Coping Self-Efficacy and Psychological Distress Following ...

Coping Self-Efficacy and Psychological Distress Following the Oklahoma City Bombing'

CHARLES C. BENIGHT? ROBERT W. FREYALDENHOVEN, JOEL HUGHES, JOHN M. RUIZ, A N D TIFFANY A. ZOSCHKE

Univer.yity of Colorado at Colorado Spring.s

WILLIAM R. LOVALLO University of Oklolioma Health Sciences Center and Department of C'eteran.c Afl'irs

Findings from a study of 27 victims of the Oklahoma City bombing are reported. This research focused on the importance of subjective appraisals of coping self-efficacy in pre- dicting psychological distress following this tragedy. Results supported the hypothesis that judgments of coping self-eficacy taken 2 months after the bombing added significantly to the explanation of general and trauma-related distress after controlling for income, social support, threat of death, and loss of resources. Coping self-eficacy judgments taken I year later were also important in explaining psychological distress after controlling for loss of resources and social-support perceptions. Although coping self-efficacy percep- tions taken at 2 months were related to distress levels I year later, they did not remain sig- nificant in a regression analysis controlling for loss of resources and income. Implications of these findings for post-terrorist bombing interventions are discussed.

At 9:02 a.m. on April 19, 1995, the world was suddenly shocked by one of the worst terrorist acts in United States history. One hundred sixty-eight people were killed and hundreds more were injured as a terrorist bomb ripped through the Alfred P. Murrah Federal Building in Oklahoma City, Oklahoma. Research on the mental health impact of disasters has rapidly proliferated in the last several decades and has demonstrated that serious psychological ramifications can occur following a variety of disasters (Adams & Adams, 1984; Rubonis & Bickman, 1991) and terrorist attacks in particular (Shalev, 1992). The primary aim of this research is to evaluate the importance of subjective appraisals of coping self- efficacy in predicting psychological distress after this tragedy.

Coping self-efficacy (CSE) is defined as a person's subjective appraisal of his or her ability to cope with the environmental demands of a stressful or traumatic

I We would like to thank all ofthe victims ofthis tragedy who gave oftheir time and effort to help us complete this study. We would also like to thank Sara Jo Nixon for her help in completing this research.

Torrespondence concerning this article should be addressed to Charles C. Benight, Department of Psychology, University of Colorado at Colorado Springs, Colorado Springs, CO 80933-71 50. e- mail: [email protected].

1331

Journal ofApplied Social Psychology, 2000, 30, 7 , pp. 1331-1344. Copyright 0 2000 by V. H. Winston 8 Son, Inc. All rights reserved.

1332 BENIGHT ET AL.

situation (Bandura, 1986). Research on CSE judgments taken in the early after- math of natural disasters has been supportive of the importance of these judg- ments as a predictor of psychological distress (Benight et al., 1997; Benight et al., 1999; Benight, Swift, Sanger, Smith, & Zeppelin, 1999; Murphy, 1987). Murphy demonstrated that individuals with higher levels of global self-efficacy reported lower levels of psychological distress following the Mount St. Helens volcano disaster. Benight and colleagues have looked at more specific sel f-effi- cacy judgments tied directly to the coping demands of the disaster environment. These studies have shown that CSE is an important variable i n predicting psy- chological distress and possibly physical distress following natural disasters cross-sectionally and longitudinally.

In healthy community volunteers, Benight. Ironson, et al. (1999) and Benight, Swift, et al. (1999) found that higher specific CSE judgments were associated with lower reports of psychological distress following Hurricane Andrew and Hurricane Opal, respectively. In the Hurricane Andrew study, CSE judgments accounted for an additional 10% of the variance after controlling for income, education, gender, age, estimated damage, threat of death, and lost resources. In addition, CSE demonstrated a direct pathway to psychological dis- tress acutely and to psychological distress 9 months later in a causal model anal- ysis. The results for the Hurricane Opal study were similar, with CSE judgments accounting for an additional 26% of the variance for Impact of Event Scale (IES; Horowitz, Weiner, & Alvarez, 1979) scores after controlling for lost resources, threat of death, gender, and social support. This study also demon- strated that CSE judgments served as a mediator between several important fac- tors (e.g., lost resources, social support, optimism) and distress. Benight et al. (1997) looked at CSE in a sample of HIV-infected men following Hurricane Andrew and also found positive results. CSE judgments accounted for an addi- tional 5 1% of the variance over and above CD4 cell counts, threat of death, esti- mated damage, income, and education in predicting post-traumatic stress disorder (PTSD) symptoms. CSE also explained an additional 28% of the vari- ance in predicting a hormonal ratio (norepinephrine/cortisol) associated with PTSD.

