items from the M-PSS. In the disaster sample, only the sample from the Waldo Canyon Wildfire study collected data for annual income (n1 = 189, n2 = 155, n3 = 123).

Procedure

Clinic staff screened eligible patients and asked if they would be interested in participating. A research assistant described the details of the study and obtained informed consent and HIPAA agreement. The baseline questionnaire packet was complete at home and was picked up the following day. Participants were contacted by phone one week prior to the 6-week and 3-month assessments. Research assistants retrieved study packets from participants at each time point.

Sample 2: Disaster survivors

Participants were survivors of several natural disasters and part of randomized controlled trials of a web-based intervention for disaster recovery. The first study was completed with Hurricane Ike survivors. Fifty-six respondents completed the Time 1 online survey. Among those, 45 respondents completed the Time 2 survey. The second sample was recruited following the Bastrop Fire in Texas. In this sample, 28 respondents completed the Time 1 survey; 27 respondents the Time 2 survey; and 15 respondents completed Time 3. Approximately a third of the sample had accessed a mental health provider post-disaster (n = 9). The last group of survivors was collected following the Waldo Canyon Fire in Colorado. A total of 189 respondents completed the Time 1 survey; 155 respondents completed the Time 2 survey; and 123 respondents completed the Time 3 survey (see Table 1 for demographic information of the total sample). The average scores on our post-traumatic stress symptom measure (maximum 50) suggest a mildly distressed combined sample (Mt1 = 18.88; Mt2 = 17.40; Mt3 = 17.08).

Procedure

Each disaster study followed very similar procedures and used many of the same measures. The studies differed slightly in terms of assessment timing for post-tests and follow-ups. For the Hurricane Ike study participants were recruited through the University of Texas Medical Branch (UTMB) from a larger study being conducted on stress and coping in Hurricane Ike survivors. Individuals 21 years or older who had access to the Internet and met distress criteria (i.e., scored five or greater on the TSQ [Brewin et al., 2002] or a score higher than 20 on the PSS [Cohen, Kamarck, & Mermelstein, 1983]) were invited to participate. For the other two wildfire studies no distress inclusion criteria were utilized.

Hurricane Ike data collection began 11 months post hurricane. Participants were sent a recruitment email referring them to the primary investigator. For details on the procedure for this study please see Steinmetz et al. (2012). The Bastrop fire participants (n = 28) were recruited approximately 2 months after the fire through a multi-pronged approach. Through coordination with the Texas P.R.I.D.E. Crisis Counseling Program, participants were contacted by counselors or through radio spots, print ads, and community meetings. Participants contacted a researcher by e-mail or phone and were given study information. The participants completed the initial survey and then were randomly assigned to either the website access or usual care group. Participants were assessed at baseline, 2 weeks, and 60 days. Each participant received $25 per survey set completion. The final disaster sample (n = 189) was collected approximately 45 days following the Waldo Canyon Fire. Participants were recruited through print media, television, and community response email list-serves. Participants went directly to the pre-test surveys online and were automatically randomly assigned to an experimental group. Participants completed the baseline assessment, a 30-day post-test and a 60-day follow-up assessment.

Table 1 depicts the demographics for this combined sample of disaster survivors (N1 = 273, N2 = 227, N3 = 138). At Time 1, the mean age of participants was 45.76 years old (SD = 14.08), and 72.5% of them were females. Most participants were Caucasian (92.35%).

Sample 3: Undergraduate sample

Students taking undergraduate psychology courses from a mid-sized university in Colorado were offered extra credit or cash for participation in a study to evaluate the efficacy of a new web-based intervention for trauma recovery. Only participants who had been exposed to a major traumatic event in the last 12 months and met Criterion A from the DSM-IV-TR (2000) were included in the analysis. Of the initial 300, 242 met the inclusion criteria. The majority of the sample was Caucasian (78.9%) and female (82.2%). The mean age of participants was 22.24 years (SD = 5.64). The mean score for the post-traumatic stress symptoms measure demonstrated moderately high levels of distress (M = 38.05) on the PCL-C. Table 1 displays demographic information for the sample.

Procedure

The participants completed the survey through an online research program. Individuals were then randomly assigned to one of three groups (web, paper version of the website, and waitlist control). A post-test was completed after a 30-day period.

Measures

All three samples utilized the CSE-T. The original 20-item version of the CSE-T was designed by identifying common trauma-related demands that survivors faced across the context specific trauma CSE measures created by the first author (e.g., Hurricane CSE, Domestic Violence CSE, see Benight, Harding-Taylor, Midboe, & Durham, 2004 or Benight, Ironson, & Durham, 1999). Participants were asked to rate their capability to handle a series of post-traumatic demands on a 7-point scale ranging from 1 (not at all capable) to 7 (totally capable). Please see items in Table 2.

