Content uploaded by Stevan E Hobfoll
Author content
All content in this area was uploaded by Stevan E Hobfoll on Sep 08, 2015
Content may be subject to copyright.
Content uploaded by Stevan E Hobfoll
Author content
All content in this area was uploaded by Stevan E Hobfoll on Sep 08, 2015
Content may be subject to copyright.
NOTICE CONCERNING
COPYRIGHT RESTRICTIONS
The copyright law of the United States [Title 17, United
States Code] governs the making of photocopies or other
reproductions of copyrighted material.
Under certain conditions specified in the law, libraries and
archives are authorized to furnish a photocopy or other
reproduction. One of these specified conditions is that the
reproduction is not to be used for any purpose other than
private study, scholarship, or research. If a user makes a
request for, or later uses, a photocopy or reproduction for
purposes in excess of “fair use” that use may be liable for
copyright infringement.
The institution reserves the right to refuse to accept a
copying order if, in its judgment, fulfillment of the order
would involve violation of copyright law. No further
reproduction and distribution of this copy is permitted by
transmission or any other means.
The Combined Stress of Family Life, Work, and
War in Air Force Men and Women: A Test of
Conservation of Resources Theory
Stevan E. Hobfoll
Rush Medical College
Amiram D. Vinokur, Penny F. Pierce, and
Lisa Lewandowski-Romps
University of Michigan
We examined the effect of family life, work, and war-related stressors on
posttraumatic stress disorder (PTSD) symptoms, depression symptoms, per-
ceived physical health, and perceived functioning in a large sample of 796
randomly selected Air Force men and women who were deployed during the
period of October 7, 2001 to September 2004. As predicted by conservation
of resources (COR) theory, we found that these combined sources of stress
directly affected PTSD and depression symptoms and that resource loss,
more than resource gain, mediated these effects. We further found that
greater PTSD and depression symptoms were related to diminished physical
health and poorer functioning. Contrary to what COR theory might predict,
resource losses did not have a greater direct effect on perceived health and
functioning than resource gain. However, because stressors had so much
greater effect on resource loss than on gain, and resource loss had so much
more direct effect on PTSD and depression symptoms than resource gains, it
is still the stronger mediator of the effects of stressors on perceived health
and functioning than resource gain. Critical for military policy, our results
This article was published Online First July 9, 2012.
Stevan E. Hobfoll, Department of Behavioral Science, Rush Medical College; Amiram D.
Vinokur and Lisa Lewandowski-Romps, Institute for Social Research, University of Michigan;
Penny F. Pierce, Institute for Social Research and School of Nursing, University of Michigan.
This research (MDA 905-04-1-TS07) was sponsored by a TriService Nursing Research
Program (TSNRP) grant to Penny F. Pierce, Uniformed Services University of the Health
Sciences. However, the information or content and conclusions do not necessarily represent the
official position or policy of, nor should any official endorsement be inferred by, the TSNRP,
the Uniformed Services University of the Health Sciences, the U.S. Department of Defense, or
the U.S. Government.
Correspondence concerning this article should be addressed to Stevan E. Hobfoll, 1645 W.
Jackson Blvd., Suite 400, Chicago, IL 60612. E-mail: Stevan_Hobfoll@Rush.edu
International Journal of Stress Management © 2012 American Psychological Association
2012, Vol. 19, No. 3, 217–237 1072-5245/12/$12.00 DOI: 10.1037/a0029247
217
highlight the greater importance of limiting resource loss cycles than pro-
moting gain cycles, even if each has a role in supporting or undermining
well-being, psychosocial and work functioning, and deployment readiness.
Keywords: Resources, losses, gains, health, functioning
Given the prolonged nature of many military conflicts and the employ-
ment of an all-volunteer force, U.S. military personnel are more likely to be
subjected to the exposure of war-related trauma and must meet the challenge
of balancing their military commitment with other life domains (Gold et al.,
2007). Exposure to war-related violence and threat may compound the
influence of job stressors, financial stressors, and work-family conflict. This
also means that posttraumatic stress disorder (PTSD) and depression are not
the sole outcomes of this array of stressors. Rather, military personnel, their
families, and the military are also concerned with military personnel’s
physical health and functioning. Indeed, because stress and PTSD are related
to attrition from the military, this becomes a critical problem for retaining a
healthy fighting force (Hoge, Auchterlonie, & Milken, 2006). It follows that
a more comprehensive picture of war-related stress for military personnel
should include examination of these multiple sources of stress as well as the
multiple outcomes of the combined sources of stress. This would simultane-
ously tie together the literature on traumatic stress and war and the literature
on traditional occupational stress (e.g., dealing with bosses, colleagues, and
paperwork), which would be more contextually representative of the realities
of military life.
Clearly, the threat of death or injury, and witnessing horrible war scenes
are a major source of military stress (Solomon, Mikulincer, & Hobfoll, 1986;
Tanielian & Jaycox, 2008). Studies have repeatedly found that PTSD and
depression are common sequelae of military involvement (Vasterling et al.,
2010) and that these effects are often long lasting (Hoge et al., 2004; Kessler,
2000; Kulka et al., 1990). In general, PTSD is a severe reaction to the threat
of death or harm, characterized by reexperiencing the trauma as if it is still
occurring, avoidance, and hyperarousal. In turn, depression is characterized
by feelings of sadness, malaise, and reduced energy and may vary from quite
mild to extremely severe. Studies of returning troops from Iraq and Afghan-
istan estimated the troops to have a PTSD prevalence of 14 –16% (Hoge et
al., 2004; Tanielian & Jaycox, 2008), and levels of depression symptoms are
also reported as being substantial (Hoge et al., 2004), although these are
probably lower for Air Force (AF) personnel. War-related PTSD and depres-
sion are painful, affect many aspects of life, reduce military readiness, and
generate long-term costs for the individual and society (Hoge et al., 2004).
War-related and more mundane occupational stressors experienced by
military personnel have individual and combined likely negative health
218 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
consequences (Elder, Shanahan, & Clipp, 1997; Neria & Koenen, 2003).
Moreover, in their military and civilian roles, military personnel need to
function at high levels, but those who experience PTSD and depression have
been found to have significantly impaired functioning (Gimbel & Booth,
1994; Gold et al., 2007; Hoge, Terhakopian, Castro, Messer, & Engel, 2007;
Vogt, King, Savarese, & Suvak, 2004). Prior studies, including one from this
same data set, suggest that military stress is similar to other high-stress
occupations, with high levels of burnout that may spill over into other areas
of postmilitary employment if underlying levels of functioning are dimin-
ished (Vinokur, Pierce, Lewandowski-Romps, Hobfoll & Galea, 2011). This
suggests that war-related stressors and life stressors that accompany military
personnel would result in PTSD and depression, which in turn would be
related to poorer physical health and poorer functioning. Indeed, the causes
of PTSD and depression are overlapping; for example, job loss was one of the
most important predictors of PTSD after the World Trade Center attack,
although job loss is typically not seen as a traumatic event (Galea et al.,
2002).
