Resilience, Traumatic Brain Injury, Depression, and Posttraumatic StressAmong Iraq/Afghanistan War Veterans

Timothy R. Elliott and Yu-Yu HsiaoTexas A&M University

Nathan A. KimbrelDurham Veterans Affairs Medical Center, Durham, North

Carolina and Duke University Medical Center

Eric C. Meyer and Bryann B. DeBeerVA VISN 17 Center of Excellence for Research on Returning

War Veterans, Central Texas Veterans HealthCare System,Waco, Texas, and Texas A&M University College of Medicine

Suzy Bird GulliverWarriors Research Institute at Baylor Scott and White Health,

Waco, Texas, and Texas A&M University College of Medicine

Oi-Man KwokTexas A&M University

Sandra B. MorissetteVA VISN 17 Center of Excellence for Research on Returning

War Veterans, Central Texas Veterans HealthCare System,Waco, Texas, and Texas A&M University College of Medicine

Objective: We examined the prospective influence of the resilient, undercontrolled, and overcontrolledpersonality prototypes on depression and posttraumatic stress disorder (PTSD) symptoms among Iraq/Afghanistan war veterans. After accounting for the possible influence of combat exposure, we expectedthat the resilient prototype would predict lower depression and PTSD over time and would be associatedwith adaptive coping strategies, higher social support, lower psychological inflexibility, and higherself-reported resilience relative to overcontrolled and undercontrolled prototypes, independent of trau-matic brain injury (TBI) status. Method: One hundred twenty-seven veterans (107 men, 20 women;average age � 37) participated in the study. Personality was assessed at baseline, and PTSD anddepression symptoms were assessed 8 months later. Path analysis was used to test the direct and indirecteffects of personality on distress. Results: No direct effects were observed from personality to distress.The resilient prototype did have significant indirect effects on PTSD and depression through its beneficialeffects on social support, coping and psychological inflexibility. TBI also had direct effects on PTSD.Conclusions: A resilient personality prototype appears to influence veteran adjustment through itspositive associations with greater social support and psychological flexibility, and lower use of avoidantcoping. Low social support, avoidant coping, and psychological inflexibility are related to overcontrolledand undercontrolled personality prototypes, and these behaviors seem to characterize veterans whoexperience problems with depression and PTSD over time. A positive TBI status is directly andprospectively associated with PTSD symptomology independent of personality prototype. Implicationsfor clinical interventions and future research are discussed.

Keywords: resilience, PTSD, depression, combat exposure, traumatic brain injury

This article was published Online First July 27, 2015.Timothy R. Elliott and Yu-Yu Hsiao, Department of Educational

Psychology, Texas A&M University; Nathan A. Kimbrel, DurhamVeterans Affairs Medical Center, Durham, North Carolina, and Depart-ment of Psychiatry and Behavioral Sciences, Duke University MedicalCenter; Eric C. Meyer and Bryann B. DeBeer, VA VISN 17 Center ofExcellence for Research on Returning War Veterans, Central TexasVeterans HealthCare System, Waco, Texas, and Department of Psychi-atry, Texas A&M University College of Medicine; Suzy Bird Gulliver,Warriors Research Institute at Baylor Scott and White Health, Waco,Texas, and Department of Psychiatry, Texas A&M University Collegeof Medicine; Oi-Man Kwok, Department of Educational Psychology,Texas A&M University; Sandra B. Morissette, VA VISN 17 Center of

Excellence for Research on Returning War Veterans, Central TexasVeterans HealthCare System, and Department of Psychiatry, TexasA&M University College of Medicine.

This study is one in a series conducted from a larger project investigatingadjustment of veterans from Operation Enduring Freedom and OperationIraqi Freedom. A complete list of publications from the project is availableupon request from Dr. Morissette at Sandra.Morissette@va.gov.

This work was completed with support from the Veterans Health Ad-ministration. The views are those of the authors and do not necessarilyrepresent the views of the Department of Veterans Affairs.

Correspondence concerning this article should be addressed to TimothyR. Elliott, PhD, Department of Educational Psychology, 4225 TAMU,College Station, TX 77843-4225. E-mail: telliott@tamu.edu

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

Rehabilitation Psychology © 2015 American Psychological Association2015, Vol. 60, No. 3, 263–276 0090-5550/15/$12.00 http://dx.doi.org/10.1037/rep0000050

263

Impact and Implications• Although resilience is often discussed as a beneficial factor in the

posttraumatic stress disorder (PTSD) and traumatic brain injuryliterature, the present study is the first to examine a resilientpersonality prototype and its association with adjustment overtime among veterans with these conditions.

• The results of the present study indicate that screening withpersonality instruments to determine resilient, undercontrolled,and overcontrolled traits may be used to assess therapeutic prog-nosis and to potentially allocate treatment resources to veteranswho may be particularly at risk for chronic, unremitting problemsover time.

• For veterans with overcontrolled and undercontrolled personalityprofiles, social support, avoidant coping and psychological inflex-ibility are modifiable factors that can be addressed in treatment toreduce symptoms of distress associated with PTSD anddepression.

Introduction

Posttraumatic stress disorder (PTSD) and depression are fre-quently experienced by Iraq/Afghanistan war veterans. Evidenceindicates that 12% to 20% of Iraq/Afghanistan war veterans havea diagnosis of PTSD and about 15% endorse symptoms of depres-sion (Hoge et al., 2004; Seal, Bertenthal, Miner, Sen, & Marmar,2007). Approximately 23% of Iraq/Afghanistan veterans developclinically significant levels of PTSD; Fulton et al., 2015). Higherrates of PTSD and other mental health problems have been ob-served among veterans who served in Iraq than among thoseserving in Afghanistan (Ramchand et al., 2010). It is important tonote that depression and PTSD share several overlapping symp-toms and often co-occur among Iraq/Afghanistan veterans. Adiagnosis of PTSD and/or depression among these war veterans isassociated with functional difficulties including occupational andsocial impairment, physical health problems, and substance use(Brenner et al., 2010; Hoge et al., 2004; Thomas et al., 2010).

These disorders are further complicated by co-occurring trau-matic brain injuries (TBIs) commonly incurred through blast ex-posures. Approximately 15% to 20% of returning service membersreport a probable TBI during their deployment (Belanger, Uomoto,& Vanderploeg, 2009; Hoge et al., 2008). The Armed ForcesHealth Surveillance Center (2011) identified 212,742 cases of TBIin all armed forces members, classifying 76.7% as mild, 16.8% asmoderate, and 1.1% as severe. Concomitants of TBI often includesymptoms of depression and PTSD (Morissette et al., 2011).

The complexities and confounds of co-occurring PTSD and TBIon clinical assessment—and the corresponding need to identifyfactors that complicate or enhance recovery—are widely acknowl-edged (Bahraini et al., 2014; Vasterling, Bryant, & Keane, 2012).It is preferable to study these factors within contextual models thatexamine optimal adjustment among veterans despite existing in-dicators of risk (e.g., combat exposure; Dolan et al., 2012). Suchadjustment is likely influenced by an array of psychosocial andcontextual variables (Vasterling & Dikmen, 2012) and their influ-ence may be best investigated longitudinally in studies that includea wide range of veterans known to have either PTSD or TBI, both,or neither (Bahraini et al., 2014).

The empirical pursuit of protective factors is described as thestudy of resilience (Bonanno et al., 2012). The majority of return-

ing veterans report few if any psychological problems followingwarzone deployment. Preinjury self-reported resilience (McCauleyet al., 2013) and protective psychological characteristics (e.g.,attributions about symptoms, self-efficacy; Belanger, Barwick,Kip, Kretzmer, & Vanderploeg, 2013) are associated with lessdistress among individuals with mild TBI. Resilience appears tohave beneficial effects on adjustment among persons with chronicconditions by facilitating the experience of positive emotions andinfluencing cognitive appraisals of symptoms (Ong, Zautra, &Reid, 2010). Collectively, these studies support prevailing notionsof resilience including the “ability to sustain equilibrium andadaptive functioning under stressful circumstances” (Mancini &Bonanno, 2010, p. 259) and the ability to “bounce back” ordemonstrate “‘better than expected’ development or functioning”(Windle & Bennett, 2012, p. 219) in the wake of a negative lifeevent.

Critical reviews of the resilience literature observe that thelack of a comprehensive, well-articulated theoretical modelundermines the ability to identify the key mechanisms andcharacteristics of resilience under specific situations and todevelop strategic interventions to promote resilience (Davydov,Stewart, Ritchie, & Chandieu, 2010; Fletcher & Sarkar, 2013).Of the available models, the Block and Block (1980) theory ofego control and ego resiliency—two psychological factors re-lated to the capacity for effective adaptation to change andconflict (Block, 1993)— explicates specific personality proto-types that appear particularly advantageous for clinical appli-cations. The three prototypes—resilient, undercontrolled, andovercontrolled—are mapped out with behavioral ratings (instudies of children; Caspi & Silva, 1995) and with self-reportmeasures of personality traits from adolescents and adults(Chapman & Goldberg, 2011; Dennissen, Asendorpf, & vanAken, 2008; Letzring, Block, & Funder, 2005). Resilient indi-viduals are often typified by low neuroticism (or negativeaffectivity) and above-average scores on other personalitytraits. Undercontrolled individuals are characterized by lowconscientiousness and a moderate level of neuroticism. Over-controlled individuals are characterized by high neuroticism,low extraversion, and average scores on other factors.