Table 2.

Total Variance and Factor Loadings in Exploratory Principal Axis Factoring for CSE-T Items

Factor Loadings
TSES Items Sample 1 Sample 2 Sample 3
*1. Deal with my emotions (anger, sadness, depression, anxiety) since I experienced my trauma. .70 .84 .71
2. Deal with the impact of the trauma has had on my life. .74 .84 .81
*3. Get my life back to normal. .70 .79 .70
4. Have conversations about the traumatic experience. .55 .60 .63
5. Accept what happened. .51 .67 .70
6. Find some meaning in what happened. .56 .60 .55
7. Control distressing thoughts about the traumatic experience. .79 .80 .83
8. Control distressing images of the event/experience that come into my mind. .73 .75 .77
*9. Not “lose it” emotionally. .80 .82 .68
10. Deal with thoughts about my own vulnerability. .80 .82 .74
*11. Manage distressing dreams or images about the traumatic experience. .71 .73 .73
*12. Not be critical of myself about what happened. .70 .71 .68
*13. Be optimistic since the traumatic experience. .80 .81 .74
*14. Be supportive to other people since the traumatic experience. .63 .57 .54
*15. Control thoughts of the traumatic experience happening to me again. .68 .73 .61
16. Cope with thoughts that I can't handle things anymore. .74 .85 .72
*17. Get help from others about what happened. .62 .62 .63
18. Not lose my temper at others. .47 .56 .55
19. Be strong emotionally. .75 .86 .68
20. Not be critical of myself about how I'm handling this (e.g., “I should be dealing with this better!). .75 .81 .71
Total Variance (%) 50.58 57.62 49.92
Cronbach's alpha for 20 items .95 .96 .94
Cronbach's alpha for selected 9 items .90 .91 .87
Open in a new tab

Note. The asterisk indicates an item selected for the 9-item version of the CSE-T. Sample 1 = hospital patient sample; Sample 2 = disaster survivor sample; Sample 3 = student sample.

Measures in Sample 1

In the hospital sample peritraumatic dissociation was measured using the Peritraumatic Dissociative Experiences Questionnaire (PDEQ; Marmar, Metzler, & Otte, 2004), trauma related negative cognitions with the Post-traumatic Cognitions Inventory (PCI; Foa, Ehlers, Clark, Tolin, & Orsillo, 1999), post-traumatic stress symptoms using the PTSD Checklist-Civilian Version (PCL-C; Blanchard et al., 1996), and demographic information with items asking ethnicity, gender, marital status, employment, education, and annual income.

Measures in Sample 2

For the disaster samples, in addition to the CSE-T, The Modified PTSD Symptoms Scale (Falsetti, Resnick, Resick, & Kilpatrick, 1993) was used to assesses the degree of distress about PTSD symptoms in the past two weeks. Only the first 10 items of M-PSS were utilized in this analysis due to the inadvertent utilization of only the first 10 items from the measure in the Bastrop and Waldo Canyon samples. The final 7 items target hyperarousal symptoms. Thus, correlations with this 10-item measure should be interpreted with caution. The M-PSS has shown high internal consistency and good concurrent validity. The Cronbach's alpha coefficients for the present study were .88 at Time 1, .89 at Time 2, and .89 at Time 3. Other measures included The Center for Epidemiologic Studies Depression Scale (CED-D; Radloff, 1977) and The Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec, 1990). A five-item demographic questionnaire (age, gender, race, ethnicity, marital status, and education) was also included in the Time 1 assessment. Annual income was assessed only in the Waldo Canyon sample.

Measures in Sample 3

Post-traumatic stress symptoms were assessed using the PCL-C (the same measurement as in the hospital sample). The Ryff's Psychological Well-Being Scale (PWBS; Daraei, 2013; Ryff, 1989) and the Post-traumatic Growth Inventory (PTGI; Tedeschi & Calhoun, 1996) were used to assess positive responses to trauma.