These studies suggest that individuals’ CSE judgments during the early recovery period following a disaster are potentially important in post-disaster recovery. However, what remains unclear is whether CSE is important outside the domain of natural disasters. The only studies identified that have looked at CSE judgments and psychological distress related to trauma outside natural disasters have been completed by Solomon and colleagues (Solomon, Benben- ishty, & Mikulincer, 1991 ; Solomon, Weisenberg, Schwarzwald, & Mikulincer, 1988). Results from these studies show that low CSE judgments for military combat were correlated with greater PTSD symptoms and general psychological distress at 1 and 2 years following the war.

COPING SELF-EFFICACY AND TERRORISM 1333

The present study extends past research by evaluating the importance of CSE in psychological recovery from a domestic terrorist attack. We hypothesize that CSE judgments taken within the 2 months following the bombing will explain a significant proportion of the variance in ITSD symptoms and general psychologi- cal distress over and above several control variables (e.g., social support, threat of death). In addition, we hypothesize that CSE judgments at 1 year following the bombing will also explain a significant proportion of the variance after control variables. Finally, we expect that CSE judgments sampled 2 months after the bombing will be an important predictor for distress measures taken 1 year after the attack. These hypotheses were tested using hierarchical multiple regression anal- yses. Threat of death, social support, loss of resources, and income were chosen as control variables for the Time 1 analysis based on previous literature demonstrat- ing these to be important factors in disaster recovery. For the Time 2 analysis and the longitudinal regression, only the two most strongly zero-order-correlated fac- tors were chosen as control variables because of the reduced sample size.

Method

Participants

A convenience sample of 27 victims was recruited 2 months after the bomb- ing. These participants were found through local businesses that were directly across the street from the Murrah Federal Building. This sampling strategy was chosen because individuals within this range of the bomb would have experi- enced the most trauma other than those who actually survived within the federal building. Nearly all participants (91%) reported that they felt the explosion, and 83% reported hearing it. All of them were evacuated from their workplaces. The participants responded to the statement, “I thought I might die” on a 7-point scale ranging from I (not at af f ) to 7 (absofu/efy)-the mean value was 3.7. This sug- gests that this group felt that there was a moderate chance that they might die during the explosion. Of the 27 participants, 10 reported either a 6 or a 7 on this question. A large percentage (67%) reported some type of injury that required medical attention as a result of the bombing. These injuries ranged from cuts and bruises to hearing difficulties and skull fractures. Of those who were injured, 20% required hospitalization as a result of their injuries. Three quarters of the participants knew others who were injured or killed as a result of the bombing. Thus, based on these data, this sample was highly exposed to this tragedy.

The Time 1 means for the Global Severity Index (GSI [Derogatis, 19831; M = 0.71) and the IES (Horowitz et al., 1979; M = 37.4) showed that this sample was relatively distressed. Based on the published norms for the GSI, the mean of 0.71 is lower than psychiatric outpatients (M = 1.26), and above the nonpatient mean (M= 0.31). This places the sample mean at the 90th percentile for the nonpatient

1334 BENIGHT ET AL.

norm group and at the 21st percentile, compared to the psychiatric outpatient group. When looking at the mean of 37.4 for the IES, this is slightly lower than the published norm values for stress-clinic patients (Horowitz et al., 1979) with a mean of 38.7, yet substantially higher than the medical-student norm group ( M = 9.8). The mean value of 37.4 for this sample is also comparable to those reported by Israeli soldiers with PTSD following the Lebanon war ( M = 35.1; Solomon, 1989).

The mean age of this sample was 41 years. Of these individuals, 13 partici- pants (48%) were men and 14 (52%) were women. The median income range reported for sample at Time 1 was between $40,000 and $45,000 per year. Educa- tionally, 2 (7%) reported a high school education, 7 (26%) some college, 10 (37%) college graduate, and 8 (30%) graduate education. Ethnically, almost the entire sample was Caucasian, with only I (3%) African American and 1 (3%) Native American.