The predictive utility of these personality prototypes is sup-ported by prospective studies of behavioral ratings of childrenin which the prototypes predicted emotional, behavioral, andhealth outcomes in adulthood (Caspi & Silva, 1995; Chapman& Goldberg, 2011; Dennissen et al., 2008). Resilient individu-als reported less distress and aggression than those classified aseither undercontrolled or overcontrolled. Resilient individualsfurther assume some normative adult social roles earlier (Den-nissen et al., 2008) and have a lower cardiovascular disease riskin middle age (Chapman & Goldberg, 2011). Cross-sectionalstudies with adults reveal that resilient individuals report moreadaptive, proactive problem-solving styles following the onsetof a traumatic disability, as well as a greater sense of acceptanceat medical discharge than overcontrolled individuals (Berry,Elliott, & Rivera, 2007). In contrast, compared with resilientindividuals, overcontrolled individuals report higher depressionfollowing disability onset (Berry et al., 2007) and are moresocially isolated and engage in fewer recreational pursuits inolder age (Steca, Alessandri, & Caprara, 2010). Undercon-trolled adults report more mistrust of others, including family

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

264 ELLIOTT ET AL.

members, than resilient adults (Steca et al., 2010). In summary,both theory and prior research assert that a resilient personalityprototype may be viewed as a “protective trait” that can “ac-count for stress resistance or the maintenance of positive out-comes in the face of challenge” (Ong, Bergeman, & Boker,2009, p. 1784).

In the present study, we used a prospective design and a con-textual model to examine how resilient, undercontrolled, and over-controlled prototypes predict distress over time among Iraq/Afghanistan war veterans. We expected the resilient prototype tobe associated with less distress over time, whereas we expected theovercontrolled prototype to be associated with greater distress. Themodel was construed to examine several mechanisms that mayelucidate the ways in which a resilient prototype may facilitateadjustment over time. Based on our understanding of the charac-teristics that typify resilience, we hypothesized that a resilientprototype would be associated with greater perceived social sup-port, greater sense of personal resilience, greater psychologicalflexibility (and less experiential avoidance), and more effectivecoping behaviors compared with the overcontrolled and undercon-trolled prototypes. We further hypothesized that these mediatingvariables would, in turn, be associated with lower depression andPTSD symptoms. In contrast, we expected that veterans charac-terized as undercontrolled and overcontrolled would experiencedifficulty marshalling and utilizing social support, and wouldreport fewer psychological resources, and rely on ineffectual cop-ing strategies.

In our contextual model we also examined whether a positivehistory of TBI at baseline would have a direct and deleteriousrelationship with depression and PTSD over time, as well asindirect effects through a negative association with psychologicalresources such as social support, self-reported resilience, coping,and psychological flexibility. Thus, our a priori model was de-signed to detect any potential effects of personality factors inde-pendent of TBI status on the mediating and outcome variables.Finally, we expected that the degree of combat exposure partici-pants reported at baseline would influence the primary outcomevariables of PTSD and depression at the eight-month follow-up.Therefore, combat exposure was included as a covariate to accountfor any variance in the self-report measures attributable to thisfactor.

Method

Participants and Procedure

The present study is a secondary analysis of an initial sample of145 Iraq/Afghanistan war veterans enrolled in a larger projectexamining war zone experiences and postdeployment adjustment(titled Project SERVE). Several studies have been conducted fromthe project and a complete list is available upon request from thelast author. The current study is the first from the project toexamine the association of personality prototypes to subsequentadjustment.

Participants were recruited through randomly selected mailingsto veterans enrolled in the Central Texas Veterans Health CareSystem (CTVHCS) system, with advertisements at recruitmentsites and veterans’ service organizations, and during in-servicepresentations to VA staff. Eligible participants were required to be

enrolled in the CTVHCS; however, there was no requirement thatthe veteran be receiving treatment. Prospective participants wereeligible if they were (a) an Iraq/Afghanistan war veteran, (b) ableto provide informed consent, and (c) able to complete measures inthe baseline assessment. Exclusion criteria included (a) a diagnosisof bipolar or psychotic disorder, (b) recently initiated psychiatricmedications or psychotherapy, or (c) suicidality or homicidalitywarranting crisis intervention. Initial eligibility was determined ina phone interview and later confirmed during an in-person inter-view.

Participants who responded to study recruitment strategies werescreened by telephone using a scripted interview that inquiredabout inclusion/exclusion criteria. Participants who were deemedinitially eligible from the screen were scheduled for an in-personassessment at the beginning of which informed written consentprocedures were completed. Inclusion/exclusion criteria were con-firmed during the baseline interview (e.g., diagnostic assessmentof exclusionary diagnoses). Diagnostic interviews were completedby trained master’s level assessors. Diagnostic consensus wasreached on all interviews under the supervision of a licensedclinical psychologist.

Of the 145 who enrolled in the study, 8 were excluded for adiagnosis of schizophrenia or bipolar disorder and 9 were excludeddue to missing all the items in the Multidimensional PersonalityQuestionnaire (MPQ). For the remaining 128 participants, wefurther used the criteria proposed by Patrick, Curtin, and Tellegen(2002) to identify any invalid responses with two validation scales:the Variable Response Inconsistency and the True Response In-consistency scales. One participant who did not meet the valid casecriteria was excluded. The final number of participants for subse-quent analyses is 127. The average age of this sample was 37.64(SD � 10.54) years old and the average years of education was14.17 (SD � 2.542) years. The majority were male (n � 107). Thedistribution of the participants in terms of race was as follows:White/Caucasian (n � 80), African American (n � 23), and other(n � 24). With respect to ethnicity, 33 participants identified asHispanic. Almost two thirds of the final sample (n � 82; 64%) hadservice-connected pensions for a disability.

Participants were assessed at three separate time points: a base-line interview and two follow-up mailing assessments four andeight months later, as described in detail below and depicted inFigure 2.

Measures of Primary Predictor Variables

Personality prototypes. The Multidimensional PersonalityQuestionnaire Brief Form (MPQ; Patrick et al., 2002) was admin-istered at baseline. The MPQ contains 155 items (with a true/falseresponse format) and is composed of 11 primary trait scales:Wellbeing, Social Potency, Achievement, Social Closeness, StressReaction, Aggression, Alienation, Control, Harm Avoidance, Tra-ditionalism, and Absorption. The 11 trait scales feature threehigher order dimensions that correspond to constructs of emotionand temperament: Positive Emotionality (PEM), Negative Emo-tionality (NEM), and Constraint (CON). These three dimensionsassess the defining characteristics of the resilient prototype. HigherPEM scores reflect a positive and active engagement with theenvironment and a capacity to experience positive emotions (e.g.,enthusiasm, zest). Higher scores on NEM reflect a predisposition

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

265VETERAN RESILIENCE

for anger, resentment, anxiety and negatively valenced emotionalrelationships. Higher CON scores reflect a penchant for caution,harm-avoidance, and convention; lower scores suggest impulsivityand sensation-seeking tendencies. In terms of the Big Five person-ality traits, the NEM scale may be construed as a combination ofthe Neuroticism factor and low Agreeableness, PEM may beconsidered a combination of Extraversion and Achievement Mo-tivation from the Conscientiousness factor, and CON reflects Con-scientiousness and aspects of Openness to Experience (Tellegen &Waller, 2008).

The MPQ Brief Form has been used in prior studies of resiliencyin adult development (Shiner, Masten, & Roberts, 2003) andtraumatic stress (Miller & Harrington, 2011). For this study, theCronbach’s alpha coefficient for the MPQ was .76. Alpha coeffi-cients for the 11 primary trait scales in the present study rangedfrom .64 (Traditionalism) and .69 (Absorption) to .88 (Wellbeing)and .89 (Closeness).

Assessment of traumatic brain injury. Veterans werescreened at baseline for possible deployment-related TBI with theBrief Traumatic Brain Injury Screen (Schwab et al., 2007). Thescreening interviews were conducted by trained masters ordoctoral-level clinical interviewers. Veterans screened positive forTBI based on endorsement of a head injury during deployment(e.g., from a blast, vehicular accident, fall, bullet, fragment) thatled to an alteration of consciousness (e.g., being disoriented,“dazed,” or confused, seeing “stars”), loss of consciousness, orposttraumatic amnesia. Fifty-seven veterans (44.8% of the sample)screened positive for TBI.

Combat exposure. The 18-item Full Combat Exposure Scale(FCES; Hoge et al., 2004) assesses combat experiences duringwarzone deployment (e.g., being attacked or ambushed, handlinghuman remains, incoming mortars). Among Iraq/Afghanistan vet-erans, scores on the FCES are strongly associated with PTSDsymptoms (Hoge et al., 2004). Participants completed the FCESduring the baseline assessment. Internal consistency was .92 in thecurrent study.

Measures of Psychological Resources andMediating Variables

Self-reported resilience. Veterans completed the Connor–Davidson Resilience Scale (CDRISC; Connor & Davidson, 2003)at baseline. This measure contains 25 items rated on a 5-pointLikert-type scale ranging from 0 (not true at all) to 4 (true nearlyall of the time). Higher total scores reflect greater resilience. Thepsychometric properties of the CDRISC are among the best incomparisons with other available resilience measures (Windle,Bennett, & Noyes, 2011). Internal consistency was .96. Sampleitems include “I am able to adapt when change occurs,” “I can dealwith whatever comes my way,” “I like challenges,” and “I tend tobounce back after illness, injury or other hardships.”

