Homelessness and Contact with the Criminal Legal System among U.S. Combat Veterans: An Exploration of Potential Mediating Factors

Abstract

A notable proportion of criminal legal system (CLS)-involved combat veterans have a history of homelessness. However, knowledge regarding the relationship between homelessness and CLS involvement among combat veterans is largely based on descriptive studies, and potential mechanisms linking homelessness to CLS involvement are largely unknown. Using data from the 2019-2020 National Health and Resilience in Veterans Study (n = 1,353), this study examined the association between homelessness and contact with CLS among U.S. combat veterans and evaluated the mediating roles of shared risk factors. Findings revealed that a large proportion (53.6%) of the association between homelessness and CLS involvement was accounted for by indirect associations, most notably via drug use disorder (22.1%), moral injury (11.4%), and alcohol use disorder (10.7%). These findings highlight modifiable mechanisms that may link homelessness to CLS involvement, which may help inform targeted prevention efforts to mitigate the risk for CLS involvement among combat veterans.

Keywords

criminal legal system veterans homelessness combat mediating factors

Introduction

Homelessness and criminal legal system (CLS) involvement act as risk factors for one another, such that individuals experiencing homelessness are more likely to become CLS-involved and those with former CLS involvement are more likely to become homeless (Edwards, Barnes, et al., 2021; Finlay et al., 2016; Greenberg & Rosenheck, 2008). According to the results of different studies, it is estimated that between 25% and 75% of CLS-involved veterans experience homelessness both prior to and after incarceration (Blue-Howells et al., 2019). In a national sample of veterans who received Health Care for Re-Entry Veterans outreach services, for instance, 30% of veterans in prison had a history of homelessness and were experiencing homelessness at the time of their incarceration (Tsai et al., 2014). After incarceration, national data suggest that about 25% of veterans experienced homelessness in the year after they received outreach from Veteran Justice Outreach programs (Finlay et al., 2016). Many veterans are caught in a cycle of contact with the CLS, with the majority having at least one prior incarceration and 43% having been arrested four or more times (Bronson et al., 2015).

Studies conducted on the general homeless population emphasize that CLS involvement usually results from the psychosocial risk factors associated with homelessness—such as poor health, substance abuse, and severe mental health problems—rather than the state of homelessness itself (Greenberg & Rosenheck, 2008). In addition to the association between homelessness and these psychosocial outcomes, homelessness may also prevent individuals from effectively coping with the existing stressors, which in turn may lead to criminal pathways as an alternative coping strategy. In comparison to the general population, multiple studies have found a higher prevalence of homeless individuals with mental illnesses in the CLS (Solomon & Draine, 1999; White et al., 2006). Overall, the findings of the extant research imply that homelessness may lead to CLS involvement through mediating factors such as mental health and substance use. Although there is ample evidence showing common risk factors associated with both homelessness and CLS involvement among combat veterans (Edwards, Dichiara, et al., 2021) their intersection is largely under-examined and potential mechanisms linking homelessness to CLS involvement remain largely unexplored. Elucidation of mechanisms that may mediate the relationship between homelessness and CLS involvement among combat veterans has implications for building a more comprehensive model of care and addressing psychosocial and clinical risk factors amenable to interventions such as psychotherapy or case management.

Risk Factors for Homelessness and CLS Involvement

Substance use is one of the major risk factors for CLS involvement (Finlay et al., 2019) and CLS-involved veterans have the highest rates of alcohol and drug use disorders (DUDs) within the veteran population, with rates as high as 71% and 65%, respectively (Blodgett et al., 2015). Veterans experiencing homelessness also have high rates of substance use disorders, with one study documenting that among veterans entering Veterans Affairs (VA)-supported housing, 60% had a substance use disorder diagnosis, 54% of which also had been diagnosed with a comorbid alcohol and DUD (Tsai et al., 2014). According to the 2016 Survey of Prison Inmates, about 8.6% of male veterans in state prisons and 29.6% of veterans in federal prisons had been incarcerated for drug-related offenses (Maruschak et al., 2021). Given the high rates of substance use among veterans who have experienced homelessness and incarceration, substance use may intercede the relationship between homelessness and CLS involvement.

In addition to substance use, anger and depression are other conditions that may affect the association between homelessness and CLS involvement among combat veterans. As proposed by general strain theory, anger and depression are two of the negative affective states that result from exposure to strain and may lead to criminal acts (Agnew & White, 1992). Although anger and aggression can be adaptive in threatening contexts, such as combat, they become maladaptive when individuals are no longer in danger (Miles et al., 2020). Survival Mode Theory proposes that post-traumatic stress disorder (PTSD)-related emotions and cognitions may result in ambiguous or neutral events being interpreted as threatening, and result in anger or aggressive behaviors (Novaco et al., 2012). PTSD, however, is not the sole mechanism for anger as anger has been associated with veteran homelessness even after adjustment for PTSD, depression, and substance use (Adler et al., 2022). One explanation for the link between anger and homelessness is that anger may undermine veterans’ ability to hold a job and maintain social connections that may act as sources of shelter (Adler et al., 2022). Depression, however, is the most prevalent non-substance use mental disorder among CLS-involved veterans, with estimates as high as 51% (Blodgett et al., 2015). Similar to the general strain theory, the depression-driven hypothesis of crime suggests that externalizing behaviors such as crime and substance use may be an expression of emotional distress or a coping strategy for depression (Elizabeth Kim et al., 2019). Although both anger and depression are associated with both arrest and homelessness in veterans, individually, they may also mediate the relationship between them. To the best of our knowledge, no study has examined whether depression may mediate the relationship between homelessness and CLS involvement.

Many homeless individuals are socially isolated or have limited social support networks. Among homeless veterans, research suggests that greater social support is associated with positive outcomes (O’Connell & Rosenheck, 2016). Although there are limited studies of social support among CLS-involved veterans, its association with crime, incarceration, and recidivism in the civilian population is well documented (Alward et al., 2020; Beaver et al., 2014). It is possible that homeless veterans with levels of higher social support are less likely to engage in behaviors that place them at risk for incarceration (e.g., violence, theft), or they may be shielded from situations that are implicated (e.g., trespassing, public drunkenness).

In addition to the previously reviewed factors, there are other conditions unique to the military experience that may affect homelessness and CLS involvement among combat veterans. PTSD, for instance, is one of these risk factors with an estimated lifetime prevalence among homeless veterans being about 30%, and homelessness has been proposed to be a potentially traumatic event in and of itself (Tsai et al., 2020). Notably, veterans with PTSD are also at an increased risk for CLS involvement, with a recent meta-analysis documenting that veterans with PTSD have 50% greater odds of incarceration compared with veterans without PTSD (Taylor et al., 2020). Estimates suggest that about 30% of veterans in jail and about 25% of male veterans in prison have been diagnosed with PTSD (Bronson et al., 2015). General strain theory proposes that exposure to traumatic events may indirectly increase antisocial behavior via posttraumatic negative affect (Agnew & White, 1992). As such, PTSD may increase contact with the CLS via increased engagement in externalizing behaviors such as aggressive or impulsive/dysregulated behavior or substance use.