Procedure

At Time I , participants completed a psychosocial questionnaire and a semi- structured interview lasting between 1 hr and 1.5 hr. Approximately 1 year after the bombing, the participants were sent a questionnaire packet and a self- addressed stamped envelope to be returned in the mail. Out of the initial 27 par- ticipants, 17 returned the second questionnaire for a response rate of 63%. An independent t test was conducted testing responders from nonresponders on all three psychological distress outcome measures. No statistical differences were identified. However, the nonresponder group averages were consistently higher than those of the responder group.

Measures

CSE. CSE was measured using a 43-item scale designed to assess the main situational demands of coping with a bombing. Items were divided into four main areas of coping demands: behaviors, thoughts, images, and emotions. Under each of these areas, specific demands that participants were facing were assessed. The items were all answered on a 7-point Likert-type scale ranging from 1 (not at all capable) to 7 (total[y capable). The items were related to issues such as doing my job well, managing thoughts of people dying, thoughts of personal injury. dis- tressing dreams, feelings of helplessness, and so forth. The internal reliability coefficients for this measure were .96 at Time 1 and .97 at Time 2.

L$e threat. Participants were asked whether they thought they were going to die during the bombing. Answers ranged from 1 (nor at all) to 7 (absolutely).

Loss ofresources. Loss of resources was assessed in line with Hobfoll’s ( 1989) model of stress. Participants answered 40 questions on a 5-point scale

COPING SELF-EFFICACY AND TERRORISM 1335

ranging from 0 (no loss) to 4 (extreme amount of loss). The list of resources utilized for this study related to material resources (e.g., car, furniture) and experiential resources (e.g.. time for sleep, daily routine). The score for lost resources is determined by summing the items.

Social support. The Interpersonal Support Evaluation List (ISEL; Cohen, Mermelstein, Kamarck, & Hoberman, 1985) was utilized to measure social sup- port. The ISEL is a 40-item assessment that measures four main areas of social support. These include tangible support, belonging support, appraisal support, and self-esteem support. Reliability and validity have been supported for this measure (Cohen et al., 1985). Respondents answer in a TrueIFalse format. An overall summed score was utilized in this study as a global measure of social- support perceptions.

Symptom Checklist-90, revised (SCL-90R). The SCL-90R (Derogatis, 1983) was used to assess psychological distress following the bombing. The SCL-90R is a 90-item self-report measure designed to assess general psychological distress in psychiatric and medical patients. The questions relate to the past week, and an average level of overall psychological distress is attained (i.e., GSI). The SCL- 90R has been utilized in several studies investigating psychological reactions to trauma (Baum, Gatchel, & Schaeffer, 1983). Adequate reliability and validity have been reported (Derogatis, 1983).

PTSD sympromatology. PTSD was assessed utilizing the frequency of experi- encing 17 PTSD symptoms derived from the American Psychiatric Association's (APA's) revised third edition of the Diagnostic and Statistical Manual (DSM-III- R; APA, 1987). Participants were asked to assess how often over the past week they have experienced the series of symptoms associated with PTSD diagnosis (e.g., intrusive thoughts, avoidant response, hyperarousal). Participants indicated how often they have experienced each symptom on a 5-point scale ranging from 0 (not at all) to 4 (every day). A summed score for the frequency of all of the symptoms was utilized as a measure of PTSD symptom reporting. For Time 2 participants, we added a severity assessment that was rated on a 5-point scale ranging from 0 (nor at all distressing) to 4 (exrremely distressing) for each symp- tom experienced.

ZES. The IES (Horowitz et al., 1979) is a 15-item scale that assesses the emo- tional impact of a traumatic event on a person by looking at intrusive thoughts, emotional numbing, and avoidance. Respondents answer items related to the fre- quency of each symptom experienced on a 4-point scale ranging from 1 (nor a/ all) to 4 (ofren). This scale has shown adequate reliability and validity and is extensively used in trauma research (Horowitz, Field, & Classen, 1993).

