PTSD as a Risk Factor for Chronic Gang Membership During Adolescence and Early Adulthood: A Group-Based Trajectory Modeling Approach

Abstract

This study sought to identify developmental patterns of gang membership among a sample of juvenile offenders during adolescence and early adulthood. Examine the relevance of PTSD for predicting development. The Pathways to Desistance data were used in analyses. Group-based trajectory modeling was used to identify developmental patterns of gang membership. Multinomial logistic regression examined if PTSD predicts trajectory group membership. A five-group model best fit the gang membership data. Meeting lifetime criteria for PTSD at baseline predicted membership to a trajectory group characterized by chronic gang membership. Results indicate that screening for PTSD may help to identify juvenile offenders who may be at-risk for becoming chronic gang members. Providing trauma-informed care may help to reduce the impact of PTSD on gang membership risk by providing treatment that helps juvenile offenders manage emotions and establish and maintain social support relationships outside of a gang context.

Keywords

gangs development PTSD

Introduction

Gangs present a major problem facing society today, as gang crime often results in collateral damage in the form of accidental victims and generally involves more serious offenses than non-gang crime (Klein et al., 2006). One important consideration in investigating issues related to gang crime is understanding gang membership from a developmental perspective. Generally, risk for gang membership is highest in adolescence (Esbensen & Carson, 2012; Pyrooz, 2014), however, there certainly exist gang members who demonstrate a more chronic course of continued gang involvement into adulthood. Risk for antisocial behavior decreases following adolescence (Laub et al., 1998; Moffitt, 1993), so these individuals who continue their involvement with gangs demonstrate continuity in serious criminal behavior and present a major concern for the criminal justice system. These things make identifying risk factors predicting chronic gang membership of critical importance for criminal justice professionals seeking to tackle issues related to gang crime in the United States.

Mental illness may help to understand chronic gang membership, as mental health problems are a robust risk factor for offending and mental illness exists at elevated prevalence among gang members (Ardino et al., 2013; Harris et al., 2013; Wojciechowski, 2017; Wood & Dennard, 2017). Post-traumatic stress disorder (PTSD) presents one type of mental illness that may be particularly salient for understanding chronic gang membership, as trauma exposure (or threat of exposure) is a necessary antecedent to the disorder and has also been found to be related to gang membership (American Psychiatric Association, 2013; Gilman et al., 2017; Miller, 2001; Peterson et al., 2004). Despite these potential links, there has yet to be any study which examines risk for chronic gang membership and the role of PTSD for understanding the developmental course of involvement with gangs.

Gang Membership from a Developmental Perspective

Criminology has experienced a paradigm shift in recent decades, progressing from a discipline focused on static influences of risk factors on offending during adolescence to a field that understands offending as a developmental process characterized by change and continuity across time (Blumstein, 1986; Laub, 2004). This move to a life-course approach to offending has resulted in the use of theories that account for shifts in risk/protective factor exposure leading to shifts in crime and differing salience of these factors for influencing offending based on period of the life-course. Moffitt (1993) presents one of the more prominent frameworks for studying antisocial behavior in this manner, delineating two patterns of antisocial behavior that evolve across time: adolescent-limited and life-course persistent. The former was posited to comprise the vast majority of youth whose path was characterized by later onset of delinquent behavior in adolescence and desistence as adulthood is entered. The latter was characterized by early onset of delinquent behavior and a more chronic course of engagement across the life-course. Risk and protective factor profiles early in life and social feedback processes associated with these early life differences were expected to differentiate the two groups. Obviously, the life-course persistent offenders presented the group of greatest concern to criminologists, as their chronic course also involved engagement in more serious offending behaviors. While more contemporary research indicates that heterogeneity in offending patterns across the life-course is much more complex than this (Cihan et al., 2017; Lum & Vovak, 2018; Wojciechowski, 2017), this framework provides a useful heuristic for understanding the relevance of a life-course perspective of offending in a general sense and also gang membership more specifically in the case of the present study. Gang membership may be a particularly salient antisocial behavior of interest as it pertains to the life-course persistent offender taxonomy because of the increased seriousness of offending that generally characterizes gangs (Klein et al., 2006).

Like all other criminological outcomes of interest, gang membership may be understood as a developmental process. These processes exist in spite of the widely-held myth of the permanence of gang membership (Howell, 2012; Krohn & Thornberry, 2008). Gang membership may be characterized by all of the life-course processes associated with general offending. Individuals may enter, leave, or change levels of involvement during different points in the life-course (Pyrooz, 2014; Pyrooz et al., 2013). As such, certain factors may be associated with risk for chronic gang membership. Historically there has existed some divergence regarding the gang “problem.” Empirical research has historically focused on adolescent gang membership assuming it as an age-graded process that individuals eventually age out of (Decker et al., 2013; Krohn & Thornberry, 2008). Alternatively, gang data maintained by law enforcement agencies has historically focused on adult gang members (National Gang Center, 2013). This divergence may be indicative of a bifurcated focus on adolescent-limited gang members and life-course persistent gang members. This may also indicate the relevance of studying this phenomenon as a continuous process from a life-course perspective. Doing so may help to understand why some individuals develop into chronic gang members who may potentially continue engagement in highly serious antisocial behavior well into adulthood.

In recent years, there has been a renewed focus on studying gang membership as a developmental process. Numerous qualitative studies examine the life-course processes involving gang members’ choices to join or leave a gang (Bolden, 2012; Carson & Esbensen, 2016; Jackson, 2016). Research by Pyrooz et al. (2013) indicated that the more embedded one was within a gang’s social structure, the more likely that an individual would demonstrate slower desistence from gang involvement. Melde and Esbensen (2011) indicate that joining a gang may act as a turning point in the life-course, resulting in changes to social controls and psychological constructs that partially account for an increase in delinquency. These changes were found to be relatively lasting, as desistence from gang membership did not result in such psychological, social, and behavioral changes. While all of these highlighted studies provide examples of the ways that gang research has been reoriented within a life-course framework in recent years, there has been little research studying the actual heterogeneity in patterns of gang membership during different periods of the life-course.

Pyrooz (2014) provides the only study that investigated these differences in patterns of gang membership. This research indicated that individuals differed in their timing of joining a gang, with some individuals reporting early initiation, whereas others reported joining a gang in adulthood. This study was thought to perhaps provide some explanation for the bifurcation of the focus on adolescent versus adult gang activity by data source. This study also provided further evidence for dispelling the myth that gang membership is some permanent status, as many individuals eventually desisted from gang involvement. The Pyrooz (2014) study also utilized a general population sample of adolescents and emerging adults to study this phenomenon. While these results then are generalizable, they may lack focus on the group of same-aged individuals who may be at greatest risk for gang involvement, that is, juvenile offenders.