Moral injury, a character wound resulting from perpetrating, witnessing, or learning about events that violate one’s personal values, beliefs, or expectations (Litz et al., 2009), and/or experiencing betrayal by a trusted authority (Shay, 2014), has received less attention as a risk factor for homelessness and CLS involvement. Although related to PTSD, moral injury is a distinct construct characterized by guilt, shame, social isolation, anhedonia, and anger (Bryan et al., 2018). In this way, the moral injury could explain additional variance in functional impairment, such as self-destructive behavior, beyond that which is explained by PTSD (Griffin et al., 2019). Individuals may utilize aggressive or violent behavior to downregulate emotional experiences such as shame that occur in response to a sense of inadequacy or social failure (Elison et al., 2014). Although moral injury has largely been excluded from investigations as a risk factor for incarceration, a growing body of literature suggests that it may be salient among CLS-involved veterans (Martin et al., 2020). Homelessness may increase exposure to potentially morally injurious events (PMIEs) or exacerbate the already existing moral injury, which may lead to incarceration and interaction with the CLS (Martin et al., 2020).

The Current Study

Extant studies examining the relationship between homelessness and CLS involvement have been limited to veterans in treatment, housing, or justice programs at the VA. The current study analyzed data from a nationally representative survey of U.S. combat veterans to evaluate the modifiable or dynamic clinical factors that mediate the relationship between homelessness and CLS involvement. Based on prior research, the authors hypothesized that

Hypothesis 1 (H1): Homelessness would be associated with greater odds of CLS contact.

Hypothesis 2 (H2): Substance use problems, anger, depression, moral injury, PTSD, and social support would mediate this association.

Because combat exposure is associated with higher rates of homelessness (Tsai et al., 2013) and CLS involvement (Larson & Norman, 2014), as well as increased risk of PTSD, substance use (Bray et al., 2013), moral injury, depression, and problematic anger (Miles et al., 2017), the sample was restricted to combat veterans.

Method

Data and Sample

Data came from the National Health and Resilience in Veterans Study, a nationally representative, population-based study of U.S. military veterans conducted between November 18, 2019, and March 8, 2020 (Tsai et al., 2020). The sample was drawn from the Ipsos KnowledgePanel, a probability-based research panel of over 50,000 households, which consists of a nationally representative sample of all U.S. adults (both civilians and veterans). Panel members were recruited through national random samples by telephone and postal mail. Households were provided with access to the internet and computer hardware if needed. Dual sampling frames included listed and unlisted telephone numbers and households with and without telephone and internet access (Tsai et al., 2020). Of the 7,860 veterans in the panel, 4,069 (51.7%) completed a 50-min anonymous web-based survey. Census data from the 2019 Veterans Supplement of the Current Population Survey were used to compute the post-stratification weights to allow for the generalizability of study results to the national population of U.S. veterans. Of the 4,069 veterans in the sample, 1,353 (33.2%) reported that they served in a combat or war zone and were included in the current study. All participants in the survey provided informed consent and the Institutional Review Board (IRB) approval was given by the Human Subjects Committee of the VA Connecticut Healthcare System.

Measures

Consistent with the previously used measurement techniques (Chen et al., 2016; Edwards, Dichiara, et al., 2021), contact with CLS was assessed by a binary indicator of whether the respondents have ever been arrested. Respondents who reported they were arrested were assigned a value of “1,” otherwise they were assigned a value of “0.” Similar to its use in the National Epidemiologic Survey of Alcohol and Related Conditions – III (NESARC-III) (Chen et al., 2016), history of prior homelessness was assessed with the following question: “In your entire adult life, have you ever been homeless (i.e., stayed in a shelter, transitional housing, outdoors, or some other unstable housing situation)?” Veterans who responded affirmatively were coded as “1,” while those who reported no history of homelessness were coded as “0.”

Seven measures were examined as potential mediating factors between homelessness and contact with CLS. A modified self-report version of the alcohol use disorder (AUD) module from the Mini International Neuropsychiatric Interview (MINI) for DSM-5 (Grant et al., 2015; Sheehan et al., 2016) was used to assess lifetime AUD. A modified self-report version of the DUD module from the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association, 2013) version of the MINI Neuropsychiatric Interview (Panza et al., 2022; Sheehan et al., 2016) was used to measure lifetime DUD. Respondents who were positively screened for mild, moderate, and severe DUD were assigned the value of “1.” Drugs included stimulants, cocaine, narcotics, hallucinogens, phencyclidine, inhalants, cannabis, tranquilizers, and miscellaneous (e.g., steroids, nonprescription sleep or diet pills). All items from the respective AUD and DUD modules of the MINI Neuropsychiatric Interview for DSM-5 were administered and there were no modifications to item content other than that the items were administered in a self-report instead of an interview format.

The hostility subscale of the Symptom Checklist-90-Revised (Derogatis, 1994) was used to assess anger (α = .89). Respondents were asked to report the extent to which they have been distressed or bothered by six hostility symptoms in the past month, including (a) having urges to break or smash things, (b) having urges to beat, injure, or harm someone, (c) feeling easily annoyed or irritated, (d) temper outbursts that they could not control, (e) getting into frequent arguments, and (f) shouting or throwing things. Response categories for each item ranged from “0 = not at all” to “4 = extremely.” Although labeled as hostility, the construct is more aligned with anger, as it measures the thoughts, feelings, and actions characteristic of anger (Dillon et al., 2020). To assess lifetime major depressive disorder (MDD), a dichotomous measure was created based on a modified self-report version of the major depressive disorder module from the DSM-5 version of the MINI Neuropsychiatric Interview (Panza et al., 2022; Sheehan et al., 2016). Just as the measures of AUD and DUD, all items from the MDD module of the MINI Neuropsychiatric Interview for DSM-5 were administered.

Social support was measured by a 5-item scale from the Medical Outcomes Study Social Support Survey (Sherbourne & Stewart, 1991), asking respondents to what extent the following supports were available to them if needed: (a) Someone to confide in or talk to about their problems, (b) Someone to get together with for relaxation, (c) Someone to help them with daily chores if they were sick, (d) Someone to turn to for suggestion, and (e) Someone to love and make you feel wanted (α = .91). Response categories ranged from “0 = none of the time” to “4 = all of the time.” The social support measure was reverse-coded, so the higher score shows a lower social support level. Moral injury was assessed by the Moral Injury Events Scale (MIES; Nash et al., 2013), a 9-item self-report measure that assesses exposure to PMIEs such as witnessing others’ immoral acts, violating one’s own moral codes, and feeling betrayed by leaders and fellow service members (α = .90). Response categories for each item ranged from “1 = strongly disagree” to “6 = strongly agree.” PTSD was assessed using the Posttraumatic Stress Disorder Checklist for DSM-5 (Weathers et al., 2013), a 20-item self-report assessment of PTSD symptoms developed by the National Center for PTSD. Total scores ranged from 0 to 80, and a dichotomous measure with a cut-off score of 33 was used as indicative of lifetime PTSD (α = .97; Blevins et al., 2015).