Data Analysis

Missing data were imputed using the maximum likelihood estimation using AMOS for participants who responded to any items of the measurements. Less than 1% of the values were replaced for Time 1, 2, 3, respectively for Sample 1 and 2. There were no missing values in the undergraduate student sample. A Velicer's minimum average partial test (MAP; O'Connor, 2000; Velicer, Eaton, & Fava, 2000) was used to identify the number of components in the 20 items version of the CSE-T in the hospital sample at Time 1. Exploratory principal components analyses were performed to compute factor loadings of the items for each sample based on the results of the Velicer's MAP test. A confirmatory factor analysis was performed on the items selected in the hospital sample to test a model fit for a shortened version of the CSE-T in each sample using AMOS. Cutoff criteria for the fit indices were: CFI greater than .90 (Hu & Bentler, 1999), RMSEA lower than .10 (Browne & Cudeck, 1993), and SRMR lower than .08 (Hu & Bentler, 1999). Cronbach's alphas were calculated to assess internal consistency, and Pearson's correlations were calculated between CSE-T and other associated variables to assess criterion related validity. A principal components analysis was performed using selected CSE-T items and the 10 items of PDEQ in the hospital sample to evaluate discriminant validity. Finally, we conducted a test of invariance for the model representing a short version of the CSE-T across all samples using AMOS (Byrne, 2001). Across all samples, we compared the fit indices among an unconstrained model (model 1) and a model with factor loadings constrained to be equal (model 2), a model with intercepts constrained to be equal (model 3), a model with variances constrained to be equal (model 4), and a final model combining constrained models that showed invariance with the unconstrained model (model 5).

Results

Exploratory Principal Components Analysis

The Velicer's MAP test on the 20 items of the CSE-T revealed one component among those items in the hospital sample at Time 1. An exploratory principal components analysis (PCA) was performed on the 20 items of the CSE-T in each sample. Based on the results of the Velicer's MAP test, the number of components was set at 1. Table 2 displays the total variance and factor loadings for each sample.

Item Selection

Item selection was based on three approaches. First, we examined the item-total correlations. Second, we analyzed the factor loadings based on the PCA. Finally, we looked at each item identified for removal and evaluated the clinical utility of the item.

Using the hospital-based sample at time 1 as our reference group due to the level of acute post-traumatic distress, we examined redundancy among the 20 CSE-T items. There were several high corrected item-total correlations along with high Cronbach's alphas, indicating item redundancy. Nine items (items 2, 7, 8, 9, 10, 13, 16, 19, and 20) had a corrected item-total correlation greater than .70 (Stevens, 2009). All of these items were removed except for items 9 (“Not lose it emotionally.”) and 13 (“Be optimistic since the traumatic experience.”) that were retained due to clinical utility.

Factor loadings were then examined from the exploratory PCA in the hospital sample. Based on the criterion for the cut-off point of the factor loading below .617, four additional items (items 4, 5, 6, and 18) were removed. The corrected-item-total correlation analysis and the factor loading review ultimately resulted in nine items in a short version of the CSE-T (see Table 2 for selected items).

Reliability and Validity

Internal consistency for the 9-item version of the CSE-T indicated good reliability (see Table 3). Examination of test-retest reliability in the hospital sample over a 6-week period demonstrated a strong relationship (r = .72), high reliability over a 3-month period (r = .60), and high reliability between 6 weeks and 3 months (r = .57). In the disaster survivor sample, test-retest reliability was strong over each time period (2 weeks r = .80; 30 days r = .81; and 2 months r = .76).

Table 3.

Correlation Matrix for Variables in the Hospital and Disaster Survivor Samples

Pearson's Correlation
Variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
1 Self-efficacy T1 .80*** .76*** −.67*** −.61*** −.70*** −.49*** −.42*** −.35*** −.65*** −.61*** −.46***
2 Self-efficacy T2 .72*** .81*** −.70*** −.76*** −.74*** −.36*** −.33*** −.31*** −.59*** −.65*** −.51***
3 Self-efficacy T3 .60*** .57*** −.69*** −.70*** −.80*** −.32*** −.37*** −.40*** −.56*** −.60*** −.60***
4 PT stress symptoms T1 .83*** .82*** .26*** .24*** .33*** .62*** .62*** .56***
5 PT stress symptoms T2 −.45*** −.66*** −.35** .81*** .25*** .28*** .31*** .59*** .61*** .56***
6 PT stress symptoms T3 −.48*** −.47*** −.72*** .62*** .27** .34*** .35*** .58*** .62*** .65***
7 PT cognition T1 −.70*** −.61*** −.37*** .65*** .52***
8 PT cognition T2 −.52*** −.73*** −.34** .83*** .51*** .80***
9 PT cognition T3 −.49*** −.48*** −.70*** .49*** .76*** .63*** .54***
10 PT dissociation −.39*** −.34** −.33* .57*** .57*** .42*** .52*** .43***
11 Worry T1 .73*** .59*** .42*** .31*** .32***
12 Worry T2 .57*** .33*** .36*** .41***
13 Worry T3 .32*** .32*** .48***
14 Depression T1 .70*** .55***
15 Depression T2 .70***
16 Depression T3
Open in a new tab

Note. The lower diagonal region of the matrix represents values for the hospital sample. The upper diagonal region of the matrix represents values for the disaster survivor sample. Scores of PT stress symptoms for the disaster sample was based on 10 items of the M-PSS. n for the Hospital sample: T1 = 74, T2 = 69, T3 = 60. n for the disaster survivor sample T1 = 273, T2 = 227, T3 = 138. T1 = Time 1; T2 = Time 2; T3 = Time 3; PT = posttraumatic.