Juvenile offenders have already demonstrated a proclivity for antisocial behavior, as evidenced by involvement with the juvenile justice system. Given the relationship between gang membership and antisocial behavior (Gordon et al., 2004; Harper et al., 2008; Hill et al., 2001), this group may be most at-risk. Focusing attention on developmental patterns of gang membership among this indicated group then may help target resources and treatment more efficiently in order to tackle issues related to gang activity. The present study sought to address this concern by examining developmental patterns of gang membership among a sample of juvenile offenders. While this study does indeed indicate the importance of studying differences in gang membership patterns, this study is mainly descriptive and focused only on the age-grading of membership, rather than attempting to explain why these individuals may become chronic gang members following onset. The present study sought to examine a potentially salient risk factor for chronic gang membership in PTSD.

PTSD as a Risk Factor for Chronic Gang Membership

Mental illness has been observed at elevated rates among gang members (Coid et al., 2013). Macfarlane (2019) presents several potential arguments for why this relationship may exist: (1) gangs preferentially select individuals with mental issues, (2) gang membership facilitates the development of mental health problems, and (3) gang membership enhances underlying mental health issues to the point that they become apparent. Some research has indicated support for the first postulate, as individuals suffering from mental illness may have greater risk of being selected into gangs. For example, in the case of attention-deficit hyperactivity disorder, Craig et al., (2002) found higher symptom scores increased risk for chronic gang membership and the symptoms preceded gang membership. Similar research has identified this mechanism underlying gang membership for antisocial personality disorder also (Egan & Beadman, 2011). However, there is a dearth of research focusing examination on PTSD as a predictor of gang membership and chronic gang membership. Using longitudinal data for understanding chronic patterns of gang membership may allow for identifying such pre-existing risk factors that predict chronic gang membership. If observed, this would provide evidence that individuals suffering from mental illness are selecting into gangs, rather than gang membership leading to onset or exacerbating mental health issues. The present study focuses on the first postulate, as there exist numerous reasons to believe that PTSD may increase risk for chronic gang membership.

Research has indicated that PTSD exists at elevated prevalence among gang members (Wood & Dennard, 2017). Victims of trauma often demonstrate deficits in establishing strong and healthy relationships (Bell et al., 2019; Fontana & Rosenheck, 2010; Saunders & Edelson, 1999; Wingo et al., 2017). Numerous underlying mechanisms have been posited to contribute to this increased risk. When individuals are exposed to interpersonal trauma, they have had their rights violated by another person. This may lead to a learned distrust of others that inhibits the establishment of healthy social relationships. Traumatic stress exposure has also been found to result in biological damage to the stress processing systems of the body (McEwen, 1998; Stam, 2007). This can lead to the uninhibited release of stress hormones, even when an individual is not presented with a provocative stimulus (McEwen, 2000). This can leave the individual in a constant state of fight or flight, leading to constant appraisal of the social environment scanning for potential threats or harm (Kleshchova et al., 2019; Lanius et al., 2017). Individuals may then lash out at otherwise neutral parties having perceived them to be threatening. This is evidenced in the elevated risk for engagement in violent behavior among individuals with PTSD (MacManus et al., 2015; Wojciechowski, 2017). It is easy to see how these biological and behavioral processes can hinder one’s ability to form strong and healthy social ties.

Beyond social relationships, trauma has also been found to be related to issues with identity formation and maintenance (American Psychiatric Association, 2013; Silverstein, 1994). Numerous explanations have been offered for this problem. Some focus on how experiences of trauma may lead to re-conceptualizations of one’s relationship with the world following near-death or simply harmful experiences that are more existential in nature (Thompson & Walsh, 2010; Weems et al., 2016). On a more extreme end, complex and recurring traumatic experiences can lead to dissociation experiences that can tear the identity into pieces in an attempt to preserve core characteristics of the person (Ringrose, 2018; Sinason, 2002). Issues with establishing and maintaining identity have been found to be associated with mental distress (Hatano & Sugimura, 2017; Meca et al., 2019). All of these proposed mechanisms are likely contributors to PTSD-sufferers’ diminished capacities, but gangs may present a means of addressing these social support and identity needs that cannot be found elsewhere.

Gang membership may provide solutions to the highlighted problems experienced by individuals with PTSD. Gangs present an established web of social relationships with entrenched roles for members of the group (Flannery et al., 1998; Gordon, 2000; Ngo et al., 2017). The fact that a hierarchal structure with identifiable roles and identities for members is already established may make the transition into a gang identity much simpler than attempting to establish one’s own identity when those abilities are already diminished due to experiences with trauma. Further, gangs often act as sources of social support by members and a lack of such relationships have been cited by gang members who lack such relationships as a reason why they joined a gang (Quinn et al., 2017; Sinclair & Grekul, 2012). Relatedly, individuals who have been victims of trauma often fear subsequent victimization both as a learned response to trauma and because of dysfunction in the stress processing systems of the body that regulate systems meant to aid in ensuring that revictimization does not occur (Jovanovic et al., 2010; Stam, 2007). For this reason, individuals with PTSD often indicate that they lack feelings of secureness and safety (Fullerton et al., 2015; Grieger et al., 2004). Gangs are often cited by members as being protective and providing such a sense of safety (Ortiz, 2019; Skarbek, 2012; Sobel & Osoba, 2009). While this may be somewhat counterintuitive because of the increased risk for exposure to violence that gang members face (Gilman, et al., 2017; Miller, 2001; Peterson et al., 2004), it is still identified by gang members as a major reason why gangs present a functional option for individuals with PTSD. In terms of risk for chronic membership, these highlighted issues presented by PTSD sufferers may make transitioning away from gang involvement more difficult than for individuals without PTSD. A lack of social connections outside of the gang may make desistence more difficult, as the social support provided by the gang cannot be substituted. Further, the lack of safety reported by individuals suffering from PTSD should be as chronic as the disorder itself, so individuals with PTSD may not feel safe if they leave the gang, thus, contributing to risk for chronic gang membership. Despite all of these highlighted mechanisms that may increase risk for chronic gang membership among PTSD sufferers, there has yet to be any study examining these processes.