Psychological inflexibility. The Acceptance and ActionQuestionnaire–II (AAQ-II; Bond et al., 2011) assessed baselinelevels of psychological inflexibility and experiential avoidance,which refers to being unwilling or unable to remain in contact withunwanted private experiences (e.g., cognitions, emotions, physio-logical sensations) and taking steps to avoid the experience (Hayeset al., 2004). The AAQ-II is a seven-item self-report measure thatuses a 7-point Likert-type scale (1 � never true, 7 � always true).

Sample items include “I am afraid of my feelings” and “Emotionscause problems in my life.” Higher total scores indicate greaterpsychological inflexibility.

The AAQ-II exhibits a single-factor structure, good internalconsistency, good test—retest reliability, and good convergentassociations with the original AAQ (Bond et al., 2011). AAQ-IIscores have been uniquely predictive of PTSD severity in cross-sectional research with a subset of the current sample (Meyer,Morissette, Kimbrel, Kruse, & Gulliver, 2013) and in prospectiveresearch with college students exposed to a campus shooting(Kumpula, Orcutt, Bardeen, & Varkovitzky, 2011). Internal con-sistency was .94 in the current study.

Social support. The DRRI-Postdeployment Social Supportscale (PDSS) from the Deployment Risk and Resilience Inventory(King, King, Vogt, Knight, & Samper, 2006) was used to assessperceived availability of social support. The PDSS was adminis-tered four months after the initial assessment. This 15-item self-report scale uses a 5-point Likert-type scale from 1 (stronglydisagree) to 5 (strongly agree). The scale was validated for usewith Iraq/Afghanistan veterans (Vogt, Proctor, King, King, &Vasterling, 2008). The PDSS assesses perceived emotional andinstrumental forms of social support from family, friends, employ-ers and the community following deployment. The scale has dem-onstrated good internal consistency (.87) in prior research withveterans (Vogt et al., 2008). Internal consistency was .90 in thepresent study.

Coping. The Brief COPE (B-COPE, Carver, 1997) assessedcoping styles at the 4-month assessment. The Brief COPE is ashortened version of the Full COPE (Carver, Scheier, & Wein-traub, 1989). The Brief COPE consists of 28 items (two items eachin 14 subscales). Responses on the Brief COPE are rated on afour-point Likert-type scale that ranges from 0 (I haven’t beendoing this at all) to 3 (I’ve been doing this a lot). Following aprevious factor analysis of the B-COPE (Grosso et al., 2014), twofactors were created by summing the subscales of the B-COPE:action-oriented and avoidant coping. Higher scores on both scalesreflect a greater proclivity for that coping style. Internal coeffi-cients were .83 and .73 for the action-oriented coping and theavoidant coping factors, respectively.

Mental Health Symptoms

Depression. The Beck Depression Inventory–II (BDI-II;Beck, Steer, & Brown, 1996) was administered eight months afterthe initial assessment. The BDI-II is a widely used self-reportmeasure that contains 21 items rated on a 4-point Likert-type scale.Higher scores indicate greater severity of depressive symptoms.The BDI-II has demonstrated excellent psychometric properties inprior research (Beck, Steer, Ball, & Ranieri, 1996; Beck, Steer, &Brown, 1996). Internal consistency was .96 in the current study.

PTSD symptoms. The PTSD Checklist–Military Version(PCL-M; Weathers, Litz, Herman, Huska, & Keane, 1993) wasadministered eight months after the initial assessment. The 17-itemPCL-M is a common self-report measure of military-related PTSDsymptoms experienced during the past month. The PCL-M hasexcellent internal consistency and validity (Blanchard, Jones-Alexander, Buckley, & Forneris, 1996). Across studies of veterans,PCL-M cutoff scores to screen for PTSD range from 28 to 60(Keen, Kutter, Niles, & Krinsley, 2008). Generally, a cutoff score

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

266 ELLIOTT ET AL.

of 50 is used among studies of male Vietnam veterans (Forbes,Creamer, & Biddle, 2001; Weathers et al., 1993). Internal consis-tency was .97.

Results

Cluster Analysis

Cluster analysis was conducted with SPSS 22.0 to identify thethree personality prototypes. In doing hierarchical clustering, eachparticipant was at first treated as one independent cluster andcombined in pairs in a series of steps. Ward’s method along withsquared euclidean distance measures were used to minimize thetotal within-cluster sums of squares for each step (Sharma, 1996).By observing the largest shift in the total within-cluster sums ofsquares between steps, the number of clusters can be determined(Aldenderfer & Blashfield, 1984). After defining the number ofclusters, the initial cluster centers from Ward’s method were usedin the follow-up nonhierarchical K-means analysis to optimize thecluster classification. Additionally, the level of agreement betweenthe preliminary Ward’s method and the K-mean classificationresults were estimated by Cohen’s Kappa. We adopt the guidelineproposed by Landis and Koch (1977) in which Kappa valuesbetween .41 to .60 as moderate, .61 to .80 as substantial, and �.81as almost perfect agreement. Finally, we labeled the clusters byinspecting the PEM, NEM, and CON average scores for eachcluster.

The agglomeration schedule shows that the largest change inwithin-cluster sum of squares existed between the 124th stepand the 125th step. Therefore, the optimal cluster solution existswhen the participants combined till the 127th step, indicatingthat the suggested number of clusters is 127 � 124 � 3. Wefurther applied the three-cluster solutions from the Waldmethod to the K-means clustering analysis to assign each par-ticipant into one of the three clusters. Figure 1 presents thepersonality prototypes (using z scores for ease of interpretation)of the three clusters obtained from K-means cluster results. Theovercontrolled group (n � 27) was identified by its high levelof NEM and low level of PEM. The resilient group (n � 51)was identified as the group with the highest PEM and the lowestNEM. Finally, the undercontrolled group (n � 49) was charac-terized by the lowest level of Constraint. The Cohen’s Kappacoefficient for the agreement between the Ward and K-meansclassifications was .67, which is regarded as substantial agree-ment (Landis & Koch, 1977).

The pattern to the prototypes depicted in Figure 1 differs slightlyfrom a previous cluster analysis (Miller, Greif, & Smith, 2003) anda latent profile analysis (Wolf, Miller, Harrington, & Reardon,2012) of MPQ data obtained from veterans meeting screeningcriteria for PTSD. In these studies, a “low psychopathology”prototype had relatively average PEM, CON, and NEM scores(unlike the pronounced elevation on PEM and lower NEM thatcharacterize the resilient prototype in Figure 1). Perhaps differ-ences in inclusion and exclusion criteria account for some of these

Figure 1. Three personality prototypes based on the Multidimensional Personality Questionnaire in the sampleof 127 veterans. Resilient prototype � 40.1% (n � 51) of the sample, undercontrolled � 38.6% (n � 49);overcontrolled � 21.3% (n � 27). PEM � Positive Emotionality; NEM � Negative Emotionality; CON �Constraint. See the online article for the color version of this figure.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

267VETERAN RESILIENCE

variations: Participants in the Miller et al. (2003) and the Wolf etal. (2012) studies met screening criteria for PTSD and veteranswith psychotic or bipolar disorders, psychological crises, andrecently initiated therapy or medication were excluded from thepresent study.

Associations Among the Variables

The simple associations between personality type and the threetypes of manifest variables (i.e., the potential mediators, distressoutcomes, and covariates) are listed in Table 1. PEM and NEM

were significantly and inversely correlated, �.39 (p � .001).Similarly, NEM and CON were significantly correlated, �.18 (p �.05). The correlation between PEM and CON (.13) was not sig-nificant. The MPQ Unlikely Virtues scale (designed to detect“faking good”) was not significantly correlated with any of theother self-report variables (all rs ranged from �.16 to .14).

Results from one-way analyses of variance indicated signif-icant associations between personality prototype and all themanifest variables (all ps � .05). Follow-up post hoc compar-isons provided evidence of the mean score differences between

Figure 2. Saturated a priori model of personality prototypes, traumatic brain injury (TBI), and mediatingcharacteristics predicting posttraumatic stress disorder (PTSD) and depression eight months following baseline.Ref � reference group.

Table 1Mean Score Differences Among Mediators, Distress Outcomes, and the Covariate byPersonality Prototypes

Measure

Personality prototype

pPost hoc

test

Overcontrolled(Cluster 1)

Undercontrolled(Cluster 2)

Resilient(Cluster 3)

M SD M SD M SD

Mediators (4th month)Social support 47.10 9.78 47.63 8.56 57.06 8.38 �.001 1 & 2 � 3B-COPE_Action 10.08 3.92 7.91 3.42 10.20 3.97 .010 3 � 2B-COPE_Avoidant 4.24 2.73 3.20 1.97 1.77 1.96 � .001 1 & 2 � 3AAQ-II 30.26 11.74 27.17 9.92 15.71 7.84 � .001 1 & 2 � 3CDRISC 61.78 21.48 61.87 18.64 80.96 14.03 � .001 1 & 2 � 3

Distress outcomes (8th month)PCL-M 60.24 15.13 51.50 17.76 33.98 16.57 � .001 1 & 2 � 3BDI-II 32.13 12.66 24.20 12.99 11.08 9.84 � .001 1 � 2 � 3

Covariate (baseline)Combat exposure 31.08 16.48 23.26 14.98 18.34 11.56 .001 1 � 3

Note. B-COPE � Brief-COPE; B-COPE Action � Action coping score; B-COPE_Avoidant � Avoidantcoping score; AAQ-II � Acceptance and Action Questionnaire–II; CDRISC � Connor–Davidson ResilienceScale; PCL-M � PTSD Checklist–Military; BDI-II � Beck Depression Inventory–II.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

268 ELLIOTT ET AL.

any two of the personality subgroups. Both the overcontrolledand undercontrolled groups had higher psychological inflexi-bility and lower self-reported resilience at baseline, lower socialsupport at four months, and more PTSD symptoms at eightmonths compared with the resilient group. The overcontrolledgroup had the highest depression scores at the eighth month,followed by the undercontrolled and the resilient group. Theovercontrolled group had higher combat exposure scores thanthe resilient group, but these two groups were not significantlydifferent from the undercontrolled group. The resilient groupapplied more action-oriented coping strategies at four monthsthan the undercontrolled group. In contrast, the overcontrolledand undercontrolled groups reported a greater use of avoidantcoping than the resilient group.