Potentially confounding sociodemographic and military background measures were also controlled for. Respondents’ sex, race/ethnicity, and age were obtained through their self-reports. Military service duration included eight categories ranging from “1 = 6 months or less” to “8 = 20 years and over.” Educational attainment included 14 categories ranging from “1 = no formal education” to “14 = professional or doctorate degree.” Finally, religiosity, which is associated with better mental health (Tsai & Rosenheck, 2011), lower rates of substance abuse (Torchalla et al., 2014), and lower levels of crime (Chu, 2007; Jang & Johnson, 2003; Pirutinsky, 2014) was assessed by using the Duke University Religion Index (Koenig & Büssing, 2010), a 5-item scale that assesses three dimensions of religiosity, including organizational, non-organizational, and intrinsic religiosity (α = .90).

Data Analysis

Statistical analyses in the current study proceeded in several interconnected steps. First, descriptive statistics and bivariate correlations among the variables were computed. Second, since most of the dependent, independent, and moderating variables characterized low base rate events and conditions, receiver-operating characteristic (ROC) analyses were conducted to evaluate the extent to which individuals with a prior history of homelessness or CLS involvement could be differentiated from those without on the basis of key predictors. The ROC analysis facilitates the assessment of the sensitivity (true positive) and specificity (true negative) of a measure to determine whether it efficiently classifies the membership in a dichotomous outcome. True positives and false positives are virtually equivalent when the probability of a test is .50 under the curve (i.e., null hypothesis), while a value of 1.0 indicates perfect accuracy prediction (DeLisi & Vaughn, 2008). The effect sizes were determined based on the guidelines by Rice and Harris (2005), which suggest that an area under the curve (AUC) of .56 (r = .10, d = .20) represents a small effect size, .64 (r = .24, d = .50) represents a moderate effect size, and .71 (r = .37, d = .80) represents a large effect size. Then, since the dependent variable—contact with the CLS—was a dichotomous measure, a series of binary logistic regression analyses were performed. Model 1 examined the association between homelessness and contact with CLS while also controlling for potential confounders. In subsequent models (i.e., Model 2 through Model 8), the mediators were included to assess their associations separately with contact with CLS in the presence of homelessness and other covariates. In Model 9, all mediators were included along with other covariates to examine if the association between each mediator and contact with CLS changed in the presence of other mediators.

Finally, to assess mediation and reveal the direct and indirect associations between homelessness and contact with CLS via mediating variables, the Karlson–Holm–Breen (KHB) method was used. (Karlson et al., 2012; Kohler et al., 2011; Orak & Solakoglu, 2021). The KHB method permits disentangling the mediating effect from the rescaling effect in nonlinear models and makes it possible to reveal how much of the change in the coefficient of the independent variable is caused by the addition of the mediating variable (Kohler et al., 2011). It also allows for decomposing the total effect of the independent variable into the direct and indirect effects, enabling us to display the relative magnitude of each specific mediating effect while also controlling for potential confounders. Little’s MCAR (missing completely at random) test revealed that missing values (~8%) were missing completely at random and the listwise deletion method was used to handle these observations, which allowed the authors to use the KHB method to reveal the proportions of indirect effects of homelessness on contact with the CLS via each mediating variable. Post-stratification weights for all inferential analyses were applied to permit the generalizability of results to all U.S. combat veterans. An examination of pairwise correlations between variables and collinearity diagnostics (i.e., variance inflation factors and tolerance values) showed that multicollinearity was not an issue in the analyses. The VIF values ranged from 1.07 to 1.85 with a mean value of 1.26, and the tolerance values ranged from .54 to .93. For all calculations and statistical analyses, Stata 17 MP statistical software was used (StataCorp, 2021).

Results

Descriptive Statistics

Table 1 presents the descriptive statistics for the study variables. About 32% of respondents reported having contact with the CLS (i.e., arrested) at least once in their lifetime, and approximately 8% reported a history of homelessness. About 42% of respondents were identified as having AUD and about 11% had DUD. The mean score for anger was 2.17, indicating that on average, respondents had low levels of anger. About 16% of respondents were identified as having MDD. The mean score for social support was about 7.53, showing that on average, respondents had moderate-to-high levels of social support. The mean score for moral injury was about 18.35 on a scale from 9 to 54, suggesting that respondents on average, had moderate-to-low levels of moral injury. Approximately 15% of respondents had lifetime PTSD. About 6% of the sample was female and about 20% had underrepresented racial/ethnic identities (U.R.I.). Respondents’ ages ranged from 25 to 98, with a mean of 65.40. The mean score for the ordinal military years measure was about 5.14, which corresponds to an average military service duration of 6–9 years. The mean score for religiosity was about 10.63, indicating that on average, respondents had moderate religiosity levels. The mean score for the ordinal educational attainment measure was about 11.21, which corresponds to an associate’s degree.

Table 1:

Descriptive Statistics of All Study Variables

Variables	M/Percentage	SD	Minimum	Maximum	N
Contact with the CLS	32%	—	—	—	1,347
Homelessness	8%	—	—	—	1,347
Alcohol use disorder	42%	—	—	—	1,337
Drug use disorder	11%	—	—	—	1,353
Anger	2.17	3.67	0	23	1,321
Major depressive disorder	16%	—	—	—	1,343
Social support	7.53	5.27	1	21	1,328
Moral injury	18.34	10.04	9	54	1,333
PTSD	15%	—	—	—	1,296
Female	6%	—	—	—	1,353
Underrepresented racial/ethnic identities (U.R.I.)	20.10%	—	—	—	1,353
Hispanic	8.65%	—	—	—	1,353
Black, non-Hispanic	7.61%	—	—	—	1,353
2+ races, non-Hispanic	2.07%	—	—	—	1,353
Other, non-Hispanic	1.77%	—	—	—	1,353
Age	65.40	13.89	25	98	1,353
Military years (ordinal)	5.14	2.09	2	8	1,353
Religiosity	10.63	6.98	0	22	1,322
Educational attainment (ordinal)	11.21	1.46	1	14	1,353

Note. CLS = criminal legal system; PTSD = post-traumatic stress disorder.

Bivariate Associations

Bivariate associations between the key variables are presented in Table 2. Results of these analyses revealed that homelessness was positively associated with contact with the CLS and all mediating variables. In addition, all mediating variables were positively associated with CLS involvement.