This study aims to better understand how PTSD is related to gang membership is related to gang membership from a developmental perspective and addresses two research questions: (1) How do trajectories of gang membership differ during adolescence and early adulthood? and (2) Do individuals with PTSD have increased risk for chronic gang membership throughout adolescence and early adulthood?

Methods

Data

This study utilized data from all 11 waves of the Pathways to Desistance dataset. These data consist of the responses of 1,354 juvenile offenders recently adjudicated for a serious offense prior to baseline. This panel study followed these juvenile offenders across 84 months following baseline measurements, resulting in 11 total data points for each participant. The study period occurred from 2000 to 2010, with recruitment occurring during the 2000 to 2003 span. Serious offenses which qualified participants for inclusion in the study included all felony offenses, as well as several misdemeanor weapons-related and sexual assault charges. Study sites in Maricopa County, AZ and Philadelphia, PA were the locations where participants were recruited for the study. For male drug offenders, a cap on the total proportion of these individuals included in the sample was capped at 15% in order to maintain heterogeneity in baseline characteristics. A peak level of attrition of 16.8% was reached at the final data point.

Data utilized in the present study were all collected via participant self-report. The Pathways research team provided each participant with a laptop computer during the interview sessions. A member of the research team would then administer verbal prompts to the participant and the participant would input responses into the laptop computer. It was believed that this would maximize confidentiality and would increase honesty in reporting. Interviews took place in locations that were convenient for participants, like participants’ homes, criminal justice facilities, and/or libraries and other public places.

Measures

Gang membership

The main dependent variable of the present study was gang membership. A binary measure was used to assess gang membership at each observation point. This delineated participants who reported being in a gang during a given observation period from those who were not in a gang (0 = No; 1 = Yes).

PTSD

PTSD status at baseline was the main independent variable in this study. The Composite International Diagnostic Inventory was utilized to assess the likely presence or absence of PTSD diagnosis at baseline (World Health Organization, 1990). This is a fully structured interview used to assess the presence of PTSD by asking about whether or not they experience a range of symptoms of the disorder. Using computerized algorithms that allow for skip patterns, the research team was able to ask participants about the presence of PTSD symptoms in participants’ lifetimes. The algorithms were then able to track patterns of symptom endorsement that indicated likely/unlikely diagnosis of PTSD. These skip patterns also allow for discernment that these symptoms are not due to medication, alcohol/drug use, and/or some physical injury. A binary indicator was utilized to delineate participants who met criteria for a lifetime diagnosis for PTSD at baseline from those who did not (0 = No; 1 = Yes).

Control variables

Several control variables were included in analyses in order to ensure that estimations were not biased by omitted variables. The first of these control variables was gender. Gender was included because past research has indicated that risk for gang membership may be higher for males (Hill et al., 1999). Gender was operationalized as a binary variable at baseline, delineating male and female participants (0 = Male; 1 = Female).

Another covariate included in modeling was race. Racial identification has been indicated by past research as an important characteristic for gang affiliation (Esbensen et al., 2008). Race was measured at baseline as a nominal variable. Four race categories were delineated: White, Black, Hispanic, and Other Race. Several dummy variables were computed, each delineating one of the race categories from all other participants (e.g., 1 = White; 0 = All other participants. The dummy variable corresponding to White participants was omitted in order to provide a reference group.

Another control variable included in analyses was socioeconomic status (SES). This is because past research has indicated the importance of this construct for understanding risk for gang membership (Hollis, 2019). SES was measured at baseline as an interval variable composed of a weighted score combining participants’ parents’ occupational prestige and educational attainment. If both parents could provided data for this measure, then a mean of both parents’ scores was computed so that each participant has a single SES score.

Participants’ degree of association with deviant peers at baseline was also included in analyses as a control variable because past research has highlighted the influence of antisocial peers as a risk factor for gang membership (Lenzi et al., 2015). Deviant peer association was measured using a series of ordinal items on a five-point scale assessing the general number of peers who attempt to influence participants to engage in seven different antisocial acts in the prior 6 months (1 = None of them; 5 = All of them). This measure was adapted from the measures used by the Rochester Youth Development Study (Thornberry et al., 1994). A mean was then computed from the individual items in order to provide a single deviant peer association score for each participant.¹

While the influence of deviant peers is highly relevant for predicting gang membership risk, participants’ resistance to this influence as a psychological construct must also be controlled for in analyses. Research indicates that peer pressure may play a large role in predicting choice to join a gang (Alleyne & Wood, 2014), so one’s resistance to this influence is paramount for understanding this process. This construct was measured at baseline using a set of two-part prompts. Participants were first presented with a vignette with two response options and told to indicate which option best described their own behavior (e.g., “some people go along with their friends just to keep their friends happy” and “Other people refuse to go along with what their friends want to do, even though they know it will make their friends unhappy”). They then were asked to rate the degree to which the statement was accurate (sort of true vs. very true). A score ranging from one to four was then computed, with higher scores indicating greater resistance to peer influence. A mean was computed from all of the individual scores so that participants has a single score at each observation point. The baseline measurement of this variable was utilized in analyses.

Another variable included as a control covariate in analyses was age at baseline. Research has indicated that adolescence may be the peak period of the life-course for gang membership (Esbensen & Carson, 2012), necessitating inclusion of age in analyses as a control variable. Age was measured in single digit intervals.

Exposure to violence was also included as a control variable, as research indicates that gang members are at greater risk for having histories of trauma exposure (Miller, 2001). Two forms of exposure to violence were included as indicators: direct victimization and witnessed violence. Both of these variables were binary, as they delineated participants who reported ever experiencing any of either type of exposure to violence in their lifetimes prior to baseline from participants who did not report such histories (0 = No; 1 = Yes).

The final control variable included in analyses was social support. This is because research indicates that gang members’ choices to join gangs often rest on the basis of social isolation and gangs may offer protection and social support for those who lack it (Ngo et al., 2017). Social support was measured at baseline as a count of the number of caring adults who participants reported provided social support in two or more domains. This was a measure of social support depth, as it eschewed relationships that were only characterized by a single social support domain.