TBI diagnosis was significantly associated with personality pro-totype (p � .01). A majority of individuals in the overcontrolledgroup screened positive for TBI (n � 17, 63%). In contrast, amajority of individuals in the resilient group screened negative forTBI (n � 37; 71%). A higher proportion of men screened positivefor TBI than women. No associations were found between genderand personality prototype. There was no significant associationbetween personality prototypes and the distribution of service-connected pensions. TBI was not associated with race or ethnicity.

Independent sample t tests revealed several significant differ-ences between those with a positive and negative TBI history.Individuals with TBI had higher psychological inflexibility andcombat exposure at baseline, lower social support at the fourthmonth, and more PTSD and depression symptoms at the eighthmonth than those without TBI (all ps � .01). TBI history was notsignificantly related to self-reported resilience at baseline or to thecoping strategy variables (i.e., active and avoidant) at the fourthmonth.

Correlations among all the mediators, mental health symptomoutcomes, and the covariate variables are provided in Table 2. Allof the mediating variables were associated with PTSD and depres-sion (p � .05), except for action-oriented coping. Lower self-reported resilience and higher psychological inflexibility at base-line and lower social support at the fourth month were significantlyassociated with higher PTSD and depression at the eighth month.Higher avoidant coping at the fourth month was also significantly

positively associated with PTSD and depression. Combat exposurewas significantly associated with the mediating variables (exceptaction-oriented coping) and with PTSD and depression.

Path Analysis

Path analysis was used to test the proposed contextual model(see Figure 2), in which psychological inflexibility and self-reported resilience at baseline and social support, action-orientedand avoidant coping at the fourth month were positioned as po-tential mediators of the relationships between personality proto-types and TBI diagnosis at baseline to PTSD and depression eightmonths later. Combat exposure was included as a covariate in themodel. Direct and indirect effects were examined in the modelwith Mplus 7.1 (Muthén & Muthén, 1998–2012). Missing datawere handled by conducting the full information maximum like-lihood methods to incorporate information from incomplete re-sponses and retain available information from all 127 participantsin the analyses. Additionally, 1,000 bootstrapping samples wereproduced to estimate the 95% confidence interval for all the pathcoefficients. Bootstrapping technique was applied to compensatethe significant test results based on small sample (i.e., 127 in thepresent study), especially in testing indirect effect (Preacher &Hayes, 2008).

To achieve an ideal model to interpret the data, we began witha saturated model in which all possible correlations between vari-ables were specified in the model depicted in Figure 2. Then weexcluded three nonsignificant paths based on the previous univar-iate association results. The three paths deleted in the correctedmodel (see Figure 3) were the paths from TBI to self-reportedresilience (the CDRISC) at baseline and to active and avoidantcoping at the fourth month.

Fit indices were examined to assess model fit. A chi-squaresignificance test with a p value greater than .05 indicates adequatemodel fit. Two incremental fit indices, the comparative fit index(CFI) and Tucker–Lewis index (TLI) were chosen; a value of CFIand TLI exceeding .95 indicates a good fit (Hu & Bentler, 1999;Yu, 2002). Furthermore, two absolute fit indices, the root-mean-square error of approximation (RMSEA) and the standardizedroot-mean-square residual (SRMR) were used for model evalua-

Table 2Correlations Among Mediators, Distress Outcomes, and Combat Exposure

Variable 1 2 3 4 5 6 7

Mediators (4th month)1. Social support2. B-COPE_Action .2633. B-COPE_Avoidant �.499 .1204. AAQ-II �.548 �.094 .5985. CDRISC .559 .406 �.369 �.611

Distress outcomes (8th month)6. PCL-M �.676 .088 .681 .704 �.4127. BDI-II �.689 .019 .646 .731 �.530 .861

Covariate (baseline)8. Combat exposure �.294 .029 .152 .431 �.198 .473 .468

Note. Correlations greater than |.15| are significant at p � .05. B-COPE � Brief-COPE scale; B-COPE_Ac-tion � Action coping score; B-COPE_Avoidant � Avoidant coping score; AAQ-II � Acceptance and ActionQuestionnaire–II; CDRISC � Connor–Davidson Resilience Scale; PCL-M � PTSD Checklist–Military; BDI-II � Beck Depression Inventory–II.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

269VETERAN RESILIENCE

tion. A model with a RMSEA value below .05 indicates good fit,whereas RMSEA values between .05 and .08 suggest acceptable fit(Browne & Cudeck, 1992). A cutoff value of .09 for SRMR alsosuggests good model fit (Hu & Bentler, 1999). The test of thecorrected model provided excellent fit indices: �2(3) � 1.453, p �.693; CFI � 1.00; TLI � 1.00; RMSEA � 0.000; SRMR � 0.011.Thus, the path coefficients yielded from the corrected model wereinterpretable.

Direct effects. Both the standardized and unstandardizedpath coefficients for the corrected model are displayed in Table3. Personality prototype was significantly associated with mostof the mediators at baseline and at the fourth month. Comparedwith the resilient group, the overcontrolled prototype was sig-nificantly associated with more psychological inflexibility andlower self-reported resilience at baseline, and with less socialsupport and greater avoidant coping at the fourth month (allps � .01). Similarly, the undercontrolled prototype was signif-icantly associated with greater psychological inflexibility andlower self-reported resilience at baseline, and with lower socialsupport, less action-oriented coping and more avoidant copingat the fourth month than the resilient group (all ps � .01). TBIwas not associated with any mediator at baseline or at the fourthmonth.

Most of the mediators were associated with the mental healthoutcomes. Greater psychological inflexibility at baseline andlower social support and more avoidant coping at the fourth

month significantly predicted higher PTSD and depression atthe eighth month (all ps � .001). However, self-reported resil-ience (as assessed by the CDRISC) at baseline and action-oriented coping at the fourth month was not significantly pre-dictive of either PTSD or depression. A positive TBI diagnosissignificantly predicted PTSD symptoms at the eighth monthassessment. Personality prototype was not directly predictive ofeither PTSD or depression. The final model accounted for about75% of the variance in PTSD (R2 � 0.75) and 73% of thevariance in depression.

Indirect effects. All possible indirect effects of personalitytype and TBI diagnosis on distress outcomes at the eighth monthare shown in Table 4. Personality prototype had indirect effectsthrough psychological inflexibility at baseline (p � .01) andthrough social support (p � .01) and avoidant coping at the fourthmonth (p � .05) to PTSD. Similarly, the indirect effect frompersonality prototype to depression at 8 months was significant viapsychological inflexibility at baseline (p � .05), social support(p � .05), and avoidant coping at the fourth month (p � .05).Consequently, the prospective relationship of the overcontrolledand undercontrolled personality prototypes to depression andPTSD are best understood in their effects on psychological inflex-ibility, social support and avoidant coping. No significant indirecteffect was observed for TBI diagnosis on PTSD or depression inthe final model.

Figure 3. Final model depicting significant paths to posttraumatic stress disorder (PTSD) and depressionsymptoms. All path coefficients are shown in standardized form. TBI � traumatic brain injury. � p � .05.�� p � .01. ��� p � .001.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

270 ELLIOTT ET AL.

Discussion

Consistent with our initial hypotheses, veterans characterized asresilient at the baseline assessment had greater perceived socialsupport, less avoidant coping, and less psychological inflexibilitythan veterans classified as overcontrolled or undercontrolled. Theyalso reported the lowest levels of PTSD and depression. Takinginto account the variance attributable to combat experience andTBI, a resilient personality prototype was not directly predictive ofdistress assessed eight months later. Rather, a resilient personalitywas prospectively predictive of lower PTSD and depression symp-toms over time through its beneficial influence on several adaptivecharacteristics, including greater social support and lower avoidantcoping and psychological inflexibility. Thus, the protective andadaptive qualities of a resilient personality maintained a positiveinfluence over time despite prior combat exposure and the com-mon occurrence of a head injury. In contrast, facets of an under-controlled and overcontrolled personality appear to complicate

adjustment in addition to the deleterious effects that may beassociated with increased combat exposure. The indirect effects ofpersonality on adjustment over time were independent of thesignificant association of TBI status with PTSD.

The resilient group was characterized by low levels of negativeemotionality and high levels of positive emotionality on the MPQ.Positive emotions may facilitate adjustment in times of stress asthey promote flexibility and an ability to attend to and integratenew information (Cohn, Fredrickson, Brown, Mikels, & Conway,2009; Fredrickson, 2013). Experiencing positive emotions canhelp people feel closer and more connected with others (Kok et al.,2013) and increased contact and support from important individ-uals is often a characteristic of resilience (Bonanno, 2005). Posi-tive emotions appear to serve as a buffer against depression intimes of stress (Fredrickson, Tugade, Waugh, & Larkin, 2003) andevidence suggests that they promote physical health (Fredrickson,Cohn, Coffey, Pek, & Finkel, 2008; Kok & Fredrickson, 2010).