Conservation of resources (COR) theory (Hobfoll, 1989, 1998) may be
a particularly well-suited model to examine military stress because it has
guided research on war-related trauma in military personnel (King, King,
Foy, Keane, & Fairbank, 1999) and more general occupational stress (Hal-
besleben, 2006). COR theory is based on the central tenet that individuals
“strive to obtain, retain, foster, and protect those things they centrally value”
(Hobfoll, 2011, p. 117). These centrally valued entities are termed “re-
sources” and include personal, social, material, and energy resources. COR
theory further posits that resource loss is the main cause of the negative
sequelae of stressful conditions, including psychological distress, negative
health outcomes, and diminished functioning. Although the full basis for
loss’ primacy is beyond the scope of this paper (see Hobfoll, 1988, 1989,
1998, 2001), its core is found in the evolutionary need to attend especially to
losses because they critically challenge survival. In turn, this has led to a
primacy of loss in neurological, cognitive, emotional, and behavioral do-
mains. At the same time, it is the preservation of these same resources that
are the building blocks of stress resilience. To the extent people possess and
retain a strong reservoir of psychosocial and material resources, they are
more resistant to resource losses that do occur. According to COR theory,
traumatic stress is characterized by a rapid loss of psychosocial and material
resources (Hobfoll, 1991) whereas job-related demands often result in a slow
drain of resources (Hobfoll & Shirom, 1993). Over time, this means that for
military personnel, traumatic and job-related stressors will combine to result
in a depleted pool of resources that are a cause of these stress outcomes and
further undermine people’s ability to resist stressful demands because the
219Warriors Life Stress
very resources they would call upon are depleted. These are termed “resource
loss spirals” in COR theory.
COR theory also states that resource gains will contribute to improved
psychological well-being, health, and functioning, but there are relatively few
studies of resources gains (Gorgievski & Hobfoll, 2009; Wells, Hobfoll &
Lavin, 1999). Recent work in the organizational literature has pointed to the
importance of resource gains in contributing to vigor and engagement (Hal-
besleben, Harvey, & Bolin, 2009; Hobfoll, 2011), which again are important
states for individuals, their families, civilian employers, and the military.
However, a central principle of COR theory is that resource gains have
markedly weaker impact than resource losses (Hobfoll, 1989; Wells et al.,
1999).
We examined the effect of war-related stressors and job and family
stressors on several outcomes, including PTSD symptoms, depression
symptoms, perceived physical health, and perceived functioning among a
large sample of randomly selected AF men and women who were de-
ployed during the period of October 7, 2001 to September 2004. On the
basis of prior work, we made several predictions. First, (H1) we predict
that the family, occupation, and traumatic sources of stress would com-
bine to directly affect our multiple outcomes. Second, (H2) we predict
that resource loss and gain would also significantly mediate the effects of
these sources of stress on the multiple outcomes, but they could also have
direct effects (see Figure 1). In addition, on the basis of prior studies, the
model suggests that (H3) those who experience greater PTSD and de-
pression symptoms would also experience diminished physical health and
poorer functioning, which are also depicted in the model as mediating the
effects of losses and gains on these outcomes. As follows from COR
theory, it is further predicted that (H4) the effect of war-related, family,
and job stressors would be mediated through the loss pathways to a
greater extent than the gain pathways. Looking at the model in total,
support for these hypotheses would not only support COR theory but
(also) be a critical finding for the military. In particular, it would highlight
the greater importance of limiting loss cycles than promoting gain cycles,
even if both have a role in supporting or undermining well-being, psy-
chosocial, and work functioning and deployment readiness.
METHOD
Approval for the conduct of this study was provided by the Institutional
Research Board of the University of Michigan and the Uniformed Services
University of the Health Sciences.
220 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
Sampling and Data Collection Procedures
A probability sample with contact information for 2,250 AF men and
women was obtained from the Defense Manpower Data Center (DMDC) of
the U.S. Department of Defense. The sample included AF men and women
who were deployed by the AF during the period of October 7, 2001 to the
time of the sample request (September 2004). To provide better coverage and
representation of major groups of interest, the sample was constructed as a
random stratified sample to include 50% men (compared with 86% in the
deployed AF), 33.3% from the Active Duty component of the AF (compared
with 9% in the AF), 33.3% from Reserve (compared with 74% in AF), and
33.3% from the National Guard (compared with 17% in AF).
A brief recruitment letter explaining the study and inviting participation
was mailed to the 2,250 participants along with a small gift (a pen). Approx-
imately 1 week later, a 20-min telephone interview was conducted with
individuals who we had phone contact information for and/or called our
Figure 1. Longitudinal effects (standardized coefficients) of stressors, resource losses, and gains
on mental health and functioning outcomes. Stressors include job and financial stressors,
exposure to trauma, length of deployment, and work-family conflict. All solid paths or
correlations are statistically significant at .05. Curved lines represent correlations (based on the
respective covariances) between the residuals.
2
(180, N ⫽ 758) ⫽ 311.44, p ⬍ .001, TLI ⫽
.99, CFI ⫽ .99, SRMR ⫽ .03, and RMSEA ⫽ .03. R
2
⫽ variance accounted for by the
predictors.
221Warriors Life Stress
toll-free phone number. This was followed by the mailing of the first, time-1
(T-1), self-administered questionnaire (about June 2005) with another pen as
an incentive. The option was provided of completing the questionnaire on the
web, and 40% of the participants chose this option. Of the 2,250 recruitment
letters sent, 141 (6%) were returned because of inaccurate addresses, and
these individuals could not be located by other searches. The other letters
resulted in 1,451 completed telephone interviews, and of those, 1,009
(69.5%) provided data using the mailed questionnaire (605, or 60%) or a
web-based version (404, or 40%). One year later, we sent participants
announcement reminders about the time-2 follow-up data collection (T-2;
September 2006) and a pen as an incentive. The 1-year follow-up period was
intended to provide sufficient time for changes to occur in the conditions of
the deployment and its associated stresses (De Lange, Taris, Kompier,
Houtman & Bongers, 2004).
Of those 1,009 who completed the T-1 questionnaire, 796 men and women
(79%) also completed the T-2 questionnaire or its web-based equivalent (32 and
68%, respectively). Respondents had a mean age of 38.2 years, a median of 14.9
years of education, and a median household income of $55,365. The sample
included 49.9% women, 76.5% Whites, 14.4% African Americans, and 5.3%
Hispanics; 53.1% were married; and 42.4% had at least one dependent child
during deployment. In terms of military background, the sample included 25.5%
officers and 73.4% enlisted soldiers, and 25.5, 39.0, and 35.5%, respectively,
were from the Active Duty force, the Reserve, and the National Guard. As for
deployment, 65.1% were deployed at least once to the war theater. Deployment
to the theater of war was defined based on the U.S. Department of Defense
designation as deployment to at least one of the locations that included Iraq,
Afghanistan, Qatar, Kuwait, or Saudi Arabia. Deployments to nontheater loca-
tions included deployment to Europe, Southeast Asia, or other countries without
current active military conflict (e.g., Korea, Japan).