Analytic strategy

The analyses for this study proceeded in two phases. The first phase of analyses utilized group-based trajectory modeling (GBTM) to model longitudinal gang membership response data to elucidate patterns of gang membership risk across adolescence and early adulthood. GBTM entails an iterative process of fixing a series of polynomial functions of varying numbers and order (linear, quadratic, or cubic) to a set of longitudinal response data in order to discern the best fit to the longitudinal outcome data. Participants are assigned a probability of membership to each group in the model based on their own individual trajectory. Nested model fit is determined via the use of Bayesian Information Criteria (BIC) statistics for each model. As the number and order of polynomial functions fit to the data varies, BIC statistics will indicate variation in overall model fit. Beyond BIC, Nagin (2005) identifies several other relevant characteristics that are necessary to consider when determining the model with the best fit to the data also. Posterior probabilities of assignment indicate the average probability of assignment for all participants that are actually assigned membership to each group. This probability should exceed .7, indicating a high average probability of assignment to a given group and no others in the model for participants assigned to that group. Relatedly, average odds of correct classification indicate the improvement in the odds of correctly assigning a participant to the group to which they were actually assigned based on posterior probabilities of assignment. This improvement is compared to random assignment and should exceed a score of five, indicating a 500% improvement in the odds of correct assignment. 95% confidence bands should also be tightly fit around each group in the model, indicating general confidence in estimation. Finally, researchers should also use discretion for choosing parsimonious models over models which provide marginal improvement in fit with additional groups, but the new groups provide no additional nuance to the model. Full-information maximum likelihood estimation was utilized to manage missing data. Based on the data distribution, a probability distribution for modeling data must be chosen, with choices of censored-normal, zero-inflated poisson, and logit available. Because the gang membership outcome variable was binary, the logit model was chosen to model data.

The second phase of analyses explored whether PTSD and control covariates predicted assignment to trajectory groups using multinomial logistic regression. This method entails the use of trajectory group assignment as a nominal variable, with one category omitted in order to provide a reference group for covariate effects to be assessed in comparison to. Coefficients are described as relative risk ratios (RRR). RRR indicates the impact that a one unit increase on an independent variable of interest has on the risk of being assigned to a given group, relative to the risk of being assigned to the omitted reference group. Listwise deletion was used to manage missing data.

Results

Table 1 provides descriptive statistics for variables included in analyses. The sample was comprised mainly of males (Male = 86.4%; Female = 13.6%). The racial breakdown of the sample was 41.4% Black, 33.5% Hispanic, 20.2% White, and 4.4% were categorized as Other Race. 6.5% of the sample met criteria for a likely lifetime PTSD diagnosis at baseline. The average age at baseline was 16.044 years old.

Table 1.

Baseline Descriptive Statistics.

	Mean/Proportion	Standard deviation	Minimum	Maximum
PTSD diagnosis (0 = No; 1 = Yes)	0.065	NA	0	1
Direct victimization (0 = No; 1 = Yes)	0.673	NA	0	1
Witnessed violence (0 = No; 1 = Yes)	0.931	NA	0	1
Social support	1.998	1.058	0	6
Gender (0 = Males, 1 = Females)	0.136	NA	0	1
Race (Reference = Whites)
Black	0.414	NA	0	1
Hispanic	0.335	NA	0	1
Other race	0.048	NA	0	1
SES	51.409	12.299	11	77
Deviant peer association	1.321	0.926	0	4
Resistance to peer influence	1.987	0.576	0.3	3
Negative affect	0.746	0.779	0	4
Age	16.044	1.143	14	19

Results from the GBTM analyses indicated that a five-group model provided the best fit to the data. The five-group model provided better fit to the data based on BIC than the two, three, and four group models. While the six-group model provided somewhat better fit based on BIC, the inclusion of a sixth group did not provide any additional nuance to the model, as this group was comprised of participants who otherwise would fit into another group characterized by complete abstinence from gang membership, save for several in the group who indicated gang membership only for an extremely brief time. Stata also provided the error message “Warning: variance matrix is nonsymmetric or highly singular,” upon estimating the six-group model; indicating issues with estimation. Another group in this model also had very wide 95% confidence bands, indicating very little confidence in the estimated trajectory. For these reasons, the five-group model was chosen as the best fitting. The chosen model also met criteria for fit related to posterior probabilities of assignment, average odds of correct classification, and 95% confidence intervals. Table 2 provides BIC statistics for trajectory models, whereas Table 3 provides posterior probabilities of assignment for the trajectory groups in the chosen model.

Table 2.

Bayesian Information Criterion Statistics for Trajectory Models with Varying Group Numbers.

2-group	−2632.08
3-group	−2498.27
4-group	−2444.49
5-group	−2434.63
6-group	−2356.18

Table 3.

Posterior Probabilities of Assignment to Trajectory Groups.

	Accelerating	Adolescence-limited	Abstaining	Late desisting	Chronic
Accelerating	0.917	0.046	0.017	0.018	0.002
Adolescence-limited	0.024	0.838	0.094	0.040	0.005
Abstaining	0.004	0.015	0.980	<0.001	<0.001
Late desisting	0.021	0.043	0.001	0.922	0.013
Chronic	0.048	<0.001	<0.001	0.036	0.915

The chosen trajectory model is described in Figure 1. The first group in the model was characterized by slight accelerations in gang membership across the entire study period. 4.0% of the sample was assigned to this group with a linear growth function as is described as the “Accelerating” trajectory group. The next group in the model was characterized by a relatively high intercept at age 16, followed by reductions in gang membership across the study period, with no remaining participants reporting gang membership by age 20 and none thereafter. Because of this growth pattern, this group is described as the “Adolescence-Limited” group. This group was characterized by a quadratic polynomial function and 8.5% of the sample was assigned membership to this group. The third group in the model was characterized by complete abstinence from gang membership across the entire study period. This group is described as the “Abstaining” group. The vast majority of participants in the sample were assigned to the group (79.6%). This group was characterized by a linear polynomial function. The next group in the model was characterized by an age-16 intercept near 1, indicating nearly 100% gang membership at that time. This was followed by steady reductions in gang membership across the study period, with near complete desistence by age 23. This group is described as the “Late Desisting” group. This group was characterized by a quadratic polynomial function and 4.9% of the sample was assigned membership to this group. The final group in the model was characterized by steady acceleration in gang membership from age 16 to age 20, with 100% of the group reporting gang membership from age 20 through the remainder of the study period. This group is described as the “Chronic” group. This group comprised 3.0% of the sample and was characterized by a quadratic polynomial function.

Figure 1.

Gang membership trajectory model.