Table 2:

Bivariate Correlations Among Study Variables

Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1.
2.	.17*
3.	.26*	.11*
4.	.21*	.25*	.24*
5.	.11*	.25*	.17*	.19*
6.	.12*	.19*	.24*	.21*	.42*
7.	.10*	.21*	.08*	.10*	.31*	.20*
8.	.17*	.18*	.15*	.18*	.32*	.24*	.18*
9.	.08*	.23*	.12*	.11*	.55*	.44*	.22*	.34*
10.	−.08*	.02*	−.01*	.01	.09*	.13*	.01	.05	.12*
11.	−.05*	−.15*	−.08*	−.14*	−.31*	−.24*	−.12*	−.05*	−.23*	−.27*
12.	−.09*	−.13*	−.06*	−.08*	−.01	−.05*	−.33*	−.03	−.04	−.05*	.07*
13.	−.05*	−.02*	−.01*	−.08*	.01	.01	−.04	−.09*	.02	.10*	−.14*	.02
14.	−.06*	−.06*	−.04*	−.07*	−.02	−.04	−.18*	−.01	−.01	.02	.12*	.13*	.02
15.	.09	.10	−.01*	.07	.13*	.07*	.08	−.11*	.10*	.10*	−.27*	−.08*	.04	.05*
16.	−.11*	−.04*	−.04*	−.08*	−.01	.04	−.03	−.01	.01	.08*	.03	.01	.17*	.05	−.05

Note. 1: contact with the CLS; 2: homelessness; 3: AUD; 4: DUD; 5: anger; 6: MDD; 7: social support; 8: moral injury; 9: PTSD; 10: female; 11: age; 12: married; 13: military years; 14: religiosity; 15: underrepresented racial/ethnic identities (U.R.I); 16: education. AUD = alcohol use disorder; CLS = criminal legal system; DUD = drug use disorder; MDD = major depressive disorder; PTSD = post-traumatic stress disorder.

p < .05.

Table 3 presents the results of the ROC analyses evaluating key predictors of homelessness and contact with CLS, which includes the AUC as an indicator of the effect size along with the standard errors and 95% confidence intervals. The analyses for the contact with CLS showed that all predictors had a small effect size, except for the AUD (AUC = .64), which had a moderate effect size. The analyses for homelessness resulted in somewhat larger effect sizes, in which AUD had a small effect size, DUD, MDD, social support, moral injury, and PTSD had moderate effect sizes, and anger had a large effect size. Overall, although many of the predictors had low effect sizes, especially for the contact with CLS, the fact that all of them had effect sizes >.50 indicates that there are apparent distributional differences in contact with CLS and homelessness based on the key predictors.

Table 3:

Receiver-Operating Characteristic (ROC) Analyses Evaluating Key Predictors of Homelessness and Contact With CLS

Variables	AUC	SE	Asymptotic normal (95% CI)	N
Outcome: contact with CLS
Homelessness	.55	.01	[.53, .57]	1,344
AUD	.64	.01	[.61, .67]	1,336
DUD	.57	.01	[.55, .59]	1,347
Anger	.57	.02	[.54, .60]	1,316
MDD	.55	.01	[.53, .57]	1,340
Social support	.56	.02	[.52, .59]	1,323
Moral injury	.60	.01	[.57, .63]	1,328
PTSD	.54	.01	[.51, .56]	1,292
Outcome: homelessness
AUD	.60	.02	[.55, .65]	1,333
DUD	.65	.02	[.60, .69]	1,347
Anger	.73	.02	[.68, .78]	1,317
MDD	.63	.02	[.58, .68]	1,338
Social support	.70	.03	[.64, .75]	1,324
Moral injury	.67	.03	[.62, .73]	1,328
PTSD	.65	.02	[.60, .70]	1,292

Note. All AUCs are statistically significant at p < .05. CLS = criminal legal system; AUD = alcohol use disorder; DUD = drug use disorder; MDD = major depressive disorder; AUC = area under the curve; SE = standard error; CI = confidence interval.

Logistic Regression Results

Table 4 presents the results of multiple logistic regression analyses. Model 1 showed that homelessness was associated with 2.5-fold greater odds of contact with the CLS (p = .002). In subsequent models (i.e., Model 2 through Model 8), the mediating variables—AUD, DUD, anger, MDD, social support, moral injury, and PTSD, respectively—were added to the analyses to examine their associations with contact with the CLS in the presence of independent and control variables. Results of these models demonstrated that AUD (p < .001), DUD (p < .001), and moral injury (p = .002) were significantly associated with increased odds of contact with the CLS. Finally, in Model 9, all mediating variables were included to see if there would be changes in the association between each mediating variable and contact with CLS in the presence of other mediators. Results showed that AUD, DUD, and moral injury were positively and significantly associated with contact with CLS. Specifically, holding all other variables constant, DUD was associated with about 2-fold greater odds (p = .01) of contact with the CLS, AUD with 69% greater odds (p = .003), and moral injury with about 3% increased odds per unit increase in this scale (p < .02). Further analyses with standardized scores on the moral injury scale revealed that for each standard deviation increase in moral injury, the odds of CLS contact increased by about 30% (p = .01). Other mediating variables did not show significant associations with contact with the CLS, suggesting that they did not mediate the association between homelessness and contact with the CLS.

Table 4:

Logistic Regression Predicting Odds of Contact With the CLS Among Combat Veterans

Variables/Models	Model 1		Model 2		Model 3		Model 4		Model 5		Model 6		Model 7		Model 8		Model 9
Variables/Models	OR	SE	OR	SE	OR	SE	OR	SE	OR	SE	OR	SE	OR	SE	OR	SE	OR	SE
Homeless.	2.51	.76**	2.22	.71*	2.01	.64*	2.22	.68*	2.33	.71*	2.37	.72**	2.22	.68*	2.33	.74*	1.57	.56
Female	0.31	.12**	0.34	.13**	0.31	.12**	0.30	.11***	0.29	.12**	0.32	.12**	0.28	.10***	0.31	.12**	0.38	.14*
Age	0.99	.006	0.99	.006	0.99	.006	0.99	.006	0.99	.006	0.99	.006	0.99	.006	0.99	.006	1.01	.007
Married	0.64	.12*	0.70	.13	0.68	.12*	0.61	.11*	0.64	.12*	0.69	.13	0.62	.11*	0.66	.12*	0.75	.16
Military years	1.01	.04	1.01	.04	1.02	.04	1.03	.04	1.02	.04	1.01	.04	1.03	.04	0.99	.04	1.03	.05
Religion	1.00	.01	1.00	.01	1.00	.01	1.00	.01	1.00	.01	1.00	.01	1.00	.01	1.00	.01	1.01	.01
U.R.I.	1.57	.33*	1.66	.37*	1.60	.34*	1.58	.34*	1.59	.34*	1.50	.32	1.57	.34*	1.62	.35*	1.61	.38*
Education	0.92	.05	0.93	.05	0.93	.05	0.92	.05	0.91	.05	0.92	.05	0.91	.04	0.93	.05	0.94	.05
AUD			2.05	.34***													1.69	.30**
DUD					2.50	.67***											2.00	.56**
Anger							1.04	.02									1.03	.03
MDD									1.48	.33							1.05	.28
Social support											1.03	.02					1.02	.02
Moral injury													1.03	.008**			1.03	.009*
PTSD															1.18	.28	.64	.20
Constant	1.76	1.33	.96	.75	1.10	.85	1.25	.95	1.53	1.17	1.27	.99	1.11	.87	1.46	1.13	.29	.25
Pseudo R²	.05		.07		.06		.06		.05		.05		.06		.05		.09
Model χ²	35.49***		45.13***		39.49***		36.92***		40.08***		41.00***		48.19***		33.16***		66.60***
Log likelihood	−847.50		−827.03		−834.64		−826.88		−843.27		−837.53		−825.12		−813.23		−750.15
N	1,317		1,309		1,317		1,294		1,314		1,303		1,302		1,271		1,223

Note. OR = odds ratio; SE = robust standard errors; U.R.I = underrepresented racial/ethnic identities; AUD = alcohol use disorder; DUD = drug use disorder; MDD = major depressive disorder.

p < .05. **p < .01. ***p < .001.