Measures
Basic Demographic and Background Military Information
Information was obtained from the respondents and from the database
provided by DMDC. We used age, gender, income, and military rank as
control variables. The values of all of the following multi-item measures
were computed as the mean score of the respective items with higher values
representing higher levels of the measure.
222 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
Stressors Assessed at T-1
Exposure to Trauma
Exposure was measured with 18 stressful events (␣⫽.79) most likely to
occur in a theater of war such as “come under small arms fire” and “take
shelter in a bunker”. This list was based on the Vietnam era combat exposure
scale, which has been validated (e.g., Keane et al., 1989) and used in
numerous studies on veterans of the war in Vietnam (e.g., Laufer, Gallops, &
Frey-Wouters, 1984). In our revision of the original scale we added items to
incorporate a broader set of traumatic events related to participation in the
current war effort, specifically sexual harassment and sexual assault or rape.
Our survey participants were asked to check whether they experienced each
event on the list during the period from October 2001 to the present time
(June 2005) and were instructed to rate the extent to which it made them feel
afraid, horrified, or helpless on a four-point scale ranging from 1 (not at all)
to4(a great deal). A “0” rating was inserted for each event that the
participants did not experience. Many exposure items (e.g., seeing dead
bodies, involvement in accidents that produce major injuries) happened to
in-theater and nontheater personnel. The mean of all 18 items on the recoded
five-point scale was used as the score for exposure to trauma.
Job Stressors
Job stressors were measured using eight items (␣⫽.86), of which six
were originally written by Kandel, Davies, and Raveis, (1985) and applied in
research by Frone, Russell, and Cooper (1992). Two items were added to
represent aspects of job stressors more frequent in military jobs (feeling
harassed, intimidated). The items evaluated the level of daily emotional
reaction experienced on the job (e.g., relaxed, frustrated, fortunate) using a
four-point scale ranging from 1 (not at all)to4(very). The scores of the items
that imply low stressors were reversed so that a high composite mean score
indicated high job stressors.
Financial Stressors
Financial stressors were measured with an index based on three items
(␣⫽.80; Vinokur & Caplan, 1987). Respondents were asked to use five-
point scales to rate their current and anticipated economic hardship, which
included difficulties living on their household income and living with a
223Warriors Life Stress
reduced standard of living. This scale was shown to be highly correlated (r ⫽
.76) with commonly used economic hardship scales that focus on financially
stressful events, such as borrowing money to pay bills (Barrera, Caples, &
Tein, 2001).
Work-Family Conflict
This conflict was assessed with a two-item scale (␣⫽.84) used by Frone
et al. (1992). Respondents were asked to provide their ratings on a five-point
scale ranging from 1 (almost never, or never)to5(almost always, or always)
to the questions “How often does your Air Force job or career interfere with
your responsibilities at home?” and “How often does your Air Force job or
career keep you from spending the amount of time you would like to spend
with your family?”
Length of Deployment
The measure included the reported total length of deployment from
October 2001 to the present time (June 2005) to the theater of war or
elsewhere.
Losses and Gains Outcomes Assessed at T-1
Loss and Gain of Resources
Losses and gains were assessed using 13 items for each construct (␣⫽
.88 and .91, respectively) on the basis of the measures developed by Hobfoll
and Lilly (1993) following COR theory’s definition of resources (Hobfoll,
1989, 1998). A preamble to the items included the question “To what extent
did you have losses (gains) in the following areas of your life as a result of
your deployment from October 2001 to the present time?” A five-point
response scale ranging from 1 (not at all)to5(large extent) was provided for
answering each item. The content of the items covered aspects of losses in
family and social relationships (e.g., your relationship with your personal
friends), financial matters (e.g., your financial situation or financial re-
sources), career (e.g., your career advancement prospects), and personal
matters (e.g., your feeling of pride serving the country).
224 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
Proximal Outcomes Assessed at T-2
Symptoms of PTSD
The symptoms were evaluated with a 17-item scale (␣⫽.94) from the
PTSD Checklist-Military version (Weathers, Huska, & Keane, 1991). The
content of the items encompasses a broad range of distress reactions includ-
ing emotional (e.g., get very upset or anxious), cognitive (trouble keeping
your mind on what you are doing), and physical/physiological (heart pound-
ing, begin to sweat) stress reactions. Using five-point scales ranging from 1
(never)to5(very frequently), the respondents rated the frequency of expe-
riencing each symptom during the period from October 2001 to the present
time.
Depressive Symptoms
The depressive symptoms were assessed with a subset of 11 items (␣⫽
.89) based on the Hopkins Symptom Checklist (Derogatis, Lipmann, Rickels,
Uhlenhuth & Covi, 1974). Using five-point scales ranging from 1 (not at all)
to5(extremely), respondents indicated how much they had been “. . .
bothered or distressed in the last 2 weeks” by various experiences such as
crying easily, blaming yourself for things, feeling lonely, feeling no interest
in things, feeling blue, thoughts of ending your life, and feeling hopeless
about the future.
Distal Outcomes Assessed at T-2
Perceived Health
This construct was measured with a four-question scale (␣⫽.78). The
questions were based on similar items from the Medical Outcome Study
(Stewart & Ware, 1992). Using five-point rating scales ranging from 1
(never/no extent) to 5 (a very great extent), respondents provided their
answers to the following questions: “In general, would you say your health
is excellent, good, fair, or poor?”, “To what extent do you have any particular
health problems?”, “Thinking about the past 2 months, how much of the time
has your health kept you from doing the kind of things other people your age
do?” (1 ⫽ none of the time to 5 ⫽ all of the time), and “To what extent do
you feel healthy enough to carry out things that you would like to do?” (1 ⫽
never/no extent to 5 ⫽ a very great extent). The scores of the items that imply
225Warriors Life Stress
poor health were reversed so that the composite mean score indicates better
health.
Role and Emotional Functioning
Functioning was assessed with a 15-item scale (␣⫽.95). The scale was
developed by Caplan et al. (1984) and validated in other studies that ad-
dressed handling responsibilities and daily demands and making the right
decisions (Vinokur, Price, & Schul, 1995). Respondents were requested to
provide their answer to the following question: “In the last two weeks, how
well have you been doing with respect to the following activities?” using a
five-point rating scale ranging from 1 (very poorly)to5(exceptionally well).