Results

Table 1 depicts the Time I and Time 2 descriptive data for the study variables. Utilizing only those participants who were sampled at Time 1 and Time 2, paired t

1336 BENIGHT ET AL.

Table 1

Means and Standard Deviations on Demographic and Primary Variables for Time I and Time 2

Time 1 Time 2

Variable M SD M SD

Age Educationa Incomea Threat of death Social support Loss of resources Coping self-efficacy Global Severity Index Impact of Event Scale PTSD symptom frequency PTSD symptom severity

40.7 10.2 4.0 2-5 9.0 4-13 3.7 2.5

32.6* 6.5 29.0* 19.5

202.3 40.2

37.4 9.3 24.4' 14.4

0.7 1 0.57

42.3 4.0 9.3 -

26.6 18.4

208.5 0.50

34.8 16.24* 14.93

9.6 2-5 5-13 -

6.8 16.9 34.2 0.34 8.6 9.0 9.7

Note. For income, the ranges were in $5,000 increments with I = under $5.000 to I 3 = over $60.000. For education. I = some high school, 2 = high-school graduate, 3 = some college, 4 = college graduate, 5 = graduate degree. avalues presented are the median and range. ' p < .05.

tests were performed on social support, lost resources, CSE, GSI, IES, and PTSD frequency. The results show that participants reported significantly less PTSD symptom frequency at Time 2 compared to Time 1 . In addition, participants reported significantly less social support at Time 2 than at Time 1. Finally, indi- viduals indicated lower perceived resource loss at Time 2 than at Time 1. Interest- ingly, perceptions of CSE did not change significantly from Time 1 to Time 2.

Zero-order correlations among predictor and outcome variables are provided in Table 2. The hierarchical multiple regressions (Tables 3 and 4) show that CSE perceptions taken 2 months following the bombing accounted for an additional 23% of the variance in predicting general psychological distress after the control variables of threat of death, income, social support, and loss of resources, F( I , 21) = 27.59, p < .OOO. CSE added 22% of the variance for the IES total score, which was significant, F( 1,2 1) = 11.85, p < .002. Finally, CSE added 28% of the variance over and above control variables for frequency of PTSD symptoms at Time I , F( I , 2 1 ) = 46.27, p < .OOO.

COPING SELF-EFFICACY AND TERRORISM 1337

At 1 year after the bombing, CSE judgments added a significant 14%, 16%, 12%, and 26% of the variance in explaining general distress, IES total scores, PTSD symptom frequency, and PTSD symptom severity, respectively (Table 4). The explanatory contribution of CSE was over and above lost resources at Time 2 and social-support perceptions at Time 2.

In a very preliminary assessment of the predictive value of CSE judgments from early after the bombing in explaining psychological distress 1 year later, we ran another hierarchical regression with Time 1 CSE, Time 1 lost resources, and income. CSE judgments taken at Time 1 were not found to be predictive of Time 2 distress over and above lost resources and income.

Discussion

The present findings, though generally supportive, should be viewed as very preliminary because of multiple limitations of this study. The most critical limita- tion of the study is the select sample. The study participants were not randomly selected, and the demographics of the sample are relatively homogeneous. Thus, the findings from this research may not generalize to other terrorist attacks or other populations. Future research needs to replicate these initial results in a larger and more representative sample. In addition to the select nature of the sam- ple, the sample size is very limited. Finally, the primary results of the study are cross-sectional, and causation cannot be inferred. I t is unclear whether it is the perception of CSE that is driving increased distress or the other way around.

With these limitations in mind, these data do provide an initial glimpse into the association of individual self-regulatory mechanisms (i.e., CSE) with psycho- logical distress following a terrorist attack. The findings, though limited, are con- sistent with data on CSE from natural disasters (Benight et al., 1997; Benight, Ironson, et al., 1999; Benight, Swift, et al., 1999; Murphy, 1987) and from com- bat soldiers (Solomon et al., 1988, 1991). The data suggest that perceived CSE to cope with the demands following a terrorist bombing is associated with reported psychological distress. Higher perceptions of CSE are related to lower experi- ences of distress.

These data also support the role of resource depletion as an important factor in response to trauma (Hobfoll, 1991). The predictor of lost resources was the only significant predictor in the regression equations, other than CSE. The greater the level of the reported resource devastation, the greater was the distress. These findings are consistent with findings reported by Freedy, Hobfoll, and Ribbe ( I 994) where war-related resource loss was related to increases in psycho- logical distress.