Beyond the shapes of the groups themselves, it is relevant to discuss their paths in relation to one another. Both the Late Desisting and Adolescent-Limited groups demonstrated turns toward desistance, but this occurs much later for the Late Desisting group and the change pattern is much shallower. This is evidenced by the fact that more than 80% of participants assigned to the Late Desisting group were still reporting gang membership around the time that participants assigned to the Adolescent-Limited group were reaching full within-group desistance. The proportion of reported gang membership for the Late Desisting and Chronic groups were nearly analogous around this same time, as the trajectory paths of the groups crossed. It is also relevant to note that the path of the Accelerating group crossed with that of the Adolescent-Limited group around age 18 with around 20% of both groups reporting gang membership, whereas this crossing occurs around age 21 for the Accelerating and Late Desisting groups with around 40% of each group reporting gang membership. While the acceleration in gang membership of the Accelerating group is consistent across the study period, it is very gradual. This is in stark contrast to the Chronic group which demonstrated early and sharp acceleration that plateaus around age 20 as the proportion of participants assigned to this group reporting gang membership reaches 100%.

The second phase of analyses entailed the use of multinomial logistic regression to examine the relevance of PTSD and covariates for predicting trajectory group assignment. Table 4 provides model estimates. The Abstaining trajectory group was omitted from analyses in order to serve as a reference group. All covariate effects should be interpreted as impacts on risk of assignment in comparison to risk of assignment to the Abstaining group. PTSD was found to significantly predict membership the Chronic trajectory group (RRR = 2.633). White participants demonstrated greater risk of assignment to the Acceleration group, relative to Black participants. Hispanic participants demonstrated greater risk of assignment to the Adolescence-Limited, Late Desisting, and Chronic groups, relative to White participants. Other Race participants demonstrated greater risk of assignment to the Chronic, and Late Desisting groups, relative to White participants. Greater levels of baseline deviant peer association were associated with increased risk of assignment to the Accelerating, Adolescence-Limited, and Late Desisting groups. Older age at baseline was associated with increased risk of assignment to the Adolescence-Limited group. A baseline history of experiencing direct victimization was associated with increased risk of assignment to the Chronic and Late Desisting groups.

Table 4.

Covariate Impact on Relative Risk of Assignment to Gang Membership Trajectory Groups in Relative Risk Ratios.

	Accelerating	Adolescence-limited	Abstaining	Late desisting	Chronic
PTSD diagnosis (0 = No; 1 = Yes)	1.789	0.531	— (omitted)	0.305	2.633*
Direct victimization (0 = No; 1 = Yes)	1.861	1.609	—	3.681*	2.545
Witnessed violence (0 = No; 1 = Yes)	1.272	1.383	—	0.404	1.578
Social support	1.190	0.819	—	0.922	0.915
Gender (0 = Males, 1 = Females)	0.168	0.733	—	0.245	0.160
Race (Reference = Whites)
Black	0.414*	0.655	—	6.697	4.054
Hispanic	0.498	3.500***	—	21.527**	16.754**
Other race	0.300	1.721	—	21.658**	11.649*
SES	1.009	1.013	—	1.014	1.014
Deviant peer association	2.070***	1.789***	—	2.629***	1.344
Resistance to peer influence	0.684	0.704	—	0.643	0.827
Negative affect	0.728	1.204	—	1.250	0.930
Age	0.944	1.283*	—	1.140	1.271
Constant	0.076	<0.001***	—	0.001**	<0.001

p ≤ .05. **p ≤ .01. ***p.

Discussion

Findings from this study indicate the relevance of PTSD for understanding chronic gang membership. Further, a five-group model of development was also identified, with some individuals demonstrating desistence from gang involvement, while others demonstrated a more chronic course throughout adolescence and early adulthood. Juvenile offenders appeared to present greater risk for gang membership at any point in the study period compared to past research using general population samples. Meeting criteria for lifetime PTSD diagnosis at baseline was associated with increased risk of being assigned to the Chronic gang membership trajectory group. This indicates that PTSD presents a major risk factor for being involved with a gang throughout adolescence and adulthood. Racial identification also emerged in the model as a relevant predictor of chronic gang membership. All of these findings have important implications for criminal justice and mental health professionals seeking to mitigate the impact of mental illness on gang membership.

Substantive and Theoretical Implications of Results

As noted above, a five-group model of development was found to best fit the gang membership data. Around 80% of the sample was assigned membership to the Abstaining group. Even accounting for the small bit of error in modeling, around 20% of juvenile offenders in this sample demonstrated some level of gang involvement during adolescence and early adulthood. This indicates, perhaps unsurprisingly, that this population of youth are at elevated risk for involvement with the gangs. While these youth did demonstrate some degree of gang involvement, the vast majority did indeed demonstrate desistence from gang involvement by the end of the study period, even if at varying ages. However, 3% of the sample followed a relatively chronic course of gang membership throughout the entire study period. By age 20, all juvenile offenders in this group reported gang membership and all remained in a gang for the remainder of the study period. This continued involvement with gangs into adulthood is a major concern, as these individuals are more likely to be continuing engagement in serious antisocial behavior consistent with life-course persistent offenders (Moffitt, 1993). In terms of the age-graded nature of this relationship, juvenile offenders reported elevated risk for gang membership in adolescence and adulthood. At any given point in the study, between 5% and 14% of participants reported gang membership.² With only the Pyrooz (2014) study for context because of methodology, this is elevated compared to general population adolescents.

Perhaps the most important risk factor identified for predicting assignment to the Chronic trajectory group was baseline PTSD status. Participants who met criteria for a lifetime PTSD diagnosis at baseline indicated greater risk for assignment to this trajectory group, with their risk for assignment to this group being 170% greater than those participants who did not meet these criteria at baseline, net of all controls. While this is somewhat not surprising considering that past research has indicated that mental illness exists at elevated prevalence among gang members (Wood & Dennard, 2017), temporal ordering is established indicating that PTSD is a risk factor for chronic gang membership, as opposed to developed due to exposure to violence while already a gang member. Considering that victims of trauma can struggle to establish and maintain healthy social relationships (Bell et al., 2019; Fontana & Rosenheck, 2010; Saunders & Edelson, 1999; Wingo et al., 2017), gangs may offer a means of obtaining social support (Quinn et al., 2017; Sinclair & Grekul, 2012) and may address these same issues among PTSD sufferers that allow for feelings of safety following trauma exposure.