Analysis
First we conducted logistic regression attrition analysis to assess possible
bias in the characteristics of respondents who participated in the study versus
those who did not. Second, we used logistic analysis to predict attrition at T-2 on
the basis of the data from T-1. Third, structural equation modeling analyses were
conducted including the estimation of the structural equation model that was
constructed to test our hypotheses (see Figure 1). A stressor composite variable
was included based on the mean scores of the five indicators of stressors that
included work- and war-related stressors, including job stressors, work-family
conflict, financial stressors, exposure to trauma, and length of deployment. The
stressor composite variable was modeled as a latent factor with the composite
score as its sole indicator. Its error was fixed as the multiplication of its variance
by one minus the reliability. The reliability of this composite variable was
estimated based on the test-retest correlation between T-1 and T-2 (r ⫽ .69). The
modeling of the stressor latent factor followed the recommendation by
Treiblmaier, Bentler, and Mair (2011, p. 12) with regard to formative latent
factors. (A model with the stressors as causal indicators of a formative latent
factor could not produce an identified unique solution and therefore the com-
posite variable approach was applied here.)
All of the other constructs were specified in the model as latent factors
that were modeled with parcel indicators. The use of parcel indicators is
justified when the goal is to understand relations among latent variables while
at the same time reducing non-normal distributions of single items with the
more appropriate ones as parcels (see Williams & O’Boyle, 2008). Perceived
health was indicated by two parcels, and the other factors were indicated by
226 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
three parcels each. The parcels were constructed by alternating the items as
they appear in the questionnaire to be placed in different parcels. For each
construct, the reliability of the parcels were computed to ascertain that it was
nearly the same as the reliability coefficient based on the entire set of items
of the construct. Finally, the model also included four control variables—age,
gender, income, and rank—with their effects on the factors and their esti-
mated intercorrelations.
Using Mplus software (version 6.1) we estimated our model (see Figure
1) on the basis of the data from the 796 participants who responded to T-1
and T-2. The estimation of the model applied the maximum likelihood
method with the full information maximum likelihood (FIML) procedure
providing a total sample of 758 respondents. To determine model fit, we
relied on a standard recommendation to examine several fit measures.
We follow Hu and Bentler’s (1999) suggestion to consider models with
the Tucker Lewis (TLI) and confirmatory fit (CFI) indices with values
close to .95 and the misfit indices standardized root mean square residual
(SRMR) and root mean square error of approximation (RMSEA) with
values equal or less than .08 and .06, respectively, as providing evidence
of good fit.
RESULTS
Analyses of Participation at Baseline
We used a logistic regression analysis, including all of the demographic
and military background variables, to predict the participation of the 796 of
respondents in T-1 and T-2 out of the targeted sample of 2,250. Age, parental
status, rank, component, and length of deployment in months were significant
predictors (odds ratio ⫽ 1.04, 1.38, 1.75, 1.29, and 1.06, respectively; p ⬍
.05). Higher response rates were found for older participants (38.8 vs. 33.9
years), parents (42% vs. 28%), officers versus enlisted (51% vs. 32%),
Reserve and National Guard members versus Active Duty (39% vs. 27%),
and those deployed for a longer period of time (3.2 vs. 3.0 months). Female
gender and deployment in the theater of the war (vs. elsewhere) did not
significantly predict participation. Thus, although there were several demo-
graphic and military variables that biased the representation of our overall
sample, the generalizability of our findings is strengthened by the greater
coverage of particular groups in the military, such as parents, who are
typically underrepresented in research.
227Warriors Life Stress
Analyses of Attrition at Follow-Up
A total of 796 respondents completed the T-2 questionnaire for a 79%
response rate. Using a logistic analysis to predict attrition at T-2 on the basis
of the data from T-1, we found that only age significantly predicted
participation at T-2 (odds ratio ⫽ 1.03; p ⬍ .01), completers being
somewhat older than noncompleters (M ⫽ 38.5 vs. 34.6 years, respec-
tively). Furthermore, we examined attrition at T-2 by comparing the T-1
reports of the participants and the nonparticipants that included various
mental health and functioning variables such as depression, role func-
tioning, and perceived health. We found that none of these comparisons
yielded a statistically significant difference, indicating that attrition did
not play a meaningful role in altering the representativeness of the
original sample used for our analyses.
Structural Equation Modeling Analyses
The results of the estimated model are displayed in Figure 1, and the
effects of the control variables on the constructs and the correlations among
the control variables are shown in Table 1. The corresponding matrix of
correlations of the model constructs is presented in Table 2 along with the
means and standard deviations of the respective variables. The results of the
analysis for our sample and those for an analysis with weights to correct for
our oversampling of women and component of service were similar (avail-
able upon request). Here we focus on the results of our sample, which
includes a more balanced representation of gender and component groups.
The estimation of the measurement model that applies to our model
produced a
2
(159, N ⫽ 796) ⫽ 292.8, p ⬍ .001, TLI ⫽ .98, CFI ⫽ .99,
SRMR ⫽ .02, and RMSEA ⫽ .03. Our estimated structural model produced
a
2
(182, N ⫽ 758) ⫽ 358.60. However, we discovered in the modification
indices that a revised model with two additional paths from stressors to PTSD
and to depressive symptoms would fit significantly better. Figure 1 displays
the results with the added two direct paths from stressors to PTSD and to
depressive symptoms, which include a
2
(180, N ⫽ 758) ⫽ 311.44, p ⬍
.001, TLI ⫽ .99, CFI ⫽ .99, SRMR ⫽ .03, and RMSEA ⫽ .03. All of the fit
indices indicate an excellent fit of the data to the model.
As hypothesized (H1), greater stressors were predictive of greater re-
source losses (⫽.68, p ⬍ .001) and diminished resource gains (⫽–.14,
p ⬍ .01). Stressors also predict higher PTSD symptoms (⫽.43, p ⬍ .001)
and depressive symptoms (⫽.36, p ⬍ .001). In turn, resource losses predict
higher levels of PTSD (⫽.13, p ⬍ .05) and lower levels of perceived
228 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
health (⫽⫺.13, p ⬍ .01). Resource gains predicted higher role and
emotional functioning (⫽.13, p ⬍ .001). Finally, PTSD and depressive
symptoms each predicted lower perceived health (respectively, ⫽⫺.17 and
⫺.37; p ⬍ .01) as well as lower functioning (respectively ⫽⫺.17 and
⫺.43; p ⬍ .001).
We applied bootstrap statistics for testing mediation (Selig & Preacher,
2008) and found several important indirect effects indicative of mediation.