Proportions of Direct and Indirect Associations

Table 5 shows the results of the KHB analyses that are presented with coefficients and percentages representing the total effect and the proportion of the total effect accounted for by the direct and indirect effects via mediating variables. A total of 53.6% of the association between homelessness and contact with the CLS was accounted for by indirect associations, most notably 22.1% via DUD (p < .001), 11.4% via moral injury (p = .01), and 10.7% via AUD (p < .001). As observed in prior logistic regression analyses, the mediating effects of anger, MDD, social support, and PTSD were not statistically significant.

Table 5:

Decomposition of Direct and Indirect Effects of Homelessness on Contact With CLS

Direct and indirect effects	B	SE	%
Homelessness
Total effect	.97**	.34	100
Direct effect	.45	.35	46.40
Indirect effect	.52***	.14	53.60
via AUD	.10***	.05	10.75
via DUD	.21***	.09	22.06
via anger	.14	.12	14.13
via MDD	.009	.05	0.93
via social support	.06	.05	6.08
via moral injury	.11**	.06	11.42
via PTSD	−.11	.09	−11.78

Note. The model is adjusted for sex, age, marital status, military service duration, religiosity, race, and educational attainment (n = 1,223). B = coefficient; SE = robust standard errors; AUD = alcohol use disorder; DUD = drug use disorder; MDD = major depressive disorder.

p < .05. **p < .01. ***p < .001.

Discussion

The Relationship Between Homelessness and CLS Involvement

The bidirectional association between homelessness and contact with CLS is well established in previous research that has focused on veterans and the general population (Edwards, Dichiara, et al., 2021; Finlay et al., 2016; Greenberg & Rosenheck, 2008). The fact that both homelessness and contact with CLS share numerous interrelated and dynamic psychosocial risk factors, however, points to an indirect association between these outcomes. Although several risk factors for homelessness and CLS involvement have been identified in the previous research, no known study has evaluated whether these shared risk factors mediate the association between homelessness and CLS involvement among combat veterans. To address this gap, this study analyzed data from a nationally representative sample of veterans to examine the association between homelessness and contact with the CLS among combat veterans, and whether shared social and clinical factors may mediate this relationship.

Results of the current study provide compelling evidence that contact with the CLS among veterans is partly attributable to homelessness and that substance use (alcohol use and drug use) and moral injury may mediate this association. The substance use mechanisms accounted for a substantial proportion (approximately 70%) of the mediating effect. As such, these results support the study hypotheses and previous research regarding the salience of alcohol and drug use difficulties that occur in this population (Snowden et al., 2017). Furthermore, moral injury also provided an important mediating effect, suggesting that additional research on this construct as well as clinical attention could be fruitful. However, contrary to the expectations, anger, depression, social support, and PTSD were not significant mediators. Although the findings should be considered exploratory given the cross-sectional nature of the study, the lack of significant mediating effects for these mechanisms might give pause to their role as potential mediators of the connection between homelessness and CLS involvement among combat veterans, and underscores the need for follow-up research employing longitudinal designs. Taken together, this study builds upon prior research by exploring some of the key mechanisms suggested by prior descriptive studies.

Substance Use and Moral Injury as Significant Mediators

The findings on substance use are disconcerting, given the challenges that exist in reaching this population and engaging them in treatment. Despite the obvious challenges, identifying the role of substance use in homelessness and CLS involvement indicates possible, critical periods for substance use intervention. It is likely that substance use treatment during periods of homelessness may help reduce the odds of individuals becoming involved in the CLS, with a recent, systematic review suggesting that legal and criminal problem severity was reduced among veterans with substance use disorder after substance use treatment (Timko et al., 2020).

The results pertaining to the influence of moral injury highlight the clinical relevance of this experience in homelessness and CLS involvement. Homelessness represents a context wherein veterans may be at risk for greater exposure to potentially traumatic events (Tsai et al., 2020). Although not all events are currently considered to be Criterion A events by DSM-5 (e.g., homelessness itself), those characterized by betrayal and transgression of moral values may still be morally injurious and result in significant distress and dysfunction (Koenig et al., 2019). In this way, the moral injury may be more ubiquitous in capturing the lived experiences across the types of experiences homeless and incarcerated persons face rather than isolated, life-threatening, or fear-based traumatic events. This may be an explanation for why PTSD was overshadowed in the final model. Furthermore, affective components of moral injury (e.g., guilt and shame) may be compounded and reinforced by the stigma associated with homelessness and/or CLS involvement and reduce help-seeking for distress related to PMIEs or other health care needs that might lower the risk for incarceration (O’Toole et al., 2015). In fact, it is suggested that when criminal behaviors are stigmatized by society, shame increases and functions in a way that leads to criminogenic risk factors such as substance use (Tangney et al., 2011). Future studies should seek to examine moral injury among homeless veterans, and further assess to what extent PMIEs were the result of military experiences (e.g., combat) or homelessness. Results from such studies could inform adaptations of existing interventions such as PTSD treatments that address moral injury within the context of homelessness.

The Roles of Anger, MDD, Social Support, and PTSD

Study findings on the role of anger, MDD, social support, and PTSD as potential mediating mechanisms did not reveal a significant mediating effect on the relationship between homelessness and CLS involvement and require additional commentary. The null associations may be due in part to the nature of the relationship between these factors and moral injury. Disentangling the interplay between moral injury, anger, MDD, PTSD, and social support is complex given the evolving conceptualization of moral injury and distinguishing core symptoms from clinical outcomes (Griffin et al., 2019). Anger, for instance, has been described as a defining characteristic of moral injury and key symptoms of MDD, such as anhedonia, load onto the construct of moral injury (Bryan et al., 2018; Roth et al., 2022). Thus, anger and MDD may be functioning largely as proxies for moral injury; however, the correlations between moral injury and anger and MDD were moderate (r = .32, .24 respectively). It is possible that when examining MDD, anger, and PTSD concurrently with moral injury, a different relationship with CLS involvement emerges. It is important to note, however, that even when these measures were entered into the logistic regression analyses separately, they still did not have any significant associations with CLS involvement. This may suggest that the construct of moral injury better captures the “posttraumatic negative affect” hypothesized in general strain theory to contribute to CLS involvement than MDD, anger, or PTSD alone.

Secondary outcomes of moral injury include reduced social support (Jinkerson, 2016), and social isolation is a characterizing moral injury feature in empirical models (Bryan et al., 2018), which may be why the social support measure was non-significant in the final model. Findings may also be attributable to respondents’ endorsement of moderate-to-high levels of social support, and because the assessment of social support did not necessarily reflect social support during homelessness. Research suggests, however, that many homeless adults who obtain housing remain socially isolated (Tsai et al., 2012).