Table 3Direct Effects of Corrected Path Model

Effect Unst. St.Unst. 95% CI

(Bootstrap)

Personality subtype (ref: resilient) ¡ mediatorsOVER ¡ social support�� �8.23 �0.34 [�13.60, �2.86]UNDER ¡ social support��� �8.01 �0.40 [�12.03, �4.00]OVER ¡ B-COPE_Action �0.52 �0.06 [�2.66, 1.62]UNDER ¡ B-COPE_Action�� �2.43 �0.31 [�3.89, �0.97]OVER ¡ B-COPE_Avoidant��� 2.48 0.44 [1.00, 3.96]UNDER ¡ B-COPE_Avoidant�� 1.37 0.29 [0.54, 2.11]OVER ¡ AAQ-II��� 11.28 0.41 [5.78, 16.77]UNDER ¡ AAQ-II��� 10.14 0.43 [0.54, 2.21]OVER ¡ CDRISC��� �17.92 �0.37 [�27.41, �8.44]UNDER ¡ CDRISC��� �18.92 �0.46 [�25.78, �12.05]

TBI diagnosis (ref: TBI negative) ¡ mediatorsTBI positive ¡ social support �1.63 �0.08 [�5.52, 2.25]TBI positive ¡ AAQ-II 1.25 0.06 [�1.81, 4.32]

Mediators ¡ Distress outcomesSocial support ¡ PCL-M��� �0.68 �0.35 [�1.02, �0.34]COPE_Action ¡ PCL-M 0.30 0.06 [�0.29, 0.89]COPE_Avoidant ¡ PCL-M��� 2.77 0.33 [1.42, 4.12]AAQ-II ¡ PCL-M�� 0.43 0.26 [0.14, 0.73]CDRISC ¡ PCL-M 0.09 0.09 [�0.05, 0.23]Social support ¡ BDI-II ��� �0.42 �0.29 [�0.68, �0.16]COPE_Action ¡ BDI-II �0.11 �0.03 [�0.59, 0.37]COPE_Avoidant ¡ BDI-II ��� 1.70 0.28 [0.72, 2.68]AAQ-II ¡ BDI-II �� 0.32 0.25 [0.09, 0.55]CDRISC ¡ BDI-II �0.02 �0.03 [�0.14, 0.10]

Personality subtype (ref: resilient) ¡ mentalhealth symptoms

OVER ¡ PCL-M 0.07 0.00 [�9.48, 9.61]UNDER ¡ PCL-M 1.30 0.03 [�5.00, 7.60]OVER ¡ BDI-II 2.57 0.07 [�3.88, 9.02]UNDER ¡ BDI-II 1.43 0.05 [�2.91, 5.77]

TBI diagnosis (ref: TBI negative) ¡ mentalhealth symptoms

TBI positive ¡ PCL-M� 5.77 0.15 [0.77, 10.77]TBI positive ¡ BDI-II 2.31 0.08 [�1.60, 6.22]

Note. Personality subtype, traumatic brain injury (TBI), Acceptance and Action Questionnaire–II (AAQ-II),and Connor–Davidson Resilience Scale (CDRISC) were measured at baseline. Social Support, Brief-COPE(B-COPE) Action, and B-COPE Avoid were measured at 4th month. PTSD Checklist–Military (PCL-M) andBeck Depression Inventory–II (BDI-II) were measured at 8th month. CI � confidence interval; ref � referencegroup; OVER � overcontrolled; UNDER � undercontrolled; Unst. � unstandardized; St. � standardized.� p � .05. �� p � .01. ��� p � .001.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

271VETERAN RESILIENCE

There is also prospective evidence that positive emotions may beelements of resilience that promote adaptive cognitive appraisalsof symptoms experienced by persons with chronic health condi-tions (Ong et al., 2010). From the perspective of Fredrickson’s(2013) broaden-and-build model positive emotion may be thedefining feature of trait resilience as it promotes flexibility inthinking and appraisals of stress, facilitates flexible coping andproblem solving, enhances social relationships and circumventslingering effects of stress and negative emotions (Ong et al., 2009).Future research should attend to the potentially unique benefits ofpositive emotionality in resilient personality prototypes to furtherour understanding of the interplay between positive affect, psy-chological flexibility, coping and supportive relationships. Theresults of the present study indicate that screening with personalityinstruments (e.g., the MPQ) to determine resilient, undercon-trolled, and overcontrolled traits may be used to assess therapeuticprognosis and to potentially allocate treatment resources to veter-ans who may be particularly at risk for chronic, unremittingproblems over time.

The mediating effect of avoidant coping, social support, andpsychological inflexibility provides some “explanation for themechanism that drives the relationship” (Hoyt, Imel, & Chan,2008, p. 323) of the personality prototypes with depression andPTSD over time. The use of avoiding coping behaviors, a lack offlexibility, and a problematic social support systems appear to

characterize individuals who have overcontrolled and undercon-trolled personality prototypes that, in turn, reinforce and perpetuatesymptoms associated with depression and PTSD. This pattern isconsistent with the broader literature regarding the relations be-tween PTSD, depression symptoms, coping (e.g., Holahan, Moos,Holahan, Brennan, & Schutte, 2005), and psychological inflexi-bility (Kumpula et al., 2011; Meyer et al., 2013). It is important tonote that these are issues that can be addressed in strategic inter-ventions.

Avoidant coping strategies can be modified through evidence-based treatments such as prolonged exposure and cognitive pro-cessing therapy (Liverant, Suvak, Pineles, & Resick, 2012; Rauchet al., 2009). Similarly, prior research indicates that psychosocialtreatment targeting psychological inflexibility is effective in treat-ing depression in veterans (Walser, Karlin, Trockel, Mazina, &Barr Taylor, 2013) and that increases in psychological flexibilitymediate response to treatment for anxiety and depression (Berking,Neacsiu, Comtois, & Linehan, 2009; Hayes, Orsillo, & Roemer,2010). Future research is needed to address whether personalityprototypes influence the effectiveness of these treatments in re-ducing symptoms of PTSD and depression.

Social support is another modifiable mediator between person-ality subtypes and mental health symptoms. Within the veteranmental health literature, low social support is a consistent predictorof greater symptoms, suicidal ideation, and poor functional out-

Table 4Indirect Effects of Personality Subtypes on Distress Outcomes in Corrected Path Model

Effect Unst. St.Unst. 95% C.I.

(Bootstrap)

OVER ¡ Social Support ¡ PCL-M�� 5.59 0.12 [0.45, 10.74]OVER ¡ B-COPE_Action ¡ PCL-M �0.15 �0.00 [�1.12, 0.81]OVER ¡ B-COPE_Avoidant ¡ PCL-M�� 6.86 0.15 [1.65, 12.07]OVER ¡ AAQ-II ¡ PCL-M �� 4.88 0.10 [0.95, 8.81]OVER ¡ CDRISC ¡ PCL-M �1.56 �0.03 [�4.30, 1.17]UNDER ¡ Social Support ¡ PCL-M�� 5.45 0.14 [1.01, 9.89]UNDER ¡ B-COPE_Action ¡ PCL-M �0.72 �0.02 [�2.33, 0.89]UNDER ¡ B-COPE_Avoidant ¡ PCL-M� 3.80 0.10 [0.89, 6.70]UNDER ¡ AAQ-II ¡ PCL-M�� 4.39 0.11 [0.94, 7.84]UNDER ¡ CDRISC ¡ PCL-M �1.65 �0.04 [�4.43, 1.13]OVER ¡ Social Support¡ BDI-II � 3.44 0.10 [0.55, 6.88]OVER ¡ B-COPE_Action ¡ BDI-II 0.06 0.00 [�0.56, 0.68]OVER ¡ B-COPE_Avoidant ¡ BDI-II �� 4.21 0.12 [1.05, 7.37]OVER ¡ AAQ-II ¡ BDI-II � 3.57 0.10 [0.73, 6.41]OVER ¡ CDRISC ¡ BDI-II 0.40 0.01 [�1.96, 2.76]UNDER ¡ Social Support ¡ BDI-II �� 3.35 0.11 [0.43, 6.27]UNDER ¡ B-COPE_Action ¡ BDI-II 0.27 0.01 [�0.95, 1.49]UNDER ¡ B-COPE_Avoidant ¡ BDI-II � 2.33 0.08 [0.44, 4.22]UNDER ¡ AAQ-II ¡ BDI-II �� 3.21 0.11 [0.62, 5.80]UNDER ¡ CDRISC ¡ BDI-II 0.43 0.02 [�1.92, 2.77]TBI positive ¡ Social Support ¡ PCL-M 1.11 0.03 [�1.57, 3.79]TBI positive ¡ AAQ-II ¡ PCL-M 0.54 0.01 [�0.94, 2.03]TBI positive ¡ Social Support ¡ BDI-II 0.68 0.02 [�1.07, 2.44]TBI positive ¡ AAQ-II ¡ BDI-II 0.40 0.01 [�0.75, 1.54]