Racial identification was also important for understanding gang membership. Participants identifying as Hispanic or Other Race demonstrated elevated risk for reporting chronic gang membership. Post-hoc analyses indicated that the vast majority of the active gang members included in this study were located in the Maricopa County, AZ study site. Considering that this area is a known hub for gang membership by members of Hispanic and Native American ethnic groups (Joseph & Taylor, 2003; Winfree et al., 2001), it is unsurprising that race would predict gang membership in this sample in this manner.³ In this way, the race effects observed must be nested within this context and interpreted as being mainly due to the community from which the data were gathered. While this context tempers the impact of this high-magnitude race finding to some extent, this also indicates the continued importance of providing culturally competent gang prevention programming in this community and understanding that youth with these ethnic backgrounds may be targeted by gangs in this area for recruitment.⁴

Specific Policy/Programmatic Implications of Results

There are numerous potential policy/programmatic measures that can be taken to address gang issues that may be drawn from the findings of this study. First, the Chronic group presents potentially the most problematic of the trajectory groups for the reasons highlighted previously. Because of this, the juvenile offenders following the Chronic developmental path should be prioritized by criminal justice professionals. Indicating gang membership in adulthood should then be seen as a potential red flag to criminal justice professionals, as they present greater risk for chronic involvement. Addressing the risk factors identified as predicting this chronic course should be considered a paramount concern for professionals focused on reducing risk for gang membership.

As noted above, social support provided by the gangs may be one reason why youth afflicted with PTSD may be drawn to chronic gang membership. Using a trauma-informed care (TIC) approach to treatment may present one way to address these issues and ameliorate the impact of PTSD on gang membership risk among juvenile offenders. The TIC treatment philosophy rests upon the perspective that re-traumatization should be avoided during treatment in order to mitigate risk of further impact on symptomology (Hodas, 2006). Further, the perspective rests on three pillars: safety, connections, and managing emotions (Bath, 2008). Providing treatment that aids in juvenile offenders with PTSD establishing healthy social relationships may reduce their risk for turning to gangs for social support. Considering that gang involvement is associated with more serious offending (Klein et al., 2006), this could also further the deceleration and desistence process. For these reasons the treatment of juvenile offenders diagnosed with PTSD should be a paramount concern for criminal justice professionals focused on reducing the impact of gangs on communities. However, having a history of direct victimization at baseline also was associated with assignment to this Chronic group, indicating that trauma exposure itself is also relevant for screening also, as TIC may have utility for impacting chronic gang membership risk among these individuals also even if they have not developed PTSD.

Limitations

While the present study provided a unique examination of gang membership patterns and identified PTSD as a predictor of gang membership, there are also numerous limitations of this study. The first limitation pertains to the GBTM method used to elucidate developmental patterns of gang membership. Cautious interpretation of these patterns is imperative, as these only represent rough approximations of gang membership across time, rather than concrete entities which participants follow in lock-step. Despite this limitation, this method remains one of the more powerful means of analyzing longitudinal data that is currently available. Another limitation of the present study pertains to the parsing of causality between PTSD and gang membership onset. While PTSD was found to predict risk for chronic gang membership, this does not necessarily mean that PTSD onset prior to onset of gang membership for some members of this trajectory group. While PTSD certainly preceded gang membership for many in this group, there were also likely individuals in this group already belonged to gangs at or prior to baseline and their PTSD may have initiated after joining a gang. Considering that gang members are at-risk for exposure to violence (Gilman, et al., 2017; Miller, 2001; Peterson et al., 2004), this possibility seems likely. This would be consistent with Macfarlane’s (2019) (2) and (3) postulates regarding the elevated prevalence of mental illness among gang members. For this reason, PTSD should be seen definitively as a reason for the continuity of gang membership, but only interpreted more cautiously as a risk factor for actually joining a gang in the first place.

Conclusion

Findings of this study were novel in several ways. This was the first study to examine developmental patterns of gang membership among juvenile offenders. At any given point in the study period, these juvenile offenders did seem to demonstrate risk that was elevated compared to general population peers, indicating the high-risk these individuals present. These results also indicate that PTSD status at baseline was a strong predictor of chronic gang membership into adulthood. In establishing this relationship, this study was the first to establish temporal ordering in the relationship between mental illness and gang membership. While past research has indicated that gang members report elevated prevalence of mental illness (Wood & Dennard, 2017), this key component of causality had yet to be established for this relationship. Perhaps the most problematic aspect of this finding is the chronic nature of the relationship. While some participants in the sample reported desistence from gang involvement across the study period, PTSD was only related to the developmental pattern characterized by chronic gang membership continuing into adulthood. For this reason, screening and treatment of adolescent gang members with PTSD entering the criminal justice system should be prioritized. Providing treatment addressing the mental health needs of these individuals and aiding in their capacity to establish healthy social relationships outside the gang may help in working toward desistence from gang activity. Future research should continue to investigate the mechanisms that underlie the PTSD-gang relationship in order to better design and implement programming that may work toward the goal of reducing gang activity.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Thomas Wojciechowski

Notes

Author Biography

Thomas Wojciechowski is an assistant professor with the Michigan State University School of Criminal Justice. His research mainly focuses on mental health, drug use, and human development.

References

Alleyne

Wood

J. L.

(2014). Gang involvement: Social and environmental factors. Crime & Delinquency, 60(4), 547–568.

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders: DSM-5. American Psychiatric Association.

Ardino

Milani

Di Blasio

(2013). PTSD and re-offending risk: The mediating role of worry and a negative perception of other people’s support. European Journal of Psychotraumatology, 4(1), 21382.

Bath

(2008). The three pillars of trauma-informed care. Reclaiming Children and Youth, 17(3), 17–21.

Bell

Robinson

Katona

Fett

A. K.

Shergill

(2019). When trust is lost: The impact of interpersonal trauma on social interactions. Psychological Medicine, 49(6), 1041–1046.

Blumstein

(ed.). (1986). Criminal careers and career criminals (Vol. 2). The National Academies Press.

Bolden

C. L.

(2012). Liquid soldiers: Fluidity and gang membership. Deviant Behavior, 33(3), 207–222.

Carson

D. C.

Esbensen

F. A.

(2016). Motivations for leaving gangs in the USA: A qualitative comparison of leaving processes across gang definitions. In Maxson

Esbensen

(Eds.), Gang transitions and transformations in an international context (pp. 139–155). Springer.

Cihan

Sorensen

Chism

K. A.

(2017). Analyzing the offending activity of inmates: Trajectories of offense seriousness, escalation, and de-escalation. Journal of Criminal Justice, 50, 12–18.

10.