Examining Different Components of Moral Injury

Some studies suggest that an examination of separate pathways of PMIEs—perpetration-based and betrayal-based—might provide a more nuanced understanding of moral injury (Frankfurt et al., 2018). The post hoc analyses revealed that only the component of betrayal significantly contributed to the indirect association between homelessness and CLS involvement. The betrayal may be related in part to interpersonal and military traumas experienced. An examination of different components of moral injury may provide important implications for interventions. The common therapeutic technique of Socratic Questioning or challenging the accuracy of self-blame, for instance, may be less appropriate than techniques such as trauma-informed guilt reduction therapy and adaptive disclosure, which recognizes responsibility while promoting the possibility of self-forgiveness and compassion (Gray et al., 2017; Norman et al., 2022). These considerations may be particularly relevant for further understanding moral injury as a mechanism between veteran homelessness and CLS involvement and guide intervention adaptation to contexts such as the CLS where personal responsibility is especially emphasized.

Study Limitations

Despite the uniqueness of the present study, the findings should be considered within the context of four limitations. First, the survey data were cross-sectional, which did not include any temporal order between variables and did not allow the use of measures to prevent potential spuriousness. Although many potential confounders were controlled for, it is still impossible to claim any causal associations among these variables. Second, sample characteristics restrict the generalizability of the results to other homeless adult populations. The sample was comprised of mostly male-identifying individuals and does not account for gendered pathways to CLS involvement and the unique needs of women experiencing homelessness. Relatedly, the sample was recruited from individuals currently housed, which excludes those who are currently homeless and/or incarcerated. This may pose a threat to the generalizability of the study results to the broader population of homeless individuals. Third, low scores on measures of anger, MDD, and PTSD and higher rates of social support may have artifactually increased the value of moral injury as a mechanism in the relationship between homelessness and CLS involvement. Fourth, although the 5-item social support measure assesses different aspects of perceived social support, including instrumental, tangible, and emotional support, the use of self-report and abbreviated measures is a limitation of the study, and further research using structured interviews and longer, multidimensional measures of social support is needed to evaluate the replicability and generalizability of the findings.

Conclusion

This study contributes to our understanding of the relationship between homelessness and CLS involvement among combat veterans, suggesting potentially modifiable mechanisms of CLS involvement for treatment targets such as alcohol and DUDs. Importantly, this study suggests that moral injury uniquely contributes to a mechanism like the association to CLS involvement over and above the contribution of PTSD and points to a new area of research for the field. Results of this study identify directions for future research that can continue to assist clinicians and researchers in identifying and testing potential factors associated with CLS involvement among homeless veterans.

Footnotes

Authors’ Note:

For all analyses in this study, the authors used data from The National Health and Resilience in Veterans Study (NHRVS). NHRVS is funded by the U.S. Department of Veterans Affairs National Center for PTSD, which had no role in the design, analysis, or interpretation of this study. The views expressed here are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. No funds, grants, or other support were received to assist with the preparation of this manuscript. The authors have no relevant financial or non-financial interests to disclose. The data that support the findings of this study are available from Robert Pietrzak, Ph.D., MPH at robert.pietrzak@yale.edu upon reasonable request. ().

ORCID iD

Ugur Orak

Ugur Orak, PhD, is an assistant professor in the Department of Sociology and Anthropology at Middle Tennessee State University. His research broadly addresses crime, deviance, and mental health, with a particular focus on where they intersect in a developmental context. His recent studies focused on the negative and protective impacts of military service on short- and long-term patterns of mental health, crime, and deviance.

Katherine Kelton, PhD MSPH, is a postdoctoral fellow with the National Center for Homelessness among Veterans and Geriatric Research Education Clinical Center. Kelton’s program of research has been heavily informed by her clinical experiences and public health lens and follows the common thread of how sociocultural, systemic inequities affect mental health. These topics span incarceration, homelessness, post-traumatic stress disorder, and substance use.

Michael G. Vaughn, PhD, is a professor and director of the PhD program in the School of Social Work, at Saint Louis University. His research agenda comprised of several interrelated foci such as temperament and psychopathy, drug abuse, violence, school dropout, biosocial interplay in human behavior, and health behavior and criminal justice.

Jack Tsai, PhD, serves as research director for the U.S. Department of Veterans Affairs Homeless Programs Office. He also serves as campus dean and professor of Public Health at the University of Texas Health Science Center at Houston.

Robert H. Pietrzak, PhD MPH, completed a BA in psychology at Clark University, MPH in epidemiology, and PhD in clinical psychology with a specialization in Clinical Neuropsychology at the University of Connecticut. He is the director of the Translational Psychiatric Epidemiology Laboratory in the Clinical Neurosciences Division of the U.S. Department of Veterans Affairs National Center for PTSD and Professor of Psychiatry at Yale School of Medicine and Professor of Public Health (Social and Behavioral Sciences) at Yale School of Public Health. His primary research interests include the psychosocial and genetic epidemiology of traumatic stress and resilience; dimensional models of stress-related psychopathology; and neurobiology of trauma-related disorders.

References

Adler

A. B.

LeardMann

C. A.

Yun

Jacobson

I. G.

Forbes

(2022). Problematic anger and economic difficulties: Findings from the Millennium Cohort Study. Journal of Affective Disorders, 297, 679–685. https://doi.org/10.1016/j.jad.2021.10.078

Agnew

White

H. R.

(1992). An empirical test of general strain theory. Criminology, 30(4), 475–500. https://doi.org/10.1111/j.1745-9125.1992.tb01113.x

Alward

L. M.

Caudy

M. S.

Viglione

(2020). Assessing the relative influence of individual attitudes, social supports, and neighborhood context on reentry outcomes: What changes matter most? Criminal Justice and Behavior, 47(11), 1487–1508. https://doi.org/10.1177/0093854819901158

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.).

Beaver

K. M.

Boutwell

B. B.

Barnes

J. C.

(2014). Social support or biosocial support?: A genetically informative analysis of social support and its relation to self-control. Criminal Justice and Behavior, 41(4), 453–470. https://doi.org/10.1177/0093854813504918

Blevins

C. A.

Weathers

F. W.

Davis

M. T.

Witte

T. K.

Domino

J. L.

(2015). The posttraumatic stress disorder checklist for DSM-5 (PCL-5): Development and initial psychometric evaluation. Journal of Traumatic Stress, 28, 489–498. https://doi.org/10.1002/jts.22059

Blodgett

J. C.

Avoundjian

Finlay

A. K.

Rosenthal

Asch

S. M.

Maisel

N. C.

Midboe

A. M.

(2015). Prevalence of mental health disorders among justice involved veterans. Epidemiologic Reviews, 37, 163–176. https://doi.org/10.1093/epirev/mxu003

Blue-Howells

Timko

Clark

Finlay

A. K.

(2019). Criminal justice issues among homeless Veterans. In Tsai

(Ed.), Homelessness among U.S. Veterans: Critical perspectives (pp. 109–137). Oxford University Press.