***

p < .001

**

p < .01

*

p < .05

We calculated Pearson's correlations between total mean scores of the 9-item CSE-T and other important conceptual and clinical outcomes to examine criterion validity of the CSE-T (see Table 3). CSE-T was significantly and negatively associated with post-traumatic stress symptoms in all samples. In the hospital sample, CSE-T was significantly and negatively associated with post-traumatic negative cognitions. In the disaster survivor sample, CSE-T was significantly and negatively correlated with worry and depression. In the undergraduate student sample (not shown in Table 3), CSE-T demonstrated a positive and significant correlation with psychological well-being (r = .49) and a significant negative relationship with post-traumatic stress (r = .60).

An exploratory principal components analysis with oblique rotation was performed on the 9 items from the CSE-T and 10 items of the PDEQ to examine discriminant validity. The Velicer's MAP test showed that there were two components among these items. Based on the result of the Velicer's MAP test, the number of components in the principal components analysis was set at two. Results showed that two components accounted for a total of 52.24% of the variance (eigenvalue = 9.93). One component consisted of the nine items of the CSE-T with factor loadings ranging from .62 to .80 (36.11% of the variance). The other component consisted of the 10 items of the PDEQ with factor loadings ranging from .33 to .88 (16.13% of the variance). The correlation between the two components was −.37. These results provide support for the discriminant validity of the CSE-T.

Confirmatory Factor Analysis

To examine factor loadings and a model fit for the 9-item version of the CSE-T, we conducted a confirmatory factor analysis in the hospital sample at Time 1 (n = 74). The initial analysis showed the model with uncorrelated error variances had acceptable goodness-of-fit, CFI = .948, RMSEA = .091, SRMR = .058, χ2/df = 1.61. Using the same model as in the hospital sample, the model had an inadequate fit in the disaster survivor sample (n = 273), CFI = .947, RMSEA = .100, SRMR = .042, χ2/df = 3.72. Thus, based on the modification indices, error variances of items 1 and 3 were covaried. After the modification, model fit was acceptable for the disaster survivor sample, CFI = .960, RMSEA = .089, SRMR = .039, χ2/df = 3.14. In the student sample (n = 242), results showed that the model was adequate with the same model as in the hospital sample, CFI = .943, RMSEA = .084, SRMR = .046, χ2/df = 2.68 (see Figure 1 for the model and factor loadings). Results indicated that the model was consistent across all samples, with the exception of correlated error variances for items 1 and 3 in Sample 2 (i.e., the disaster survivors).

Figure 1.

Figure 1

Open in a new tab

Factor Loadings in the Confirmatory Factor Analysis. Values for factor loadings are standardized coefficients. Values above the item names are intercepts. Values are in the order of the hospital/ disaster survivor/ undergraduate student samples. The covariance between error variances of items 1 and 3 is only present in the disaster survivor sample. All coefficients were significant at p < .001.

Test of Factor Invariance

To test the invariance of the model (i.e., consistency across different traumatic stress contexts) with the 9-item version of the CSE-T, we conducted the test for multigroup invariance across all samples (Byrne, 2001). Table 4 shows goodness-of-fit indices. Results of comparisons in the fit indices across all samples showed that the model with factor loadings constrained to be equal (model 2) did not significantly differ from the hypothesized model without any constraint (model 1); therefore, model 2 was accepted. However, further inspection of the comparisons of the fit indices indicated that the model with intercepts constrained to be equal (model 3) was significantly different from the hypothesized unconstrained model; therefore, model 3 was rejected. The model with variances constrained to be equal (model 4) was not significantly different from the hypothesized unconstrained model. Based on these results, the model with factor loadings and variances constrained to be equal (model 5) was compared to the hypothesized unconstrained model and results demonstrated no significant difference between these two models suggesting the factor loadings and variances were consistent across all samples.

Table 4.