The following unstandardized estimates of these indirect effects were signif-
icantly different from 0 using 95% confidence intervals. First, as predicted in
H2, we found that losses partially mediated the effect of stressors on PTSD
Table 1. Effects of Background Variables on the Constructs in Figure 1 and the
Correlations Among Them
Effects Age Gender
a
Income Rank
b
Stressors –.17
ⴱⴱⴱ
Losses .09
ⴱⴱ
Gains
PTSD symptoms –.09
ⴱⴱ
Depressive symptoms –.11
ⴱⴱⴱ
–.08
ⴱ
–.07 (NS)
Perceived health –.10
ⴱⴱ
.17
ⴱⴱⴱ
Role and emotional functioning .09
ⴱⴱ
Correlations
Gender .10
ⴱⴱ
Income .40
ⴱⴱ
.11
ⴱⴱ
Rank .14
ⴱⴱⴱ
–.03 .45
ⴱⴱⴱ
Note. N ⫽ 758. NS ⫽ not significant.
a
Gender: 0 ⫽ female, 1 ⫽ male.
b
Rank: 0 ⫽ enlisted, 1 ⫽ officer.
ⴱ
p ⬍ .05.
ⴱⴱ
p ⬍ .01.
ⴱⴱⴱ
p ⬍ .001.
Table 2. Correlations, Means, and Standard Deviations of Constructs and Variables
in Figure 1
a
Measure 1 2 3 4 5 6 7 8 9 10 11
1. Stressors —
2. Losses .51 .88
3. Gains –.11 –.19 .91
4. PTSD symptoms .41 .37 –.04 .95
5. Depressive symptoms .35 .30 –.06 .68 .90
6. Perceived health –.30 –.29 .10 –.45 –.49 .81
7. Functioning –.30 –.27 .17 –.50 –.56 .43 .96
8. Age –.09 .05 –.03 .02 –.03 –.06 .04 —
9. Gender
b
–.08 –.04 –.02 ⫺.09 –.18 .09 .09 .08 —
10. Income
c
–.14 –.08 –.03 –.15 –.19 .10 .19 .40 .10 —
11. Rank
d
.00 –.03 .04 –.11 –.12 .21 .08 .14 –.02 .46 —
M .00 1.54 2.20 1.53 1.41 4.06 3.66 38.48 .51 3.44 .28
SD .56 .62 .80 .66 .56 .71 .66 8.77 .50 1.26 .45
Note. N ⫽ 796. Bold figures in the diagonal are Cronbach’s ␣ reliabilities.
a
Correlations of .073 are statistically significant at p ⬍ .05. Internal consistency reliabilities
appear in the diagonal in bold font.
b
Gender: 0 ⫽ females, 1 ⫽ males.
c
Income ranging
from 1 ⫽⬍$20,000 (U.S.) to 5 ⫽ ⱖ$80,000 (U.S.).
d
Rank: 0 ⫽ enlisted, 1 ⫽ officers.
229Warriors Life Stress
(.13) and fully mediated the effects of stressors on perceived health (–.12). In
turn, as predicted in H2, gains partially mediated the effect of stressors on
functioning (–.03). Second, as predicted in H3, PTSD and depressive symp-
toms fully mediated the effects of stressors on perceived health and func-
tioning (respectively for PTSD ⫺.10 and ⫺.10; for depressive symptoms
⫺.17 and ⫺.21). Third, as also predicted in H3, PTSD also partially mediated
the effects of losses on perceived health and functioning (both ⫺.02).
Examining the relative effects of resource losses and resource gains, as
predicted in H4, resource loss had a greater effect on PTSD, depression, and
perceived health given its direct and indirect effects. Resource gains did have
a greater effect than resource losses on functioning, contrary to H4.
Testing Alternative Models
Our model includes effects that are examined longitudinally as well as
various relationships that are modeled as directional effects within each time
of data collection. Although these effects reflect our theoretical consider-
ations and hypotheses, they need to be examined more critically given the
cross-sectional nature of their location in the data collection and the possi-
bility that the direction of the effects should more adequately be represented
in the opposite direction. To examine whether other models with the direction
of these effects reversed provide better fit, we tested the following three
alternative models. First, we reversed the direction of influence from stressors
to losses and gains (i.e., modeling the losses and gains having an effect on the
stressor factor). Second, in a similar vein, we tested a model with perceived
health having an effect on PTSD and on depressive symptoms. Third, we
tested a model with functioning having an effect on PTSD and depressive
symptoms. All of these three alternative models are not nested in our base
model (see Figure 1) and therefore a strict
2
test of which is statistically
superior is not available. Instead researchers compare such non-nested mod-
els using the misfit measure of Akaike information criteria (AIC), with
smaller values indicating better fit.
The results show that the three alternative models, all with same number
of degrees of freedom, had, each, respectively, a
2
(180, N ⫽ 758) of
321.15, 327.05, and 328.83, which was slightly larger that of our base model
in Figure 1,
2
(180, N ⫽ 758) ⫽ 311.44. They all had virtually the same
goodness-of-fit values with a TLI between .98 and .99 and a CFI of .99 as
well as a RMSEA of .03. In addition, the alternative models had slightly
poorer (larger) AIC fit measures, respectively, 16,572.74, 16,578.64, and
16,580.42, compared with 16,563.03 for our base model. In conclusion, it
seems that our model provided somewhat better fit, or was more parsimoni-
ous, than the alternatives.
230 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
Finally, we also conducted a two-group comparison of the Active Duty
and the Reserve and National Guard combined. Compared with the Active
Duty force, the Reserves and the Guards experience a similar situation in the
sense that they are called for duty having to leave behind their normal civilian
environment. The two-group model with all of the paths constrained to
equality across the groups had very good fit with a
2
(407) ⫽ 627.55, p ⬍
.001, TLI ⫽ .98, CFI ⫽ .98, SRMR ⫽ .06, and RMSEA ⫽ .04. Releasing
constraints did not produce a significantly lower
2
for individual paths.
Thus, there were no meaningful differences between the two groups that
relate to our hypotheses.
DISCUSSION
The results were largely consistent with the proposed model. They were
also robust and held for generalizations to the AF in general and to Active
Duty, Reserve, and National Guard. By oversampling we were also able to
be more confident in regard to women personnel and those with families. The
results largely support COR theory’s emphasis on the primacy of resource
loss and the secondary, but important, influence of resource gains. First
reviewing the COR theory-based predictions of the model, it can be noted
that resource loss is generally a more powerful predictor than resource gains
of the effect of work and war stressors on PTSD and depressive symptoms.
This is because resource loss follows more directly from stressors and
because it has a stronger effect than resource gains on psychological distress.
Resource losses did not have a stronger direct effect on functioning than
resource gains. Nevertheless, the results support this key principle of COR
theory when considering the overall picture.
It is also important to note the close tie between stressors and resource
losses versus the weak tie between stressors and resource gains. Resource
gains are little affected by stressors, suggesting that stressors are more tied to
the loss of resources than reversals in resource gains or failure to gain
resources. Although the one effect is notable, resource gains are only related
to functioning and otherwise seems extraneous to the stress-distress process
in the model we tested. This is not to say that resource gains may have
important other functions, such as helping to sustain coping in the face of
adversity.