The identified drop in social-support perceptions from Time I to Time 2 is also intriguing. This may demonstrate the complexity involved in social-support perceptions following a severe tragedy (Norris & Kaniasty, 1996). Often, early

1338 BENIGHT ET AL.

0

S I - 9 - 1 0 9 9 0 0 0 0 c o

7. S

ocia

lsup

port

,256

,4

51

-.I9

2 -.

I90

,198

.I

04

1.00

0

.198

.0

69

.461

.3

42

.323

.6

91 -

Tim

e 1

8. S

ocia

l sup

port

.338

.5

56

-.459

-.2

78

,421

,0

11

,652

1.

000

.185

.0

03

.016

.2

81

.092

.9

57

,005

-

Tim

e 2

9. L

ostre

sour

ces

-.487

-.6

27

,612

.5

74

-.325

-.3

05

-.502

-.4

89

1.00

0

.O 10

.0

07

.009

.0

02

.098

.2

35

.008

.0

47 -

Tim

e 1

10. L

ostre

sour

ces

-.230

-.4

56

,521

,5

18

-.586

-.

I42

-.55

0 -.4

03

.860

1.

000

.375

.0

17

.005

.0

33

.013

.4

99

.022

.0

37

.OOO

-

Tim

e 2

v) rn 7 rn

n

n_

0

D

0

.030

.O

OO

.OOO

.0

02

.023

.7

13

.452

.0

03

.003

.0

03

.002

-

<

11.

PTSD

freq

uenc

y -.8

44

-.781

,6

66

,788

-.3

52

-.I9

7 -.4

15

-.371

.7

56

.660

1.

000

r

.OOO

.O

OO

.004

.O

OO

.072

.4

49

.03 1

.1

42

.OOO

.0

04 -

Tim

e 1

12. P

TSD

freq

uenc

y -.5

26

-.794

,8

55

,702

-.5

48

-.077

-.

I96

-.55

0 ,6

79

,555

,7

01

1.00

0 Ti

me 2

$ 13

. PT

SDse

verit

y -.5

71

-.860

,8

34

,721

-.5

14

-.059

-.2

46

-.533

,7

67

,656

,7

44

.927

1.

000

0

-I rn 2

.026

.O

OO

.OOO

.0

02

.050

.7

90

.376

.0

06

.001

.O

OO

.OO 1

.OOO

-

Tim

e 2

AJ

Not

e. T

ime

I, N

= 27

. Tim

e 2,

N=

17. I

talic

ized

item

s rep

rese

nt co

rrel

atio

ns, n

onita

liciz

ed it

ems r

epre

sent

sig

nific

ance

val

ues.

1340 BENIGHT ET AL.

Table 3

Multiple Regression With Control Variables and CSE for Each Dependent Variable at Time I

Variable P t p AR2 F df p ~~

Dependent variable: Global Severity Index, Time 1 Threat of death 0.12 1.30 0.21 Income 0.01 0.1 I 0.91 Social support -0.05 -0.51 0.62 Loss ofresources 0.44 3.68 0.00 0.60 8.20 4, 22 .OO CSE -0.56 -5.25 0.00 0.23 27.59 I , 21 .OO

Dependent variable: Impact of Event Scale, Time 1 Threat of death 0.16 1.15 0.26 Income 0.23 1.52 0.14 Social support 0.12 0.77 0.45 Loss ofresources 0.41 2.29 0.03 0.38 3.38 4, 22 .03 CSE -0.60 -3.44 0.00 0.22 11.85 1,21 .OO

Dependent variable: PTSD symptom frequency, Time 1

Threat of death -0.03 -0.42 0.67 Income 0.03 0.36 0.72 Social support -0.04 -0.42 0.68 Loss ofresources 0.44 4.34 0.00 0.59 7.80 4, 22 .OO CSE -0.63 -6.80 0.00 0.28 46.27 I , 21 .OO

Note. Standardized regression weights are final weights with all variables in equation

after severe disasters there is an influx of external environmental support offset- ting the internal social-support deterioration caused by the tragedy. However, over time, there may be a drop in social-support perceptions because of a per- ceived discrepancy between anticipated social support and actual response received (Kaniasty, Norris, & Murrell, 1990). This may also be related to the type of tragedy that these individuals experienced. A natural disaster such as a hurri- cane has clear environmental effects (e.g., house damage) where people can pro- vide direct assistance. The Oklahoma City tragedy was much more emotionally based, making social support perhaps more interpersonally based. More research on perceived support following large-scale emotional tragedies is needed.