Comparison of Fit Indices in the Test of Multigroup Invariance

Fit Index
Model χ 2 χ2/df CFI RMSEA ΔNFI Δ χ2
1. Hypothesized model (unconstrained) 197.53 2.47 .953 .050 - -
2. Factor loadings constrained to be equal 213.39 2.22 .953 .046 .006 15.86
3. Intercepts constrained to be equal 375.02 3.83 .889 .069 .068 177.49*
4. Variances constrained to be equal 216.28 2.21 .953 .045 .007 18.76
5. Factor loadings and variances constrained to be equal 233.70 2.05 .952 .042 .014 36.17
Open in a new tab

Note. df = degrees of freedom; CFI = comparative fit index; RMSEA = root mean square error of approximation; ΔNFI = a change in norm fit index from the hypothesized model without any constraints; Δ χ2 = a change in a chi-square statistic from the hypothesized model without any constraints. A significant Δ χ2 indicates a model that was not a good fit for the hypothesized model.

*

p < .001.

Discussion

Results of the present study support the psychometric soundness of the CSE-T. Within three different traumatic stress contexts, the CSE-T demonstrates good test-retest reliability, internal consistency, convergent, and discriminant validity. The CSE-T correlates in the hypothesized direction with important post-traumatic outcomes. The range of the longitudinal relationships between CSE-T and criterion validity outcomes demonstrates moderate to large effect sizes (range r = −.31 to −.74). Luszczynska et al. (2009) also reported large effect sizes for self-efficacy and critical outcomes following mass trauma situations.

Benight and Bandura (2004) argued that trauma exposure challenges individuals in severe and novel ways. These challenges serve as a spotlight on one's perceived capacity to cope. As demonstrated in their review (Benight & Bandura) and in a more recent meta-analysis (Luszczynska et al., 2009) individual perceptions of coping self-efficacy explain a large percentage of the variance in primary outcomes over time. The present results provide evidence that general trauma coping self-efficacy can longitudinally predict post-traumatic outcomes (both positive and negative).

Early CSE-T perceptions in our results show strong negative correlations with peritraumatic dissociation and subsequent psychological distress. This is consistent with Benight and Harper (2002) where a significant negative relationship between CSE and acute stress reactions was found. These results combined with our finding relating CSE-T with post-traumatic negative cognitions about one's self and the world suggests a constellation of cognitive processes that emerge early post-traumatically as the individual attempts to self-regulate. The CSE-T offers a way to help understand the self-regulation coping process with trauma as it evolves across time.

The strong positive association between CSE-T with psychological well-being is consistent with the expectation that one's belief in the ability to manage post-traumatic recovery should embolden one's sense of control. Enhanced control beliefs should foster a greater sense of well-being. The lack of relationship between CSE-T and post-traumatic growth (PTG), however, is contradictory to other research where a positive relationship was found between CSE and PTG for those with high post-traumatic stress (Cieslak et al., 2009). It is possible that the lack of relationship is the result of contradictory aspects (i.e., transcendent versus illusory) of PTG (Maerker & Zoellner, 2004). Future research is needed to shed clarify this finding.

Results of the present study support a one-factor structure. Factor analyses comparing the CSE-T across our samples indicate the invariant structure of the scale. A unidimensional structure is in line with SCT (Bandura, 1997). Unifactorial structure of other types of self-efficacy, such as general self-efficacy or more context specific self-efficacy (e.g., hurricane self-efficacy) also confirmed a unidimensional construct (Benight et al., 2004, 1999; Hyre et al., 2008; Lambert, Benight, Harrison, & Cieslak, 2011; Schwarzer & Jerusalem, 1995). Our results also confirm invariance of score variances across samples demonstrating equivalence of score dispersion. Interestingly, intercept invariance was not found suggesting that different samples may demonstrate unique initial levels of coping self-efficacy. This may have important clinical implications including utilizing the CSE-T as an acute screening measure for vulnerable populations (e.g., previously traumatized) (Benight, 2005).

The CSE-T demonstrates moderate (3-month) to high stability (< 3 months). One would expect CSE-T perceptions to vary over time (Bandura, 1997). Indeed, self-efficacy malleability is central to CSE-T as an intervention target. The large empirical literature supporting specific methods for enhancing self-efficacy perceptions via mastery, verbal persuasion, vicarious modeling, and physiological arousal management supports the utility of coping self-efficacy as an intervention focus (Bandura, 1997; Benight, 2005).