Beyond the scope of COR theory, other important findings of our study
emerge. Perhaps the most critical is that in our sample depression symptoms
had a significantly greater negative effect on perceived health and functioning
than PTSD symptoms. However, the overlap between PTSD and depression
symptoms (note they are correlated here, r ⫽ .63) suggests that PTSD would
still be important in military personnel’s lives.
231Warriors Life Stress
It is also notable that stressors affect PTSD and depression symptoms
and perceived health and functioning independent of resource losses and
gain. These effects were especially strong on PTSD and depression symp-
toms. These results may indicate that our measure of resource loss and
resource gain is not, in fact, assessing all of the gains and losses that our
participants are reporting. However, it may also indicate that elements of
stressors are at the state of threats that have not materialized to resource
losses or gains. Moreover, some of these stressors may be more long term and
therefore not being translated as recent losses or gains but as steady states
related to ongoing PTSD and depression symptoms. As Bonanno, Galea,
Bucciarelli, and Vlahov (2007) found in a recent study of disaster, resource
losses and lack of resources are important predictors of psychological dis-
tress, as consistent with COR theory.
The link between PTSD and depression symptoms and perceived
health may emanate from two major sources: exaggerated negative ap-
praisal of actual health and direct and indirect processes that impair actual
health. First, it is well founded that depression in particular is linked to a
negative evaluative framework (Giesler, Josephs, & Swann, 1996) such
that individuals might not actually have worse health, but be focusing on
it. However, studies have found that subjective and objective health
reports are substantively related (McCullough & Laurenceau, 2004;
Singh-Manoux et al., 2006). Further, there is also increasing evidence that
PTSD and depression symptoms result in direct health impairment
through compromise of the immune system and inflammation via activa-
tion of the hypothalamic-pituitary-adrenal axis (McDade, Hawkley, &
Cacioppo, 2006; Miller, Freedland, & Carney, 2005). In addition, those
who become more depressed and develop greater PTSD symptoms are
more likely to experience health problems due to poorer health behavior,
such as drinking, smoking, and lack of exercise (Fylkesnes, 1993; Jak-
upcak, Luterek, Hunt, Conybeare, & McFall, 2008).
Our study has several limitations as well as clear strengths. Some
results are based on the cross-sectional aspects of our model (i.e., rela-
tions within T-1 and T2) and therefore must be considered with appro-
priate caution. However, our tests of alternative models that reversed the
direction of influence reveal these models to be less compelling than the
one we present. Another limitation of the scope of our results is related to
the specific sample of AF personnel being studied. The generality of the
results could become more compelling if they replicate with other mili-
tary forces such as the Army. Self-report methodology also has limita-
tions, and triangulating the data with clinical interviews, biological mark-
ers of health, and supervisor ratings would strengthen the study findings
and conclusions.
232 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
Intervention
Our results point to potential interventions. Certain aspects of military
personnel’s lives are inevitable, and there is even an acceptance of honor in
exposure to danger in defense of the nation that is an essential element of
dedicated military service. However, this does not mean that multiple and
long deployments must occur, that financial stressors cannot be eased, and
that work-family conflict cannot be limited through intervention programs.
Next, the results suggest that intervention at the time of resource loss may be
a particularly powerful juncture for intervention. Losses can be reversed or
lessened through resiliency training of the military member and his or her
family, including community-level interventions, and these could target
specific areas of loss. Clearly such interventions would affect PTSD and
depression. It is notable that it is depression that has the greater downstream
effects on health and functioning, suggesting that although PTSD is more
newsworthy, depression is nevertheless a major target of intervention.
Another element to consider for potential interventions pointed out by
our research is that PTSD symptoms are not only driven by traumatic
exposure. Rather, as Hobfoll (1998) theorized, he and others have found in
studies of veterans (e.g., King et al., 1999) that PTSD symptoms are also
largely related to more general stressors occurring in people’s lives, com-
promising their personal and interpersonal resilience. As such, PTSD symp-
toms are influenced by the presence or lack of the same personal and
interpersonal resiliency variables that affect all kinds of stress, and not just
traumatic stress. Said another way, the PTSD literature has tended to focus on
the trauma and trauma-related symptoms, but successful intervention must
address the trauma and the more everyday stressful aspects of people’s lives.
This caution extends beyond military trauma research, as studies of victims
of the 2005 Katrina hurricane and flood disaster found that loss of one’s home
and work were as highly related to PTSD as was exposure to threat of death,
losing loved ones, and witnessing death and destruction (Zwiebach, Rhodes,
& Roemer, 2010).
Conclusions
We applied COR theory to the ecologies of war-related trauma and the
more family and occupational stressors that military personnel endure while
living through wartime deployments. This application provided a robust
model for understanding their stressors, distress, health, and functioning. This
is rather unique because the trauma literature and the occupational stress
literature are published in different journals, approached by different re-
233Warriors Life Stress
searchers, and tend not to consider overlapping theory. That COR theory has
been so successfully applied to both domains makes it one possible theory for
illuminating an understanding of the special stress of military personnel, but
other theories should also be considered. For example, Frederickson’s (1998)
“broaden and build” theory, in which positive emotions lead to greater
resiliency, may help our understanding of the positive, challenge elements of
military service. Bonanno’s (2004) model of predictors of resiliency has
similarly been applied to traumatic and nontraumatic stressors. Likewise,
Schaufeli and Bakker’s (2004) dual channel theory of burnout and engage-
ment as simultaneous processes would also help us to understand the negative
effect of military involvement as well as the energizing and highly motivated
elements of engagement that come with a military experience or career. We
see our application of COR theory as one step in what we hope becomes a
broader investigatory journey as trauma exposure is better integrated with a
consideration of the multiple aspects of military members’ and their families’
lives that contribute in unique ways along the path of recovery from wartime
service.
REFERENCES
Barrera Jr., M., Caples, H., & Tein, J. Y. (2001). The psychological sense of economic
hardship: Measurement models, validity, and cross-ethnic equivalence for urban families.
American Journal of Community Psychology, 29, 493–517. doi:10.1023/A:
1010328115110
Bonanno, G. A. (2004). Loss, trauma, and human resilience: Have we underestimated the
human capacity to thrive after extremely aversive events? American Psychologist, 59,
20 –28. doi:10.1037/0003-066X.59.1.20
Bonanno, G. A., Galea, S., Bucciarelli, A., & Vlahov, D. (2007). What predicts psychological
resilience after disaster? The role of demographics, resources, and life stress. Journal of
Consulting and Clinical Psychology, 75, 671– 682. doi:10.1037/0022-006X.75.5.671
Caplan, R. D., Abbey, A., Abramis, D. J., Andrews, F. M., Conway, T. L., & French, J. R. P.