Bray

R. M.

Brown

J. M.

Williams

(2013). Trends in binge and heavy drinking, alcohol-related problems, and combat exposure in the US military. Substance Use & Misuse, 48(10), 799–810. https://doi.org/10.3109/10826084.2013.796990

10.

Bronson

Carson

Noonan

Berzofsky

(2015). Veterans in prison and jail, 2011–12. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. https://bjs.ojp.gov/content/pub/pdf/vpj1112.pdf

11.

Bryan

C. J.

Bryan

A. O.

Roberge

Leifker

F. R.

Rozek

D. C.

(2018). Moral injury, posttraumatic stress disorder, and suicidal behavior among National Guard personnel. Psychological Trauma: Theory, Research, Practice, and Policy, 10(1), 36–45. https://doi.org/10.1037/tra0000290

12.

Chen

C. M.

Slater

M. E.

Castle

I.-J. P.

Grant

B. F.

(2016). Alcohol use and alcohol use disorders in the United States: Main findings from the 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III). U.S. Alcohol Epidemiologic Data Reference Manual, 10(16-AA-8020), i–114. https://pubs.niaaa.nih.gov/publications/NESARC_DRM3/NESARC3DRM.pdf

13.

Chu

D. C.

(2007). Religiosity and desistance from drug use. Criminal Justice and Behavior, 34(5), 661–679. https://doi.org/10.1177/0093854806293485

14.

DeLisi

Vaughn

M. G.

(2008). The Gottfredson–Hirschi critiques revisited: Reconciling self-control theory, criminal careers, and career criminals. International Journal of Offender Therapy and Comparative Criminology, 52(5), 520–537. https://doi.org/10.1177/0306624X07308553

15.

Derogatis

L. R.

(1994). SCL-90-R: Administration, scoring and procedures manual (3rd ed.). NCS Pearson.

16.

Dillon

K. H.

Medenblik

A. M.

Mosher

T. M.

Elbogen

E. B.

Morland

L. A.

Beckham

J. C.

(2020). Using interpretation bias modification to reduce anger in veterans with posttraumatic stress disorder: A pilot study. Journal of Traumatic Stress, 33, 857–863. https://doi.org/10.1002/jts.22525

17.

Edwards

E. R.

Barnes

Govindarajulu

Geraci

Tsai

(2021). Mental health and substance use patterns associated with lifetime suicide attempt, incarceration, and homelessness: A latent class analysis of a nationally representative sample of U.S. veterans. Psychological Services, 18(4), 619–631. https://doi.org/10.1037/ser0000488

18.

Edwards

E. R.

Dichiara

Gromatsky

Tsai

Goodman

Pietrzak

(2021). Understanding risk in younger veterans: Risk and protective factors associated with suicide attempt, homelessness, and arrest in a nationally representative veteran sample. Military Psychology, 34(2), 175–186. https://doi.org/10.1080/08995605.2021.1982632

19.

Elison

Garofalo

Velotti

(2014). Shame and aggression: Theoretical considerations. Aggression and Violent Behavior, 19(4), 447–453. https://doi.org/10.1016/j.avb.2014.05.002

20.

Elizabeth Kim

B. K.

Gilman

A. B.

Kosterman

Hill

K. G

. (2019). Longitudinal associations among depression, substance abuse, and crime: A test of competing hypotheses for driving mechanisms. Journal of Criminal Justice, 62, 50–57. https://doi.org/10.1016/j.jcrimjus.2018.08.005

21.

Finlay

A. K.

Owens

M. D.

Taylor

Nash

Capdarest-Arest

Rosenthal

Blue-Howells

Clark

Timko

(2019). A scoping review of military veterans involved in the criminal justice system and their health and healthcare. Health & Justice, 7(1), Article 6. https://doi.org/10.1186/s40352-019-0086-9

22.

Finlay

A. K.

Smelson

Sawh

McGuire

Rosenthal

Blue-Howells

Timko

Binswanger

Frayne

S. M.

Blodgett

J. C.

Bowe

Clark

S. C.

Harris

A. H. S.

(2016). U.S. Department of Veterans Affairs Veterans Justice Outreach Program: Connecting justice-involved veterans with mental health and substance use disorder treatment. Criminal Justice Policy Review, 27(2), 203–222. https://doi.org/10.1177/0887403414562601

23.

Frankfurt

S. B.

DeBeer

B. B.

Morissette

S. B.

Kimbrel

N. A.

La Bash

Meyer

E. C.

(2018). Mechanisms of moral injury following military sexual trauma and combat in post-9/11 U.S. war veterans. Frontiers in Psychiatry, 9, Article 520. https://doi.org/10.3389/fpsyt.2018.00520

24.

Grant

B. F.

Goldstein

R. B.

Saha

T. D.

Chou

S. P.

Jung

Zhang

Pickering

R. P.

Ruan

W. J.

Smith

S. M.

Huang

Hasin

D. S.

(2015). Epidemiology of DSM-5 alcohol use disorder: Results from the national epidemiologic survey on alcohol and related conditions III. JAMA Psychiatry, 72(8), 757–766. https://doi.org/10.1001/jamapsychiatry.2015.0584

25.

Gray

M. J.

Nash

W. P.

Litz

B. T

. (2017). When self-blame is rational and appropriate: The limited utility of Socratic questioning in the context of moral injury: Commentary on Wachen et al. (2016). Cognitive and Behavioral Practice, 24(4), 383–387. https://doi.org/10.1016/j.cbpra.2017.03.001

26.

Greenberg

G. A.

Rosenheck

R. A.

(2008). Jail incarceration, homelessness, and mental health: A national study. Psychiatric Services, 59(2), 170–177. https://doi.org/10.1176/ps.2008.59.2.170

27.

Griffin

B. J.

Purcell

Burkman

Litz

B. T.

Bryan

C. J.

Schmitz

Villierme

Walsh

Maguen

(2019). Moral injury: An integrative review. Journal of Traumatic Stress, 32(3), 350–362. https://doi.org/10.1002/jts.22362

28.

Jang

S. J.

Johnson

B. R.

(2003). Strain, negative emotions, and deviant coping among African Americans: A test of General Strain Theory. Journal of Quantitative Criminology, 19, 79–105. https://doi.org/10.1023/A:1022570729068

29.

Jinkerson

J. D.

(2016). Defining and assessing moral injury: A syndrome perspective. Traumatology, 22, 122–130. https://doi.org/10.1037/trm0000069

30.

Karlson

K. B.

Holm

Breen

(2012). Comparing regression coefficients between same-sample nested models using logit and probit: A new method. Sociological Methodology, 42(1), 286–313. https://doi.org/10.1177/0081175012444861

31.

Koenig

H. G.

Büssing

(2010). The Duke University Religion Index (DUREL): A five-item measure for use in epidemiological studies. Religions, 1(1), 78–85. https://doi.org/10.3390/rel1010078

32.

Koenig

H. G.

Youssef

N. A.

Pearce

(2019). Assessment of moral injury in veterans and active duty military personnel with PTSD: A review. Frontiers in Psychiatry, 10, Article 443. https://doi.org/10.3389/fpsyt.2019.00443

33.