Note. The reference group for the personality subtypes was the resilience group. Personality subtype, Accep-tance and Action Questionnaire–II (AAQ-II), and Connor–Davidson Resilience Scale (CDRISC) were measuredat baseline. Social Support, Brief-COPE (B-COPE) Action, and B-COPE Avoid were measured at 4th month.PCLM and BDI-II were measured at 8th month. CI � confidence interval; OVER � overcontrolled;UNDER � undercontrolled; B-COPE � Brief COPE scale; CDRISC � Connor–Davidson Resilience Scale;PCL-M � PTSD Checklist–Military; BDI-II � Beck Depression Inventory–II; TBI � traumatic brain injury;Unst. � unstandardized; St. � standardized.� p � .05. �� p � .01. ��� p � .001.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

272 ELLIOTT ET AL.

comes (DeBeer, Kimbrel, Meyer, Gulliver, & Morissette, 2014;Pietrzak, Russo, Ling, & Southwick, 2011; Tsai, Harpaz-Rotem,Pietrzak, & Southwick, 2012). Although social support is modifi-able, there are few interventions for veterans that directly addressimproving social support within the larger context of treatmentsaimed to reduce mental health symptoms. As part of a recovery-based approach, over the past several years the Veterans HealthAdministration has provided nationwide access to integrated peersupport programs within mental health services for veterans withsevere mental illness. Within civilian populations, peer supportprograms have demonstrated benefits such as greater treatmentengagement, better quality of life, increased social functioning,greater life satisfaction, fewer inpatient visits, and greater commu-nity engagement (Clarke et al., 2000; Craig, Doherty, Jamieson-Craig, Boocock, & Attafua, 2004; Min, Whitecraft, Rothbard, &Salzer, 2007).

It is interesting that despite the direct association of TBI toPTSD, TBI status had a limited influence on all other self-reportvariables in the contextual model. The study did not assess re-peated concussions (that often occur with multiple exposures toblasts during deployment) and we did not account for pre- orpostdeployment TBI. It is possible that individuals who experi-enced several TBIs might display more distress (and fewer psy-chological resources) than those with a single episode. Reliance onthe self-reported incidence of TBI is another limitation of thestudy. Similarly, the lack of predictive effects for self-reportedresilience raises questions about the prospective value of theConnor–Davidson instrument. Although higher scores on theCDRISC were associated with other self-report measures in amanner consistent with clinical and theoretical expectations, themeasure did not predict later adjustment after controlling for otherpredictors. Apparently, the CDRISC shares considerable overlapwith other adaptive psychological characteristics assessed in thecurrent study. The resilient personality prototype was indirectlyobtained from the MPQ. It is possible that the fairly transparent,“face valid” nature of the CDRISC may limit its utility as measureof personal resilience, particularly in prospective designs.

Findings should be interpreted within the context of limitingfactors. First, the sample included veterans who were primarilyCaucasian, male, and enrolled in the CTVHCS; thus, the degree towhich these findings generalize to other veterans is unclear. Sec-ond, the sample size was relatively small, although retention rateswere very high across time (over 90% across 8 months), especiallyfor a primarily mental health sample. These limitations are furtheroffset by the use of a prospective design and widely used, validatedself-report measures.

The present study included no predeployment personality data.No measure of symptom exaggeration was used in this study;therefore, we do not know the degree to which our results mayhave been affected by an “overreporting” of symptoms. The dif-ferences in combat exposure between two of the personality pro-totypes raises some concern about the possible effect of combatexposure on responses to the MPQ. Although the possible “con-taminating” effects of combat exposure on the MPQ cannot bedismissed, we can assume these effects would be systematic,affecting responses to all of the self-report measures in the study.In addition, including combat exposure as a covariate in the modelserved to partial out the variance directly attributable to combatexposure in the mediator and outcome variables. In this context,

then, our results reflect theoretically and clinically compellingfindings. Finally, we restricted our interest in combat exposure tothe total score from the Full Combat Exposure Scale and, in theprocess, we did not isolate specific aspects of combat exposure suchas the number of times in combat for study, nor did we have infor-mation about the length or duration of deployment. The degree towhich these specific factors may have affected the results isunknown.

References

Aldenderfer, M. S., & Blashfield, R. K. (1984). Cluster analysis. BeverlyHills, CA: Sage.

Armed Forces Health Surveillance Center. (2011). DoD numbers fortraumatic brain injury [Data file]. Retrieved from http://www.dvbic.org/TBI-Numbers.aspx

Bahraini, N. H., Breshears, R. E., Hernández, T. D., Schneider, A. L.,Forster, J. E., & Brenner, L. A. (2014). Traumatic brain injury andposttraumatic stress disorder. Psychiatric Clinics of North America, 37,55–75. http://dx.doi.org/10.1016/j.psc.2013.11.002

Beck, A. T., Steer, R. A., Ball, R., & Ranieri, W. (1996). Comparison ofBeck Depression Inventories -IA and -II in psychiatric outpatients.Journal of Personality Assessment, 67, 588 –597. http://dx.doi.org/10.1207/s15327752jpa6703_13

Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck DepressionInventory manual (2nd ed.). San Antonio, TX: Psychological Corpora-tion.

Belanger, H. G., Barwick, F. H., Kip, K. E., Kretzmer, T., & Vanderploeg,R. D. (2013). Postconcussive symptom complaints and potentially mal-leable positive predictors. The Clinical Neuropsychologist, 27, 343–355.http://dx.doi.org/10.1080/13854046.2013.774438

Belanger, H. G., Uomoto, J. M., & Vanderploeg, R. D. (2009). TheVeterans Health Administration’s (VHA’s) Polytrauma System of Carefor mild traumatic brain injury: Costs, benefits, and controversies. TheJournal of Head Trauma Rehabilitation, 24, 4–13. http://dx.doi.org/10.1097/HTR.0b013e3181957032

Berking, M., Neacsiu, A., Comtois, K. A., & Linehan, M. M. (2009). Theimpact of experiential avoidance on the reduction of depression intreatment for borderline personality disorder. Behaviour Research andTherapy, 47, 663–670. http://dx.doi.org/10.1016/j.brat.2009.04.011

Berry, J. W., Elliott, T. R., & Rivera, P. (2007). Resilient, undercontrolled,and overcontrolled personality prototypes among persons with spinalcord injury. Journal of Personality Assessment, 89, 292–302. http://dx.doi.org/10.1080/00223890701629813

Blanchard, E. B., Jones-Alexander, J., Buckley, T. C., & Forneris, C. A.(1996). Psychometric properties of the PTSD Checklist (PCL). Behav-iour Research and Therapy, 34, 669–673. http://dx.doi.org/10.1016/0005-7967(96)00033-2

Block, J. (1993). Studying personality the long way. In D. C. Funder, R. D.Parke, C. Tomlinson-Keasey, & K. Widaman (Eds.), Studying livesthrough time (pp. 9–41). Washington, DC: American PsychologicalAssociation. http://dx.doi.org/10.1037/10127-018

Block, J. H., & Block, J. (1980). The role of ego control and ego resiliencyin the organization of behavior. In W. A. Collins (Ed.), The MinnesotaSymposium on Child Psychology: Vol. 13. Development of cognition,affect, and social relations (pp. 39–101). Hillsdale, NJ: Erlbaum.

Bonanno, G. A. (2005). Resilience in the face of potential trauma. CurrentDirections in Psychological Science, 14, 135–138. http://dx.doi.org/10.1111/j.0963-7214.2005.00347.x

Bonanno, G. A., Mancini, A. D., Horton, J. L., Powell, T. M., Leardmann,C. A., Boyko, E. J., . . . Smith, T. C. (2012). Trajectories of traumasymptoms and resilience in deployed U.S. military service members:Prospective cohort study. The British Journal of Psychiatry, 200, 317–323. http://dx.doi.org/10.1192/bjp.bp.111.096552

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

273VETERAN RESILIENCE

Bond, F. W., Hayes, S. C., Baer, R. A., Carpenter, K. M., Guenole, N.,Orcutt, H. K., . . . Zettle, R. D. (2011). Preliminary psychometricproperties of the Acceptance and Action Questionnaire–II: A revisedmeasure of psychological inflexibility and experiential avoidance. Be-havior Therapy, 42, 676–688. http://dx.doi.org/10.1016/j.beth.2011.03.007

Brenner, L. A., Ivins, B. J., Schwab, K., Warden, D., Nelson, L. A., Jaffee,M., & Terrio, H. (2010). Traumatic brain injury, posttraumatic stressdisorder, and postconcussive symptom reporting among troops returningfrom Iraq. The Journal of Head Trauma Rehabilitation, 25, 307–312.http://dx.doi.org/10.1097/HTR.0b013e3181cada03

Browne, M. W., & Cudeck, R. (1992). Alternative ways of assessing modelfit. Sociological Methods & Research, 21, 230–258. http://dx.doi.org/10.1177/0049124192021002005

Carver, C. S. (1997). You want to measure coping but your protocol’s toolong: Consider the Brief COPE. International Journal of BehavioralMedicine, 4, 92–100. http://dx.doi.org/10.1207/s15327558ijbm0401_6

Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing copingstrategies: A theoretically based approach. Journal of Personality andSocial Psychology, 56, 267–283. http://dx.doi.org/10.1037/0022-3514.56.2.267

Caspi, A., & Silva, P. A. (1995). Temperamental qualities at age threepredict personality traits in young adulthood: Longitudinal evidencefrom a birth cohort. Child Development, 66, 486–498. http://dx.doi.org/10.2307/1131592

Chapman, B. P., & Goldberg, L. R. (2011). Replicability and 40-yearpredictive power of childhood ARC types. Journal of Personality andSocial Psychology, 101, 593–606. http://dx.doi.org/10.1037/a0024289