(1984). Tranquilizer use and well being: A longitudinal study of social and psychological
effects. (Tech. Rep. No. Series), Ann Arbor: Institute for Social Research. Retrieved from
http://hdl.handle.net/2027/mdp.39015016218128.
De Lange, A. H., Taris, T. W., Kompier, M. A. J., Houtman, I. L. D., & Bongers, P. M. (2004).
The relationships between work characteristics and mental health: Examining normal,
reversed and reciprocal relationships in a 4-wave study. Work & Stress, 18, 149–166.
doi:10.1080/02678370412331270860
Derogatis, L. R., Lipmann, R. S., Rickels, K., Uhlenhuth, E. H., & Covi, L. (1974). The
Hopkins Symptoms Checklist (HSCL): A measure of primary symptom dimensions. In P.
Pichot (Ed.), Psychological measurements in psychopharmacology: Modern problems in
pharmacopsychiatry. (Vol. 7, 79 –110). Basel, Switzerland: Karger.
Elder Jr., G. H., Shanahan, M. J., & Clipp, E. C. (1997). Linking combat and physical health:
The legacy of World War II in men’s lives. The American Journal of Psychiatry, 154,
330 –336.
234 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
Fredrickson, B. L. (1998). What good are positive emotions? Review of General Psychology,
2, 300 –319. doi:10.1037/1089-2680.2.3.300
Frone, M. R., Russell, M., & Cooper, M. L. (1992). Antecedents and outcomes of work-family
conflict: Testing a model of the work-family interface. Journal of Applied Psychology, 77,
65–78. doi:10.1037/0021-9010.77.1.65
Fylkesnes, K. (1993). Determinants of health care utilization: Visits and referrals. Scandinavian
Journal of Social Medicine, 21, 40 –50. doi:10.1177/140349489302100107
Galea, S., Ahern, J., Resnick, H., Kilpatrick, D., Bucuvalas, M., Gold, J., & Vlahov, D. (2002).
Psychological sequelae of the September 11 terrorist attacks in New York City. The New
England Journal of Medicine, 346, 982–987. doi:10.1056/NEJMsa013404
Giesler, R. B., Josephs, R. A., & Swann, W. B. (1996). Self-verification in clinical depression:
The desire for negative evaluation. Journal of Abnormal Psychology, 105, 358 –368.
doi:10.1037/0021-843X.105.3.358
Gimbel, C., & Booth, A. (1994). Why does military combat experience adversely affect marital
relations? Journal of Marriage and the Family, 56, 691–703. doi:10.2307/352879
Gold, J. I., Taft, C. T., Keehn, M. G., King, D. W., King, L. A., & Samper, R. E. (2007). PTSD
symptom severity and family adjustment among female Vietnam veterans. Military
Psychology, 19, 71– 81. doi:10.1080/08995600701323368
Gorgievski, M. J., & Hobfoll, S. E. (2009). Work can burn us out or fire us up: Conservation
of resources in burnout and engagement In J. R. B. Halbesleben (Ed.), Handbook of stress
and burnout in health care. Hauppauge, NY: Nova Science Publishers.
Halbesleben, J. R. B. (2006). Sources of social support and burnout: A meta-analytic test of the
conservation of resources model. Journal of Applied Psychology, 91, 1134–1145. doi:
10.1037/0021-9010.91.5.1134
Halbesleben, J. R. B., Harvey, J., & Bolino, M. C. (2009). Too engaged? A conservation of
resources view of the relationship between work engagement and work interference with
family. Journal of Applied Psychology, 94, 1452–1465. doi:10.1037/a0017595
Hobfoll, S. E. (1988). The ecology of stress. Washington, DC: Hemisphere.
Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress.
American Psychologist, 44, 513–524. doi:10.1037/0003-066X.44.3.513
Hobfoll, S. E. (1991). Traumatic stress: A theory based on rapid loss of resources. Anxiety
Research, 4, 187–197. doi:10.1080/08917779108248773
Hobfoll, S. E. (1998). Stress, culture and community: The psychology and philosophy of stress.
New York, NY: Plenum Press.
Hobfoll, S. E. (2001). The influence of culture, community, and the nested-self in the stress
process: Advancing conservation of resources theory. Applied Psychology: An Interna-
tional Review, 50, 337– 421. doi:10.1111/1464-0597.00062
Hobfoll, S. E. (2011). Conservation of resource caravans and engaged settings. Journal of
Occupational and Organizational Psychology, 84, 116 –122. doi:10.1111/j.2044-
8325.2010.02016.x
Hobfoll, S. E., & Lilly, R. S. (1993). Resource conservation as a strategy for community
psychology. Journal of Community Psychology, 21, 128 –148. doi:10.1002/1520-
6629(199304)21:2⬍128::AID-JCOP2290210206⬎3.0.CO;2-5
Hobfoll, S. E., & Shirom, A. (1993). Stress and burnout in the workplace: Conservation of
resources. In R. T. Golembiewski (Ed.), Handbook of organization behavior (pp. 41– 46).
New York, NY: Marcel Dekker.
Hoge, C. W., Auchterlonie, J. L., & Milliken, C. S. (2006). Mental health problems, use of
mental health services, and attrition from military service after returning from deployment
to Iraq or Afghanistan. Journal of the American Medical Association, 295, 1023–1032.
doi:10.1001/jama.295.9.1023
Hoge, C. W., Castro, C. A., Messer, S. C., McGurk, D., Cotting, D. I., & Koffman, R. L. (2004).
235Warriors Life Stress
Combat duty in Iraq and Afghanistan, mental health problems, and barriers to care. The
New England Journal of Medicine, 351, 13–22. doi:10.1056/NEJMoa040603
Hoge, C. W., Terhakopian, A., Castro, C. A., Messer, S. C., & Engel, C. C. (2007). Association
of posttraumatic stress disorder with somatic symptoms, health care visits, and absentee-
ism among Iraq war veterans. The American Journal of Psychiatry, 164, 150 –153.
doi:10.1176/appi.ajp.164.1.150
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.
doi:10.1080/10705519909540118
Jakupcak, M., Luterek, J., Hunt, S., Conybeare, D., & McFall, M. (2008). Posttraumatic stress
and its relationship to physical health functioning in a sample of Iraq and Afghanistan war
veterans seeking post deployment VA health care. Journal of Nervous and Mental
Disease, 196, 425– 428. doi:10.1097/NMD.0b013e31817108ed
Kandel, D. B., Davies, M., & Raveis, V. H. (1985). The stressfulness of daily social roles for
women: Martial, occupational and household roles. Journal of Health and Social Behav-
ior, 26, 64 –78. doi:10.2307/2136727
Keane, T. M., Fairbank, J. A., Caddell, J. M., Zimering, R. T., Taylor, K. L., & Mora, C. A.