Important limitations of our results must be considered. First, the hospitalized sample has a relatively small sample size for a CFA. We chose this sample, however, due to its significant trauma exposure and distress levels. The model fit was acceptable in the other samples providing evidence for model stability. Second, the CSE-T was not evaluated against the Chesney et al. (2006) measure or a general coping self-efficacy assessment precluding any determination of incremental validity. And last, further research on the divergent validity of our measure is needed relative to other important constructs including general optimism and self-esteem.

In sum, the present results support the psychometric properties of the CSE-T scale and its theoretically assumed structure. Future studies are needed that compare the predictive power of the CSE-T against other relevant measures. Further, research is needed to test whether individuals scoring higher on the CSE-T actually demonstrate more effective coping. Indeed, more complete theoretically based coping models that including trauma exposure, CSE, coping behaviors, and important trauma outcomes will help to advance this literature. The ultimate value of this measure awaits future studies that focus on elucidating critical self-regulatory processes involved with overcoming life's most challenging experiences.

Acknowledgments

The disaster studies were supported by NIMH grant #1 R41MH082498-01 and NIMH grant #2 R42MH092498-03 to the first author.

Contributor Information

Charles C. Benight, Trauma, Health, and Hazards Center and Psychology Department, University of Colorado at Colorado Springs.

Kotaro Shoji, Trauma, Health, and Hazards Center, University of Colorado at Colorado Springs.

Lori E. James, Psychology Department, University of Colorado at Colorado Springs

Edward E. Waldrep, Psychology Department, Kent State University

Douglas L. Delahanty, Psychology Department, Kent State University

Roman Cieslak, Trauma, Health, and Hazards Center, University of Colorado at Colorado Springs and Psychology Department, University of Social Sciences and Humanities, Warsaw, Poland.