COPING SELF-EFFICACY AND TERRORISM 1341

Table 4

Multiple Regression With Control Variables and Time 2 CSE for All Dependent Variables at Time 2

~

P M2 F df p Variable P t

Dependent variable: Global Severity Index, Time 2 Social support -0.07 -0.38 .71 Loss of resources 0.44 2.44 .03 .54 8.13 2, 14 .01 CSE -0.47 -2.37 .03 .I4 5.60 1, 13 .03

Dependent variable: Impact of Event Scale, Time 2 Social support 0.02 0.11 .92 Loss of resources 0.38 1.80 .10 .40 4.72 2, 14 .03 CSE -0.51 -2.22 .05 .I6 4.93 1, 13 .05

Dependent variable: PTSD symptom frequency, Time 2 Social support -0.06 -0.28 .78 Loss of resources 0.49 2.65 .02 .55 8.52 2, 14 .OO CSE -0.43 -2.11 .06 .12 4.46 1, 13 .06

Dependent variable: PTSD symptom severity, Time 2 Social support 0.09 0.56 .59 Loss of resources 0.47 3.06 .01 .54 7.12 2, 12 .01 CSE -0.63 -3.80 .OO .26 14.41 1, 11 .OO

Note. Standardized regression weights are final weights with all variables in equation.

Before addressing the implications of these findings, i t is important to question the possible redundancy of the CSE measure with distress. Based on the bivariate correlations that showed CSE to be correlated higher with the GSI (r = -30) at Time 1 than the IES was with the GSI ( r = .69), it could be argued that the CSE measure is simply a redundant measure of distress. In order to test this hypothesis, partial correlations were computed between CSE and two of the distress outcome variables (GSI and PTSD), while controlling for the distress associated with the IES. The rationale for these analyses was that if CSE were still significantly related to these distress variables after removing the shared variance of distress, then it is contributing something unique. Importantly, the correlations between CSE and Time 1 GSI and FTSD remain significant even after removing the distress associated with the IES (GSI: partial r = -.63, p = .001; PTSD: partial r = -.70,p < .OOO). CSE Time 2 correlations with GSI, PTSD frequency, and PTSD severity, while controlling for Time 2 IES scores, also remained significant (GSI: partial r = -.57, p = .002; PTSD frequency: r = -.49,

1342 BENIGHT ET AL.

p = .01 I ; F'TSD severity: partial r = -.67, p < .OOO). Thus, it appears that the CSE measure for this study is adding to our understanding of distress, rather than sim- ply serving as a redundant measure of distress.

lmplications of these findings are offered with the caveat that future studies are clearly needed to substantiate or refute the present results. The measurement of CSE provides a unique view of individual self-perceptions of coping capabil- ity across a wide range of functioning, including thought control, image control, behavioral action, and emotional containment. This type of information may help to provide insight into areas of need where a person feels highly inefficacious. For example, one individual may feel especially incompetent in managing work demands after the bombing, whereas another individual may be struggling more with managing feelings of loss related to the trauma. This type of information may be more informative for treatment strategies than relying only on distress indicators such as F'TSD symptoms.

In addition to providing potential direction for interventions. CSE perceptions may interact with certain interventions to produce negative outcomes. The cur- rent postdisaster interventions, often titled critical incident stress debriefing (CISD), have focused on a psychoeducational process, including a recounting of the tragedy within a group format. CSE perceptions may interact with this type of intervention with some at-risk individuals feeling more distress when asked to discuss in detail his or her experience with the tragedy (i.e., feeling highly inefi- cacious to deal with memories from the bombing or thoughts from the bombing). Indeed, because of the emotional and physiological reactivity linked with poor CSE perceptions, recounting the experience may further exacerbate self-apprais- als of inability to cope (Bandura, 1986). Future studies are needed before further recommendations or implications can be offered.

References

Adams, P. R., & Adams, G. R. (1984). Mount Saint Helens' ashfall: Evidence for a disaster stress reaction. American Psychologist, 39,252-260.

American Psychiatric Association. ( 1987). Diagnostic and statistical manlrcd of mental disorders (3rd ed., rev.). Washington, DC: Author.

Bandura, A. (1 986). Socia1,foundations of thought and action. Englewood Cliffs, NJ: Prentice-Hall.