Coid

J. W.

Ullrich

Keers

Bebbington

DeStavola

B. L.

Kallis

Yang

Reiss

Jenkins

Donnelly

(2013). Gang membership, violence, and psychiatric morbidity. American Journal of Psychiatry, 170(9), 985–993.

11.

Craig

W. M.

Vitaro

Gagnon

Tremblay

R. E.

(2002). The road to gang membership: Characteristics of male gang and nongang members from ages 10 to 14. Social Development, 11(1), 53–68.

12.

Decker

S. H.

Melde

Pyrooz

D. C.

(2013). What do we know about gangs and gang members and where do we go from here? Justice Quarterly, 30(3), 369–402.

13.

Egan

Beadman

(2011). Personality and gang embeddedness. Personality and Individual Differences, 51(6), 748–753.

14.

Esbensen

F. A.

Brick

B. T.

Melde

Tusinski

Taylor

T. J.

(2008). The role of race and ethnicity in gang membership. In Genert

F. V.

Peterson

Lien

(Eds.), Street gangs, migration and ethnicity (pp. 117–139). Willan.

15.

Esbensen

F. A.

Carson

D. C.

(2012). Who are the gangsters? An examination of the age, race/ethnicity, sex, and immigration status of self-reported gang members in a seven-city study of American youth. Journal of Contemporary Criminal Justice, 28(4), 465–481.

16.

Flannery

D. J.

Huff

C. R.

Manos

(1998). Youth gangs: A developmental perspective. In Gullotta

T. P.

Adams

G. R.

Montemayor

(Eds.), Delinquent violent youth: Theory and interventions (pp. 175–204). Sage.

17.

Fontana

Rosenheck

(2010). War zone veterans returning to treatment: Effects of social functioning and psychopathology. The Journal of Nervous and Mental Disease, 198(10), 699–707.

18.

Fullerton

C. S.

Mash

H. B. H.

Benevides

K. N.

Morganstein

J. C.

Ursano

R. J.

(2015). Distress of routine activities and perceived safety associated with post-traumatic stress, depression, and alcohol use: 2002 Washington, DC, sniper attacks. Disaster Medicine and Public Health Preparedness, 9(5), 509–515.

19.

Gilman

A. B.

Howell

J. C.

Hipwell

A. E.

Stepp

S. D.

(2017). The reciprocal relationship between gang involvement and victimization by peers: Findings from the Pittsburgh Girls Study. Journal of Developmental and Life-Course Criminology, 3(2), 151–167.

20.

Gordon

R. A.

Lahey

B. B.

Kawai

Loeber

Stouthamer-Loeber

Farrington

D. P.

(2004). Antisocial behavior and youth gang membership: Selection and socialization. Criminology, 42(1), 55–88.

21.

Gordon

R. M.

(2000). Criminal business organizations, street gangs and wanna-be groups: A Vancouver perspective. Canadian Journal of Criminology, 42(1), 39–60.

22.

Grieger

T. A.

Fullerton

C. S.

Ursano

R. J.

(2004). Posttraumatic stress disorder, depression, and perceived safety 13 months after September 11. Psychiatric Services, 55(9), 1061–1063.

23.

Harper

G. W.

Davidson

Hosek

S. G.

(2008). Influence of gang membership on negative affect, substance use, and antisocial behavior among homeless African American male youth. American Journal of Men’s Health, 2(3), 229–243.

24.

Harris

Elkins

Butler

Shelton

Robles

Kwok

Simpson

Young

D. W.

Mayhew

Brown

Sargent

(2013). Youth gang members: Psychiatric disorders and substance use. Laws, 2(4), 392–400.

25.

Hatano

Sugimura

(2017). Is adolescence a period of identity formation for all youth? Insights from a four-wave longitudinal study of identity dynamics in Japan. Developmental Psychology, 53(11), 2113–2126.

26.

Hill

K. G.

Howell

J. C.

Hawkins

J. D.

Battin-Pearson

S. R.

(1999). Childhood risk factors for adolescent gang membership: Results from the Seattle Social Development Project. Journal of Research in Crime and Delinquency, 36(3), 300–322.

27.

Hill

K. G.

Lui

Hawkins

J. D.

(2001). Early precursors of gang membership: A study of Seattle youth. US Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention.

28.

Hodas

G. R.

(2006). Responding to childhood trauma: The promise and practice of trauma informed care. Pennsylvania Office of Mental Health and Substance Abuse Services, 177.

29.

Hollis

M. E.

(2019). Disaggregating gang activity: An exploratory study of the socio-demographic context of gang activity. Crime, Law and Social Change, 71(4), 441–458.

30.

Howell

J. C.

Griffiths

(2012). Gangs in America’s communities. Thousand Oaks: Sage.

31.

Jackson

(2016). Gang risk factors among urban Jamaican youth: A qualitative analysis. International Journal of Criminal Justice Sciences, 11(2), 132.

32.

Joseph

Taylor

(2003). Native-American youths and gangs. Journal of Gang Research, 10(2), 45–54.

33.

Jovanovic

Norrholm

S. D.

Blanding

N. Q.

Davis

Duncan

Bradley

Ressler

K. J.

(2010). Impaired fear inhibition is a biomarker of PTSD but not depression. Depression and Anxiety, 27(3), 244–251.

34.

Klein

M. W.

Weerman

F. M.

Thornberry

T. P.

(2006). Street gang violence in Europe. European Journal of Criminology, 3(4), 413–437.

35.

Kleshchova

Rieder

J. K.

Grinband

Weierich

M. R.

(2019). Resting amygdala connectivity and basal sympathetic tone as markers of chronic hypervigilance. Psychoneuroendocrinology, 102, 68–78.

36.

Krohn

M. D.

Thornberry

T. P.

(2008). Longitudinal perspectives on adolescent street gangs. In Liberman

(Ed.), The long view of crime: A synthesis of longitudinal research (pp. 128–160). Springer.

37.

Lanius

R. A.

Rabellino

Boyd

J. E.

Harricharan

Frewen

P. A.

McKinnon

M. C.

(2017). The innate alarm system in PTSD: Conscious and subconscious processing of threat. Current Opinion in Psychology, 14, 109–115.

38.

Laub

J. H.

(2004). The life course of criminology in the United States: The American Society of Criminology 2003 presidential address. Criminology, 42(1), 1–26.

39.

Laub

J. H.

Nagin

D. S.

Sampson

R. J.

(1998). Trajectories of change in criminal offending: Good marriages and the desistance process. American Sociological Review,63, 225–238.

40.