References

  1. Baker SP, O'Neill B, Haddon W, Jr, Long WB. The injury severity score: A method for describing patients with multiple injuries and evaluating emergency care. Journal of Trauma, Injury, Infection and Critical Care. 1974;14(3) Retrieved from http://trid.trb.org/view.aspx?id=133403. [PubMed] [Google Scholar]
  2. Bandura A. Self-efficacy: The exercise of control. Henry Holt; New York, NY: 1997. [Google Scholar]
  3. Benight CC. Clinical implications of coping self-efficacy in early interventions for trauma. Guidance and Counselling. 2005;21:6–12. [Google Scholar]
  4. Benight CC, Bandura A. Social cognitive theory of posttraumatic recovery: The role of perceived self-efficacy. Behaviour Research and Therapy. 2004;42(10):1129–1148. doi: 10.1016/j.brat.2003.08.008. doi:10.1016/j.brat.2003.08.008. [DOI] [PubMed] [Google Scholar]
  5. Benight CC, Cieslak R, Molton IR, Johnson LE. Self-evaluative appraisals of coping capability and posttraumatic distress following motor vehicle accidents. Journal of Consulting and Clinical Psychology. 2008;76(4):677–685. doi: 10.1037/0022-006X.76.4.677. doi:10.1037/0022-006X.76.4.677. [DOI] [PubMed] [Google Scholar]
  6. Benight CC, Freyaldenhoven RW, Hughes J, Ruiz JM, Zoschke TA, Lovallo WR. Coping self-efficacy and psychological distress following the Oklahoma City Bombing. Journal of Applied Social Psychology. 2000;30(7):1331–1344. doi:10.1111/j.1559-1816.2000.tb02523.x. [Google Scholar]
  7. Benight CC, Harding-Taylor AS, Midboe AM, Durham RL. Development and psychometric validation of a domestic violence coping self-efficacy measure (DVCSE). Journal of Traumatic Stress. 2004;17(6):505–508. doi: 10.1007/s10960-004-5799-3. doi:10.1007/s10960-004-5799-3. [DOI] [PubMed] [Google Scholar]
  8. Benight CC, Harper ML. Coping self-efficacy perceptions as a mediator between acute stress response and long-term distress following natural disasters. Journal of Traumatic Stress. 2002;15(3):177–186. doi: 10.1023/A:1015295025950. doi:10.1023/A:1015295025950. [DOI] [PubMed] [Google Scholar]
  9. Benight CC, Ironson G, Durham RL. Psychometric properties of a hurricane coping self-efficacy measure. Journal of Traumatic Stress. 1999;12(2):379–386. doi: 10.1023/A:1024792913301. doi:10.1023/A:1024792913301. [DOI] [PubMed] [Google Scholar]
  10. Blanchard EB, Jones-Alexander J, Buckley TC, Forneris CA. Psychometric properties of the PTSD Checklist (PCL). Behaviour Research and Therapy. 1996;34(8):669–673. doi: 10.1016/0005-7967(96)00033-2. doi:10.1016/0005-7967(96)00033-2. [DOI] [PubMed] [Google Scholar]
  11. Brewin CR, Rose S, Andrews B, Green J, Tata P, McEverdy C, Foa EB. Brief screening instrument for post-traumatic stress disorder. The British Journal of Psychiatry. 2002;181(2):158–162. doi: 10.1017/s0007125000161896. doi:10.1192/bjp.181.2.158. [DOI] [PubMed] [Google Scholar]
  12. Browne MM, Cudeck R. Alternative ways of assessing model fit. In: Bollen KA, Long JS, editors. Testing structural equation models. SAGE; Newbury Park, CA: 1993. [Google Scholar]
  13. Byrne BM. Structural equation modeling with AMOS: Basic concepts, applications, and programming. 2nd Edition Lawrence Erlbaum; Mahwah, NJ: 2001. [Google Scholar]
  14. Chesney MA, Neilands TB, Chambers DB, Taylor JM, Folkman S. A validity and reliability study of the coping self-efficacy scale. British Journal of Health Psychology. 2006;11(3):421–437. doi: 10.1348/135910705X53155. doi:10.1348/135910705X53155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Cieslak R, Benight CC, Lehman V. Coping self-efficacy mediates the effects of negative cognitions on posttraumatic distress. Behaviour Research and Therapy. 2008;46(7):788–798. doi: 10.1016/j.brat.2008.03.007. doi:10.1016/j.brat.2008.03.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Cieslak R, Benight C, Schmidt N, Luszczynska A, Curtin E, Clark RA, Kissinger P. Predicting posttraumatic growth among Hurricane Katrina survivors living with HIV: The role of self-efficacy, social support, and PTSD symptoms. Anxiety, Stress and Coping. 2009;22(4):449–463. doi: 10.1080/10615800802403815. doi:10.1080/10615800802403815. [DOI] [PubMed] [Google Scholar]
  17. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. Journal of Health and Social Behavior. 1983;24(4):385–396. doi:10.2307/2136404. [PubMed] [Google Scholar]
  18. Daraei M. Social correlates of psychological well-being among undergraduate students in Mysore City. Social Indicators Research. 2013;114(2):567–590. doi:10.1007/s11205-012-0162-1. [Google Scholar]
  19. Falsetti SA, Resnick HS, Resick PA, Kilpatrick DG. The Modified PTSD Symptom Scale: A brief self-report measure of posttraumatic stress disorder. The Behavior Therapist. 1993;16:161–162. [Google Scholar]
  20. Flatten G, Walte D, Perlitz V. Self-efficacy in acutely traumatized patients and the risk of developing a posttraumatic stress syndrome. GMS Psycho-Social Medicine. 2008;5 Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2736512/ [PMC free article] [PubMed] [Google Scholar]
  21. Foa EB, Ehlers A, Clark DM, Tolin DF, Orsillo SM. The Posttraumatic Cognitions Inventory (PTCI): Development and validation. Psychological Assessment. 1999;11(3):303–314. doi:10.1037/1040-3590.11.3.303. [Google Scholar]