Clarke, G. N., Herinckx, H. A., Kinney, R. F., Paulson, R. I., Cutler, D. L.,Lewis, K., & Oxman, E. (2000). Psychiatric hospitalizations, arrests,emergency room visits, and homelessness of clients with serious andpersistent mental illness: Findings from a randomized trial of two ACTprograms vs. usual care. Mental Health Services Research, 2, 155–164.http://dx.doi.org/10.1023/A:1010141826867

Cohn, M. A., Fredrickson, B. L., Brown, S. L., Mikels, J. A., & Conway,A. M. (2009). Happiness unpacked: Positive emotions increase lifesatisfaction by building resilience. Emotion, 9, 361–368. http://dx.doi.org/10.1037/a0015952

Connor, K. M., & Davidson, J. R. T. (2003). Development of a newresilience scale: The Connor–Davidson Resilience Scale (CD-RISC).Depression and Anxiety, 18, 76–82. http://dx.doi.org/10.1002/da.10113

Craig, T., Doherty, I., Jamieson-Craig, R., Boocock, A., & Attafua, G. (2004).The consumer-employee as a member of a Mental Health Assertive Out-reach Team: I. Clinical and social outcomes. Journal of Mental Health, 13,59–69. http://dx.doi.org/10.1080/09638230410001654567

Davydov, D. M., Stewart, R., Ritchie, K., & Chaudieu, I. (2010). Resil-ience and mental health. Clinical Psychology Review, 30, 479–495.http://dx.doi.org/10.1016/j.cpr.2010.03.003

DeBeer, B. B., Kimbrel, N. A., Meyer, E. C., Gulliver, S. B., & Morissette,S. B. (2014). Combined PTSD and depressive symptoms interact withpost-deployment social support to predict suicidal ideation in OperationEnduring Freedom and Operation Iraqi Freedom veterans. PsychiatryResearch, 216, 357–362. http://dx.doi.org/10.1016/j.psychres.2014.02.010

Dennissen, J. J., Asendorpf, J. B., & van Aken, M. A. G. (2008). Childhoodpersonality predicts long-term trajectories of shyness and aggressivenessin the context of demographic transitions in emerging adulthood. Jour-nal of Personality, 76, 67–100. http://dx.doi.org/10.1111/j.1467-6494.2007.00480.x

Dolan, S., Martindale, S., Robinson, J., Kimbrel, N. A., Meyer, E. A.,Kruse, M. I., . . . Gulliver, S. B. (2012). Neuropsychological sequelae ofPTSD and TBI following war deployment among OEF/OIF veterans.Neuropsychological Review, 22, 21–34.

Fletcher, D., & Sarkar, M. (2013). Psychological resilience: A review andcritique of definitions, concepts, and theory. European Psychologist, 18,12–23. http://dx.doi.org/10.1027/1016-9040/a000124

Forbes, D., Creamer, M., & Biddle, D. (2001). The validity of the PTSDchecklist as a measure of symptomatic change in combat-related PTSD.Behaviour Research and Therapy, 39, 977–986. http://dx.doi.org/10.1016/S0005-7967(00)00084-X

Fredrickson, B. L. (2013). Positive emotions broaden and build. Advancesin Experimental Social Psychology, 47, 1–53.

Fredrickson, B. L., Cohn, M. A., Coffey, K. A., Pek, J., & Finkel, S. M.(2008). Open hearts build lives: Positive emotions, induced throughloving-kindness meditation, build consequential personal resources.Journal of Personality and Social Psychology, 95, 1045–1062. http://dx.doi.org/10.1037/a0013262

Fredrickson, B. L., Tugade, M. M., Waugh, C. E., & Larkin, G. R. (2003).What good are positive emotions in crises? A prospective study ofresilience and emotions following the terrorist attacks on the UnitedStates on September 11th, 2001. Journal of Personality and SocialPsychology, 84, 365–376. http://dx.doi.org/10.1037/0022-3514.84.2.365

Fulton, J. J., Calhoun, P. S., Wagner, H. R., Schry, A. R., Hair, L. P.,Feeling, N., . . . Beckham, J. C. (2015). The prevalence of posttraumaticstress disorder in Operation Enduring Freedom/Operation Iraqi Freedom(OEF/OIF) veterans: A meta-analysis. Journal of Anxiety Disorders, 31,98–107. http://dx.doi.org/10.1016/j.janxdis.2015.02.003

Grosso, J. A., Kimbrel, N. A., Dolan, S., Meyer, E. C., Kruse, M. I.,Gulliver, S. B., & Morissette, S. B. (2014). A test of whether copingstyles moderate the effect of PTSD symptoms on alcohol outcomes.Journal of Traumatic Stress, 27, 478–482. http://dx.doi.org/10.1002/jts.21943

Hayes, S. A., Orsillo, S. M., & Roemer, L. (2010). Changes in proposedmechanisms of action during an acceptance-based behavior therapy forgeneralized anxiety disorder. Behaviour Research and Therapy, 48,238–245. http://dx.doi.org/10.1016/j.brat.2009.11.006

Hayes, S. C., Strosahl, K., Wilson, K. G., Bissett, R. T., Pistorello, J.,Toarmino, D., . . . McCurry, S. M. (2004). Measuring experientialavoidance: A preliminary test of a working model. The PsychologicalRecord, 54, 553–578.

Hoge, C. W., Castro, C. A., Messer, S. C., McGurk, D., Cotting, D. I., &Koffman, R. L. (2004). Combat duty in Iraq and Afghanistan, mentalhealth problems, and barriers to care. The New England Journal ofMedicine, 351, 13–22. http://dx.doi.org/10.1056/NEJMoa040603

Hoge, C. W., McGurk, D., Thomas, J. L., Cox, A. L., Engel, C. C., &Castro, C. A. (2008). Mild traumatic brain injury in U.S. Soldiersreturning from Iraq. The New England Journal of Medicine, 358, 453–463. http://dx.doi.org/10.1056/NEJMoa072972

Holahan, C. J., Moos, R. H., Holahan, C. K., Brennan, P. L., & Schutte,K. K. (2005). Stress generation, avoidance coping, and depressive symp-toms: A 10-year model. Journal of Consulting and Clinical Psychology,73, 658–666. http://dx.doi.org/10.1037/0022-006X.73.4.658

Hoyt, W. T., Imel, Z. E., & Chan, F. (2008). Multiple regression andcorrelation techniques: Recent controversies and best practices. Reha-bilitation Psychology, 53, 321–339. http://dx.doi.org/10.1037/a0013021

Hu, L.-T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes incovariance structure analysis: Conventional criteria versus new alterna-tives. Structural Equation Modeling, 6, 1–55. http://dx.doi.org/10.1080/10705519909540118

Keen, S. M., Kutter, C. J., Niles, B. L., & Krinsley, K. E. (2008).Psychometric properties of PTSD Checklist in sample of male veterans.Journal of Rehabilitation Research and Development, 45, 465–474.http://dx.doi.org/10.1682/JRRD.2007.09.0138

King, L. A., King, D. W., Vogt, D. S., Knight, J., & Samper, R. E. (2006).Deployment Risk and Resilience Inventory: A collection of measures forstudying deployment-related experiences of military personnel and vet-

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

274 ELLIOTT ET AL.

erans. Military Psychology, 18, 89 –120. http://dx.doi.org/10.1207/s15327876mp1802_1

Kok, B. E., Coffey, K. A., Cohn, M. A., Catalino, L. I., Vacharkulksemsuk,T., Algoe, S. B., . . . Fredrickson, B. L. (2013). How positive emotionsbuild physical health: Perceived positive social connections account forthe upward spiral between positive emotions and vagal tone. Psycho-logical Science, 24, 1123–1132. http://dx.doi.org/10.1177/0956797612470827

Kok, B. E., & Fredrickson, B. L. (2010). Upward spirals of the heart:Autonomic flexibility, as indexed by vagal tone, reciprocally and pro-spectively predicts positive emotions and social connectedness. Biolog-ical Psychology, 85, 432–436. http://dx.doi.org/10.1016/j.biopsycho.2010.09.005

Kumpula, M. J., Orcutt, H. K., Bardeen, J. R., & Varkovitzky, R. L. (2011).Peritraumatic dissociation and experiential avoidance as prospectivepredictors of posttraumatic stress symptoms. Journal of Abnormal Psy-chology, 120, 617–627. http://dx.doi.org/10.1037/a0023927

Landis, J. R., & Koch, G. G. (1977). The measurement of observeragreement for categorical data. Biometrics, 33, 159–174. http://dx.doi.org/10.2307/2529310

Letzring, T. D., Block, J., & Funder, D. C. (2005). Ego control and egoresiliency: Generalization of self-report scales based on personalitydescriptions from acquaintances, clinicians, and the self. Journal ofResearch in Personality, 39, 395–422. http://dx.doi.org/10.1016/j.jrp.2004.06.003

Liverant, G. I., Suvak, M. K., Pineles, S. L., & Resick, P. A. (2012).Changes in posttraumatic stress disorder and depressive symptoms dur-ing cognitive processing therapy: Evidence for concurrent change. Jour-nal of Consulting and Clinical Psychology, 80, 957–967. http://dx.doi.org/10.1037/a0030485

Mancini, A. D., & Bonanno, G. A. (2010). Resilience to potential trauma:Toward a life span approach. In J. W. Reich, A. J. Zautra, & J. Hall(Eds.), Handbook of adult resilience: Concepts, methods, and applica-tions (pp. 258–282). New York, NY: Guilford Press.