Lenzi

Sharkey

Vieno

Mayworm

Dougherty

Nylund-Gibson

(2015). Adolescent gang involvement: The role of individual, family, peer, and school factors in a multilevel perspective. Aggressive Behavior, 41(4), 386–397.

41.

Lum

Vovak

(2018). Variability in the use of misdemeanor arrests by police agencies from 1990 to 2013: An application of group-based trajectory modeling. Criminal Justice Policy Review, 29(6–7), 536–560.

42.

Macfarlane

(2019). Gangs and adolescent mental health: A narrative review. Journal of Child & Adolescent Trauma, 12(3), 411–420.

43.

MacManus

Rona

Dickson

Somaini

Fear

Wessely

(2015). Aggressive and violent behavior among military personnel deployed to Iraq and Afghanistan: Prevalence and link with deployment and combat exposure. Epidemiologic Reviews, 37(1), 196–212.

44.

McEwen

B. S.

(1998). Stress, adaptation, and disease: Allostasis and allostatic load. Annals of the New York Academy of Sciences, 840(1), 33–44.

45.

McEwen

B. S.

(2000). Allostasis and allostatic load: Implications for neuropsychopharmacology. Neuropsychopharmacology, 22(2), 108–124.

46.

Meca

Rodil

J. C.

Paulson

J. F.

Kelley

Schwartz

S. J.

Unger

J. B.

Lorenzo-Blanco

E. I.

Des Rosiers

S. E.

Gonzalez-Backen

Baezconde-Garbanati

Zamboanga

B. L.

(2019). Examining the directionality between identity development and depressive symptoms among recently immigrated Hispanic adolescents. Journal of Youth and Adolescence, 48(11), 2114–2124.

47.

Melde

Esbensen

F. A.

(2011). Gang membership as a turning point in the life course. Criminology, 49(2), 513–552.

48.

Miller

(2001). One of the guys: Girls, gangs and gender. Oxford University Press.

49.

Moffitt

T. E.

(1993). Adolescence-limited and life-course-persistent antisocial behavior: a developmental taxonomy. Psychological Review, 100(4), 674–701.

50.

Nagin

D. S.

(2005). Group-based modeling of development. Harvard University Press.

51.

National Gang Center. (2013). National youth gang survey analysis. https://www.nationalgangcenter.gov

52.

Ngo

Calhoun

Worthington

Pyrch

Este

(2017). The unravelling of identities and belonging: Criminal gang involvement of youth from immigrant families. Journal of International Migration and Integration, 18(1), 63–84.

53.

Ortiz

J. M.

(2019). Gangs and environment: A comparative analysis of prison and street gangs. American Journal of Qualitative Research, 2(1), 97–117.

54.

Peterson

Taylor

T. J.

Esbensen

F. A.

(2004). Gang membership and violent victimization. Justice Quarterly, 21(4), 793–815.

55.

Pyrooz

D. C.

(2014). “From your first cigarette to your last dyin’day”: The patterning of gang membership in the life-course. Journal of Quantitative Criminology, 30(2), 349–372.

56.

Pyrooz

D. C.

Sweeten

Piquero

A. R.

(2013). Continuity and change in gang membership and gang embeddedness. Journal of Research in Crime and Delinquency, 50(2), 239–271.

57.

Quinn

Pacella

M. L.

Dickson-Gomez

Nydegger

L. A.

(2017). Childhood adversity and the continued exposure to trauma and violence among adolescent gang members. American Journal of Community Psychology, 59(1–2), 36–49.

58.

Ringrose

J. L.

(2018). Understanding and treating dissociative identity disorder (or multiple personality disorder). Routledge.

59.

Saunders

E. A.

Edelson

J. A.

(1999). Attachment style, traumatic bonding, and developing relational capacities in a long-term trauma group for women. International Journal of Group Psychotherapy, 49(4), 465–485.

60.

Silverstein

(1994). Chronic identity diffusion in traumatized combat veterans. Social Behavior and Personality: An International Journal, 22(1), 69–79.

61.

Sinason

(Ed.). (2002). Attachment, trauma and multiplicity: Working with dissociative identity disorder. Psychology Press.

62.

Sinclair

Grekul

(2012). Aboriginal youth gangs in Canada: (de) constructing an epidemic. First Peoples Child & Family Review, 7(1), 8–28.

63.

Skarbek

(2012). Prison gangs, norms, and organizations. Journal of Economic Behavior & Organization, 82(1), 96–109.

64.

Sobel

R. S.

Osoba

B. J.

(2009). Youth gangs as pseudo-governments: Implications for violent crime. Southern Economic Journal, 75, 996–1018.

65.

Stam

(2007). PTSD and stress sensitisation: A tale of brain and body: Part 1: Human studies. Neuroscience & Biobehavioral Reviews, 31(4), 530–557.

66.

Thompson

Walsh

(2010). The existential basis of trauma. Journal of Social Work Practice, 24(4), 377–389.

67.

Thornberry

T. P.

Lizotte

A. J.

Krohn

M. D.

Farnworth

Jang

S. J.

(1994). Delinquent peers, beliefs, and delinquent behavior: A longitudinal test of interactional theory. Criminology, 32, 47–83.

68.

Weems

C. F.

Russell

J. D.

Neill

E. L.

Berman

S. L.

Scott

B. G.

(2016). Existential anxiety among adolescents exposed to disaster: Linkages among level of exposure, PTSD, and depression symptoms. Journal of Traumatic Stress, 29(5), 466–473.

69.

Winfree

L. T.

Jr. Bernat

F. P.

Esbensen

F. A.

(2001). Hispanic and Anglo gang membership in two southwestern cities. The Social Science Journal, 38(1), 105–117.

70.

Wingo

A. P.

Briscione

Norrholm

S. D.

Jovanovic

McCullough

S. A.

Skelton

Bradley

(2017). Psychological resilience is associated with more intact social functioning in veterans with post-traumatic stress disorder and depression. Psychiatry Research, 249, 206–211.

71.

Wojciechowski

T. W.

(2017). PTSD as a risk factor for the development of violence among juvenile offenders: A group-based trajectory modeling approach. Journal of Interpersonal Violence, 35(13–14), 2511–2535.

72.

Wood

Dennard

(2017). Gang membership: Links to violence exposure, paranoia, PTSD, anxiety, and forced control of behavior in prison. Psychiatry, 80(1), 30–41.

73.

World Health Organization. (1990). Composite international diagnostic interview. World Health Organization.