  22. Hirschel MJ, Schulenberg SE. Hurricane Katrina's impact on the Mississippi Gulf Coast: General self-efficacy's relationship to PTSD prevalence and severity. Psychological Services. 2009;6(4):293–303. doi:10.1037/a0017467. [Google Scholar]
  23. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal. 1999;6(1):1–55. doi:10.1080/10705519909540118. [Google Scholar]
  24. Hyre AD, Benight CC, Tynes LL, Rice J, DeSalvo KB, Muntner P. Psychometric properties of the Hurricane Coping Self-Efficacy Measure following Hurricane Katrina. The Journal of Nervous and Mental Disease. 2008;196(7):562–567. doi: 10.1097/NMD.0b013e31817d016c. doi:10.1097/NMD.0b013e31817d016c. [DOI] [PubMed] [Google Scholar]
  25. Lambert JE, Benight CC, Harrison E, Cieslak R. The Firefighter Coping Self-Efficacy Scale: Measure development and validation. Anxiety, Stress & Coping. 2011:1–13. doi: 10.1080/10615806.2011.567328. doi:10.1080/10615806.2011.567328. [DOI] [PubMed] [Google Scholar]
  26. Luszczynska A, Benight CC, Cieslak R. Self-efficacy and health-related outcomes of collective trauma: A systematic review. European Psychologist. 2009;14(1):51–62. doi:10.1027/1016-9040.14.1.51. [Google Scholar]
  27. Marmar CR, Metzler TJ, Otte C. The Peritraumatic Dissociative Experiences Questionnaire. In: Wilson JP, Keane TM, editors. Assessing psychological trauma and PTSD. 2nd ed. Guilford; New York, NY: 2004. pp. 144–167. [Google Scholar]
  28. Maercker A, Zoellner T. The Janus face of self-perceived growth: Toward a two-component model of posttraumatic growth. Psychological Inquiry. 2004;15:41–48. [Google Scholar]
  29. Meyer TJ, Miller ML, Metzger RL, Borkovec TD. Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy. 1990;28(6):487–495. doi: 10.1016/0005-7967(90)90135-6. doi:10.1016/0005-7967(90)90135-6. [DOI] [PubMed] [Google Scholar]
  30. O'Connor BP. SPSS and SAS programs for determining the number of components using parallel analysis and Velicer's MAP test. Behavior Research Methods, Instruments, and Computers. 2000;32(3):396–402. doi: 10.3758/bf03200807. doi:10.3758/BF03200807. [DOI] [PubMed] [Google Scholar]
  31. Ozer EJ, Best SR, Lipsey TL, Weiss DS. Predictors of posttraumatic stress disorder and symptoms in adults: A meta-analysis. Psychological Trauma: Theory, Research, Practice, and Policy. 2008;S(1):3–36. doi: 10.1037/0033-2909.129.1.52. doi:10.1037/1942-9681.S.1.3. [DOI] [PubMed] [Google Scholar]
  32. Park CL, Ai AL. Meaning making and growth: New directions for research on survivors of trauma. Journal of Loss and Trauma. 2006;11(5):389–407. doi:10.1080/15325020600685295. [Google Scholar]
  33. Radloff LS. The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement. 1977;1(3):385–401. doi:10.1177/014662167700100306. [Google Scholar]
  34. Ryff CD. Happiness is everything, or is it? Explorations on the meaning of psychological well-being. Journal of Personality and Social Psychology. 1989;57(6):1069–1081. doi:10.1037/0022-3514.57.6.1069. [Google Scholar]
  35. Schwarzer R, Jerusalem M. Optimistic self-beliefs as a resource factor in coping with stress. In: Hobfoll SE, deVries MW, editors. Extreme stress and communities: Impact and intervention. Kluwer Academic/Plenum Publishers; New York, NY, US: 1995. pp. 159–177. [Google Scholar]
  36. Solomon Z, Benbenishty R, Mikulincer M. The contribution of wartime, pre-war, and post-war factors to self-efficacy: A longitudinal study of combat stress reaction. Journal of Traumatic Stress. 1991;4(3):345–361. doi:10.1007/BF00974554. [Google Scholar]
  37. Steinmetz S, Benight CC, Bishop S, James L. My Disaster Recovery: A pilot randomized controlled trial of an internet intervention. Anxiety, Stress & Coping. 2012;25:593–600. doi: 10.1080/10615806.2011.604869. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Stevens J. Applied multivariate statistics for the social sciences. 5th Edition Taylor & Francis; 2009. [Google Scholar]
  39. Teasdale G, Jennett B. Assessment of coma and impaired consciousness. Lancet. 1974;304(7872):81–84. doi: 10.1016/s0140-6736(74)91639-0. doi:10.1016/S0140-6736(74)91639-0. [DOI] [PubMed] [Google Scholar]
  40. Tedeschi RG, Calhoun LG. The Posttraumatic Growth Inventory: Measuring the positive legacy of trauma. Journal of Traumatic Stress. 1996;9(3):455–471. doi: 10.1007/BF02103658. doi:10.1007/BF02103658. [DOI] [PubMed] [Google Scholar]
  41. Van der Kolk BA, McFarlane AC, Weisaeth L. Traumatic stress: The effects of overwhelming experience on mind, body, and society. Guilford; New York, NY: 1996. [Google Scholar]
  42. Velicer WF, Eaton CA, Fava JL. Construct explication through factor or component analysis: A review and evaluation of alternative procedures for determining the number of factors or components. In: Goffin RD, Helmes E, editors. Problems and solutions in human assessment. Springer; New York, NY: 2000. pp. 41–71. Retrieved from http://link.springer.com/chapter/10.1007/978-1-4615-4397-8_3. [Google Scholar]
  43. Zlomke KR. Psychometric properties of internet administered versions of Penn State Worry Questionnaire (PSWQ) and Depression, Anxiety, and Stress Scale (DASS). Computers in Human Behavior. 2009;25(4):841–843. doi:10.1016/j.chb.2008.06.00. [Google Scholar]

RESOURCES