(1989). Clinical evaluation of a measure to assess combat exposure. Psychological
Assessment: A Journal of Consulting and Clinical Psychology, 1, 53–55. doi:10.1037/
1040-3590.1.1.53
Kessler, R. C. (2000). Posttraumatic stress disorder: The burden to the individual and to society.
Journal of Clinical Psychiatry, 61 (Suppl. 5), 4 –14.
King, D. W., King, L. A., Foy, D. W., Keane, T. M., & Fairbank, J. A. (1999). Posttraumatic
stress disorder in a national sample of female and male Vietnam veterans: Risk factors,
war-zone stressors, and resilience-recovery variables. Journal of Abnormal Psychology,
108, 164 –170. doi:10.1037/0021-843X.108.1.164
Kulka, R. A., Schlenger, W. E., Fairbank, J. A., Hough, R. L., Jordan, K. B., Marmar, C. R.,
Grady, D. A. (1990). Trauma and the Vietnam War generation: Report of findings from
the National Vietnam Veterans Readjustment Study. New York, NY: Brunner/Mazel.
Laufer, R. S., Gallops, M. S., & Frey-Wouters, E. (1984). War stress and trauma: The Vietnam
veteran experience. Journal of Health and Social Behavior, 25, 65– 85. doi:10.2307/
2136705
McCullough, M. E., & Laurenceau, J. P. (2004). Gender and the natural history of self-rated
health: A 59-year longitudinal study. Health Psychology, 23, 651– 655. doi:10.1037/0278-
6133.23.6.651
McDade, T. W., Hawkley, L. C., & Cacioppo, J. T. (2006). Psychosocial and behavioral
predictors of inflammation in middle-aged and older adults: The Chicago Health, Aging,
and Social Relations study. Psychosomatic Medicine, 68, 376 –381. doi:10.1097/
01.psy.0000221371.43607.64
Miller, G. E., Freedland, K. E., & Carney, R. M. (2005). Depressive symptoms and the
regulation of proinflammatory cytokine expression in patients with coronary heart disease.
Journal of Psychosomatic Research, 59, 231–236. doi:10.1016/j.jpsychores.2005.06.004
Neria, Y., & Koenen, K. C. (2003). Do combat stress reaction and posttraumatic stress disorder
relate to physical health and adverse health practices? An 18-year follow-up of Israeli war
veterans. Anxiety, Stress, & Coping: An International Journal, 16, 227–239. doi:10.1080/
1061580021000069407
Schaufeli, W. B., & Bakker, A. B. (2004). Job demands, job resources, and their relationship
with burnout and engagement: A multi-sample study. Journal of Organizational Behavior,
25, 293–315. doi:10.1002/job.248
Selig, J. P., & Preacher, K. J. (2008, June). Monte Carlo Method for Assessing Mediation: an
Interactive Tool for Creating Confidence Intervals for Indirect Effects [Computer soft-
ware]. Retrieved from http://quantpsy.org/
Singh-Manoux, A., Martikainen, P., Ferrie, J., Zins, M., Marmot, M., & Goldberg, M. (2006).
236 Hobfoll, Vinokur, Pierce, and Lewandowski-Romps
What does self rated health measure? Results from the British Whitehall II and French
Gazel cohort studies. Journal of Epidemiology and Community Health, 60, 364 –372.
doi:10.1136/jech.2005.039883
Solomon, Z., Mikulincer, M., & Hobfoll, S. E. (1986). Effects of social support and battle
intensity on loneliness and breakdown during combat. Journal of Personality and Social
Psychology, 51, 1269 –1276. doi:10.1037/0022-3514.51.6.1269
Stewart, A. L., & Ware, J. E. (1992). Measuring functioning and well-being: The medical
outcome study approach. (pp. 345–371). Durham, NC: Duke University Press.
Tanielian, T., & Jaycox, L. H. (2008). Invisible wounds of war: Psychological and cognitive
injuries, their consequences, and services to assist recovery. Retrieved from http://
www.rand.org/pubs/monographs/2008/RAND_MG720.pdf
Treiblmaier, H., Bentler, P. M., & Mair, P. (2011). Formative constructs implemented via
common factors. Structural Equation Modeling, 18, 1–17. doi:10.1080/
10705511.2011.532693
Vasterling, J. J., Proctor, S. P., Friedman, M. J., Hoge, C. W., Heeren, T., King, L. A., & King,
D. W. (2010). PTSD symptom increases in Iraq-deployed soldiers: Comparison with
nondeployed soldiers and associations with baseline symptoms, deployment experiences,
and postdeployment stress. Journal of Traumatic Stress, 23, 41–51. doi:10.1002/jts.20487
Vinokur, A., & Caplan, R. D. (1987). Attitudes and social support: Determinants of job-seeking
behavior and well-being among the unemployed. Journal of Applied Social Psychology,
17, 1007–1024. doi:10.1111/j.1559-1816.1987.tb02345.x
Vinokur, A. D., Pierce, P. F., Lewandowski-Romps, L., Hobfoll, S., & Galea, S. (2011). Effects
of war exposure on Air Force personnel’s mental health job burnout and other organiza-
tional related outcomes. Journal of Occupational Health Psychology, 16, 3–17.
Vinokur, A. D., Price, R. H., & Schul, Y. (1995). Impact of the JOBS intervention on
unemployed workers varying in risk for depression. American Journal of Community
Psychology, 23, 39 –74. doi:10.1007/BF02506922
Vogt, D. S., King, D. W., King, L. A., Savarese, V. W., & Suvak, M. K. (2004). War-zone
exposure and long-term general life adjustment among Vietnam veterans: Findings from
two perspectives. Journal of Applied Social Psychology, 34, 1797–1824. doi:10.1111/
j.1559-1816.2004.tb02586.x
Weathers, F. W., Huska, J. A., & Keane, T. M. (1991). PTSD Checklist-Military Version.
Boston, MA: National Center for PTSD, Behavioral Science Division.
Wells, J. D., Hobfoll, S. E., & Lavin, J. (1999). When it rains it pours: The greater impact of
resource loss compared to gain on psychological distress. Personality and Social Psy-
chology Bulletin, 25, 1172–1182. doi:10.1177/01461672992512010
Williams, L. J., & O’Boyle Jr., E. H. (2008). Measurement models for linking latent variables
and indicators: A review of human resource management research using parcels. Human
Resource Management Review, 18, 233–242. doi:10.1016/j.hrmr.2008.07.002
Zwiebach, L., Rhodes, J., & Roemer, L. (2010). Resource loss, resource gain, and mental health
among survivors of Hurricane Katrina. Journal of Traumatic Stress, 23, 751–758. doi:
10.1002/jts.20579
Received June 2, 2011
Revision received May 21, 2012
Accepted May 22, 2012 䡲
237Warriors Life Stress
A preview of this full-text is provided by American Psychological Association.
Content available from International Journal of Stress Management
This content is subject to copyright. Terms and conditions apply.