Baum, A., Gatchel, R. J., & Schaeffer, M. A. ( 1 983). Emotional, behavioral, and physiological effects of chronic stress at Three Mile Island. Journal qf Con- sulting and Clinical Psychology, 51, 565-572.

Benight, C. C., Antoni, M. H.. Kilbourn, K., Ironson, G., Kumar, M. A., Fletcher, M. A., Redwine, L., Baum, A., & Schneiderman, N. (1997). Coping self- eficacy buffers psychological and physiological disturbance in HIV-infected men following a natural disaster. Health Psycholoa, 16, 248-255.

COPING SELF-EFFICACY AND TERRORISM 1343

Benight, C. C., Ironson, G., Wynings, C., Klebe, K., Burnett, K., Greenwood, D., Carver, C. S., Baum, A., & Schneideman, N. (1999). Conservation of resources and coping self-efficacy predicting distress following a natural disaster: A causal model analysis where the environment meets the mind. Anxiety, Stress, and Coping, 12, 107- 126.

Benight, C. C., Swift, E., Sanger, J., Smith, A., & Zeppelin, D. (1999). Coping self-efficacy as a mediator of distress following a natural disaster. Journal of Applied Social Psychology, 29,2443-2464.

Cohen, S., Mermelstein, R., Kamarck, T., & Hoberman, H. M. (1985). Measuring the functional components of social support. In 1. G. Sarason & B. R. Sarason (Eds.), Social support: Theoty, research, and applications (pp. 73-94). Boston, MA: Martinus Nijhoff.

Derogatis, L. R. ( 1983). The SCL-9OU manual: Scoring administration. andpro- cedures for the SCL-90R. Baltimore, MD: John Hopkins University School of Medicine, Clinical Psychometrics Unit.

Freedy, J. R., Hobfoll, S. E., & Ribbe, D. P. (1994). Life events, war, and adjust- ment: Lessons for the Middle East. Anxiety, Stress, and Coping, 7, 191-203.

Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualiz- ing stress. American Psychologist, 44, 5 13-524.

Hobfoll, S. ( I 99 I ). Traumatic stress: A theory based on rapid loss of resources. Anxiety Research, 4, 187- 197.

Horowitz, M. J., Field, N. P., & Classen, C. C. (1993). Stress response syndromes and their treatment. In L. Goldberger & S. Breznitz (Eds.), Handbook qf stress: Theoretical and clinical aspects (pp. 757-773). New York. NY: The Free Press.

Horowitz, M., Weiner, N., & Alvarez, W. (1979). Impact of Event Scale: A mea- sure of sub-jective stress. Psychosomatic Medicine, 41, 209-2 18.

Kaniasty, K. Z., Norris, F. H., & Murrell, S. A. (1990). Received and perceived social support following natural disaster. Journal of Applied Social Psychol-

Murphy, S. (1987). Self-efficacy and social support mediators of stress on mental health following a natural disaster. Western Journal of Nursing Research, 9,

Norris, F. H., & Kaniasty, K. Z. (1996). Received and perceived social support in times of stress: A test of the social support deterioration deterrence model. Journal of Personality and Social Psychologv, 71,498-5 1 1.

Rubonis, A. V., & Bickman, L. (1991). Psychological impairment in the wake of disaster: The disaster-psychopathology relationship. Psychological Bulletin,

Shalev, A. Y. (1992). Postraumatic stress disorder among injured survivors of a terrorist attack: Predictive value of early intrusion and avoidance symptoms. Journal of Nervous and Mental Disease, 180,505-509.

OD, 20, 85- 1 14.

58-86.

109, 384-399.

1344 BENIGHT ET AL.

Solomon, Z. ( 1 989). Psychological sequelae of war. Journal of Nervorrs and Mental Disease, 177,342-346.

Solomon, Z., Benbenishty, R., & Mikulincer, M. (1991). The contribution ofwar- time, pre-war, and post-war factors on self-efficacy: A longitudinal study of combat stress reaction. Journal of Traumatic Stress. 4, 345-361.

Solomon, Z., Weisenberg, M., Schwarzwald, J . , & Mikulincer, M. (1988). Com- bat stress reaction and posttraumatic stress disorder as determinants of per- ceived self-efficacy in battle. Journal of Social and Clinical Psychologv, 6, 356-370.