McCauley, S. R., Wilde, E. A., Miller, E. R., Frisby, M. L., Garza, H. M.,Varghese, R., . . . McCarthy, J. J. (2013). Preinjury resilience and moodas predictors of early outcome following mild traumatic brain injury.Journal of Neurotrauma, 30, 642–652. http://dx.doi.org/10.1089/neu.2012.2393

Meyer, E. C., Morissette, S. B., Kimbrel, N. A., Kruse, M. I., & Gulliver,S. B. (2013). Acceptance and Action Questionnaire–II scores as apredictor of posttraumatic stress disorder symptoms among war veter-ans. Psychological Trauma: Theory, Research. Practice, and Policy, 5,521–528.

Miller, M. W., Greif, J. L., & Smith, A. A. (2003). MultidimensionalPersonality Questionnaire profiles of veterans with traumatic combatexposure: Externalizing and internalizing subtypes. Psychological As-sessment, 15, 205–215.

Miller, M. W., & Harrington, K. M. (2011). Personality factors in resil-ience to traumatic stress. In S. M. Southwick, B. T. Litz, D. Charney, &M. J. Friedman (Eds.), Resilience and mental health: Challenges acrossthe lifespan (pp. 56–75). New York, NY: Cambridge University Press.http://dx.doi.org/10.1017/CBO9780511994791.006

Min, S. Y., Whitecraft, J., Rothbard, A. B., & Salzer, M. S. (2007). Peersupport for persons with co-occurring disorders and community tenure:A survival analysis. Psychiatric Rehabilitation Journal, 30, 207–213.http://dx.doi.org/10.2975/30.3.2007.207.213

Morissette, S. B., Woodward, M., Kimbrel, N. A., Meyer, E. C., Kruse,M. I., Dolan, S., & Gulliver, S. B. (2011). Deployment-related TBI,persistent postconcussive symptoms, PTSD, and depression in OEF/OIFveterans. Rehabilitation Psychology, 56, 340–350. http://dx.doi.org/10.1037/a0025462

Muthén, L. K., & Muthén, B. O. (1998–2012). Mplus user’s guide (6thed.). Los Angeles, CA: Author.

Ong, A. D., Bergeman, C. S., & Boker, S. M. (2009). Resilience comes ofage: Defining features in later adulthood. Journal of Personality, 77,1777–1804. http://dx.doi.org/10.1111/j.1467-6494.2009.00600.x

Ong, A. D., Zautra, A. J., & Reid, M. C. (2010). Psychological resiliencepredicts decreases in pain catastrophizing through positive emotions.Psychology and Aging, 25, 516 –523. http://dx.doi.org/10.1037/a0019384

Patrick, C. J., Curtin, J. J., & Tellegen, A. (2002). Development andvalidation of a brief form of the Multidimensional Personality Question-naire. Psychological Assessment, 14, 150 –163. http://dx.doi.org/10.1037/1040-3590.14.2.150

Pietrzak, R. H., Russo, A. R., Ling, Q., & Southwick, S. M. (2011).Suicidal ideation in treatment-seeking veterans of Operations EnduringFreedom and Iraqi Freedom: The role of coping strategies, resilience,and social support. Journal of Psychiatric Research, 45, 720–726.http://dx.doi.org/10.1016/j.jpsychires.2010.11.015

Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resamplingstrategies for assessing and comparing indirect effects in multiple me-diator models. Behavior Research Methods, 40, 879–891. http://dx.doi.org/10.3758/BRM.40.3.879

Ramchand, R., Schell, T. L., Karney, B. R., Osilla, K. C., Burns, R. M., &Caldarone, L. B. (2010). Disparate prevalence estimates of PTSD amongservice members who served in Iraq and Afghanistan: Possible expla-nations. Journal of Traumatic Stress, 23, 59–68.

Rauch, S. A., Defever, E., Favorite, T., Duroe, A., Garrity, C., Martis, B.,& Liberzon, I. (2009). Prolonged exposure for PTSD in a VeteransHealth Administration PTSD clinic. Journal of Traumatic Stress, 22,60–64. http://dx.doi.org/10.1002/jts.20380

Schwab, K. A., Ivins, B., Cramer, G., Johnson, W., Sluss-Tiller, M., Kiley,K., . . . Warden, D. (2007). Screening for traumatic brain injury in troopsreturning from deployment in Afghanistan and Iraq: Initial investigationof the usefulness of a short screening tool for traumatic brain injury. TheJournal of Head Trauma Rehabilitation, 22, 377–389. http://dx.doi.org/10.1097/01.HTR.0000300233.98242.87

Seal, K. H., Bertenthal, D., Miner, C. R., Sen, S., & Marmar, C. (2007).Bringing the war back home: Mental health disorders among 103,788U.S. veterans returning from Iraq and Afghanistan seen at Department ofVeterans Affairs facilities. Archives of Internal Medicine, 167, 476–482. http://dx.doi.org/10.1001/archinte.167.5.476

Sharma, S. (1996). Applied multivariate techniques. New York, NY: Wi-ley.

Shiner, R. L., Masten, A. S., & Roberts, J. M. (2003). Childhood person-ality foreshadows adult personality and life outcomes two decades later.Journal of Personality, 71, 1145–1170. http://dx.doi.org/10.1111/1467-6494.7106010

Steca, P., Alessandri, G., & Caprara, G. V. (2010). The utility of awell-known personality typology in studying successful aging: Resil-ients, undercontrollers, and overcontrollers in old age. Personality andIndividual Differences, 48, 442–446. http://dx.doi.org/10.1016/j.paid.2009.11.016

Tellegen, A., & Waller, N. G. (2008). Exploring personality through testconstruction: Development of the multidimensional personality ques-tionnaire. In. G. J. Boyle, G. Matthews, & D. H. Saklofske (Eds.), TheSage handbook of personality theory and assessment: Vol. 2. Personalitymeasurement and testing (pp. 261–292). Thousand Oaks, CA: Sage.

Thomas, J. L., Wilk, J. E., Riviere, L. A., McGurk, D., Castro, C. A., &Hoge, C. W. (2010). Prevalence of mental health problems and func-tional impairment among active component and National Guard soldiers3 and 12 months following combat in Iraq. Archives of General Psychi-atry, 67, 614–623. http://dx.doi.org/10.1001/archgenpsychiatry.2010.54

Tsai, J., Harpaz-Rotem, I., Pietrzak, R. H., & Southwick, S. M. (2012). Therole of coping, resilience, and social support in mediating the relationbetween PTSD and social functioning in veterans returning from Iraq

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

275VETERAN RESILIENCE

and Afghanistan. Psychiatry: Interpersonal and Biological Processes,75, 135–149.

Vasterling, J. J., Bryant, R. A., & Keane, T. M. (2012). Understanding theinterface of traumatic stress and mild traumatic brain injury backgroundand conceptual framework. In J. J. Vasterling, R. A. Bryant, & T. M.Keane (Eds.), PTSD and mild traumatic brain injury (pp. 3–11). NewYork, NY: Guilford Press.

Vasterling, J. J., & Dikmen, S. (2012). Mild traumatic brain injury andposttraumatic stress disorder: Clinical and conceptual complexities.Journal of the International Neuropsychological Society, 18, 390–393.http://dx.doi.org/10.1017/S1355617712000367

Vogt, D. S., Proctor, S. P., King, D. W., King, L. A., & Vasterling, J. J.(2008). Validation of scales from the Deployment Risk and ResilienceInventory in a sample of Operation Iraqi Freedom veterans. Assessment,15, 391–403. http://dx.doi.org/10.1177/1073191108316030

Walser, R. D., Karlin, B. E., Trockel, M., Mazina, B., & Barr Taylor, C.(2013). Training in and implementation of Acceptance and CommitmentTherapy for depression in the Veterans Health Administration: Therapistand patient outcomes. Behaviour Research and Therapy, 51, 555–563.http://dx.doi.org/10.1016/j.brat.2013.05.009

Weathers, F. W., Litz, B. T., Herman, D. S., Huska, J. A., & Keane, T. M.(1993, October). The PTSD Checklist: Reliability, validity, and diagnos-

tic utility. Paper presented at the annual meeting of the InternationalSociety for Traumatic Stress Studies, San Antonio, TX.

Windle, G., & Bennett, K. M. (2012). Caring relationships: How topromote resilience in challenging times. In M. Unger (Ed.), The socialecology of resilience: A handbook of theory and practice (pp. 219–231).New York, NY: Springer Science � Business Media. http://dx.doi.org/10.1007/978–1-4614–0586-3_18

Windle, G., Bennett, K. M., & Noyes, J. (2011). A methodological reviewof resilience measurement scales. Health and Quality of Life Outcomes,9, 8. http://dx.doi.org/10.1186/1477-7525-9-8

Wolf, E. J., Miller, M. W., Harrington, K. M., & Reardon, A. (2012).Personality-based latent classes of posttraumatic psychopathology: Per-sonality disorders and the internalizing/externalizing model. Journal ofAbnormal Psychology, 121, 256 –262. http://dx.doi.org/10.1037/a0023237

Yu, C. (2002). Evaluating cutoff criteria of model fit indices for latentvariable models with binary and continuous outcome (Unpublisheddoctoral dissertation). University of California, Los Angeles.

Received January 13, 2015Revision received May 27, 2015

Accepted May 29, 2015 �

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

276 ELLIOTT ET AL.