Kohler

Karlson

K. B.

Holm

(2011). Comparing coefficients of nested nonlinear probability models. Stata Journal, 11(3), 420–438. https://doi.org/10.1177/1536867x1101100306

34.

Larson

Norman

(2014). Prospective prediction of functional difficulties among recently separated veterans. Journal of Rehabilitation, Research & Development, 51(3), 415–428. https://doi.org/10.1682/jrrd.2013.06.0135

35.

Litz

B. T.

Stein

Delaney

Lebowitz

Nash

W. P.

Silva

Maguen

(2009). Moral injury and moral repair in war veterans: A preliminary model and intervention strategy. Clinical Psychology Review, 29, 695–706. https://doi.org/10.1016/j.cpr.2009.07.003

36.

Martin

B. W.

Holliday

LePage

J. P.

(2020, May 4). Trauma and diversity: moral injury among justice-involved veterans: An understudied clinical concern. International Society for Traumatic Stress Studies. https://istss.org/public-resources/trauma-blog/2020-may/trauma-and-diversity-moral-injury-among-justice-in

37.

Maruschak

L. M.

Bronson

Alper

(2021). Survey of prison inmates, 2016: Veterans in prison. Bureau of Justice Statistics. https://bjs.ojp.gov/content/pub/pdf/vpspi16st.pdf

38.

Miles

S. R.

Dillon

K. H.

Jacoby

V. M.

Hale

W. J.

Dondanville

K. A.

Wachen

J. S.

Yarvis

J. S.

Peterson

A. L.

Mintz

Litz

B. T.

Young-Mc Caughan

Resick

P. A.

, on behalf of the STRONG STAR Consortium. (2020). Changes in anger and aggression after treatment for PTSD in active duty military. Journal of Clinical Psychology, 76(3), 493–507. https://doi.org/10.1002/jclp.22878

39.

Miles

S. R.

Sharp

Tharp

A. T.

Stanford

M. S.

Stanley

Thompson

K. E.

Kent

T. A.

(2017). Emotion dysregulation as an underlying mechanism of impulsive aggression: Reviewing empirical data to inform treatments for veterans who perpetrate violence. Aggression and Violent Behavior, 34, 147–153. https://doi.org/10.1016/j.avb.2017.01.017

40.

Nash

W. P.

Marino Carper

T. L.

Mills

M. A.

Goldsmith

Litz

B. T.

(2013). Psychometric evaluation of the moral injury events scale. Military Medicine, 178(6), 646–652. https://doi.org/10.7205/MILMED-D-13-00017

41.

Norman

S. B.

Capone

Panza

K. E.

Haller

Davis

B. C.

Schnurr

P. P.

Shea

M. T.

Browne

Norman

G. J.

Lang

A. J.

Kline

A. C.

Golshan

Allard

C. B.

Angkaw

(2022). A clinical trial comparing trauma-informed guilt reduction therapy (TrIGR), a brief intervention for trauma-related guilt, to supportive care therapy. Depression and Anxiety, 39(4), 262–273. https://doi.org/10.1002/da.23244

42.

Novaco

R. W.

Swanson

R. D.

Gonzalez

O. I.

Gahm

G. A.

Reger

M. D.

(2012). Anger and postcombat mental health: Validation of a brief anger measure with U.S. soldiers post deployed from Iraq and Afghanistan. Psychological Assessment, 24(3), 661–675. https://doi.org/10.1037/a0026636

43.

O’Connell

M. J.

Rosenheck

(2016). The family ties that bind: Tangible, instrumental, and emotional support among homeless veterans. In MacDermid Wadsworth

Riggs

D. S.

(Eds.), War and family life (pp. 281–319). Springer International Publishing/Springer Nature. https://doi-org.ezp.slu.edu/10.1007/978-3-319-21488-7_15

44.

O’Toole

T. P.

Johnson

E. E.

Redihan

Borgia

Rose

(2015). Needing primary care but not getting it: The role of trust, stigma and organizational obstacles reported by homeless veterans. Journal of Health Care for the Poor and Underserved, 26(3), 1019–1031. https://doi.org/10.1353/hpu.2015.0077

45.

Orak

Solakoglu

(2021). Strain and deviant reactions in compulsory military service: An investigation of mediating mechanisms. Crime & Delinquency. Advance online publication. https://doi.org/10.1177/0011128721999337

46.

Panza

K. E.

Kline

A. C.

P. J.

Potenza

M. N.

Norman

S. B.

Pietrzak

R. H.

(2022). Epidemiology of DSM-5 alcohol use disorder in U.S. military veterans: Results from the National Health and Resilience in Veterans Study. Drug and Alcohol Dependence, 231, Article 109240. https://doi.org/10.1016/j.drugalcdep.2021.109240

47.

Pirutinsky

(2014). Does religiousness increase self-control and reduce criminal behavior?: A longitudinal analysis of adolescent offenders. Criminal Justice and Behavior, 41(11), 1290–1307. https://doi.org/10.1177/0093854814531962

48.

Rice

M. E.

Harris

G. T.

(2005). Comparing effect sizes in follow-up studies: ROC Area, Cohen’s d, and r. Law and Human Behavior, 29(5), 615–620. https://doi.org/10.1007/s10979-005-6832-7

49.

Roth

S. L.

Qureshi

Moulden

H. M.

Chaimowitz

G. A.

Lanius

R. A.

Losier

B. J.

Mckinnon

M. C.

(2022). “Trapped in their shame”: A qualitative investigation of moral injury in forensic psychiatry patients. Criminal Justice and Behavior, 49(4), 593–612. https://doi.org/10.1177/00938548211039877

50.

Shay

(2014). Moral injury. Psychoanalytic Psychology, 31, 182–191. https://doi.org/10.1037/a0036090

51.

Sheehan

Lecrubier

Janavs

Baker

Harnett Sheehan

Knapp

Sheehan

Weiller

Hergueta

Amorim

(2016). Mini International Neuropsychiatric Interview English version 7.0. 2 for DSM-5. University of Southern Florida College of Medicine.

52.

Sherbourne

C. D.

Stewart

A. L.

(1991). The MOS social support survey. Social Science & Medicine, 32(6), 705–714. https://doi.org/10.1016/0277-9536(91)90150-b

53.

Snowden

D. L.

Salas-Wright

C. P.

Vaughn

M. G.

King

(2017). Military service and crime: New evidence. Social Psychiatry & Psychiatric Epidemiology, 52, 605–615. https://doi.org/10.1007/s00127-017-1342-8

54.

Solomon

Draine

(1999). Using clinical and criminal involvement factors to explain homelessness among clients of a psychiatric probation and parole service. The Psychiatric Quarterly, 70(1), 75–87. https://doi.org/10.1023/a:1022083209201

55.

StataCorp. (2021). Stata statistical software: Release 17.

56.

Tangney

J. P.

Stuewig

Hafez

(2011). Shame, guilt, and remorse: Implications for offender populations. Journal of Forensic Psychiatry & Psychology, 22(5), 706–723. https://doi.org/10.1080/14789949.2011.617541

57.

Taylor

E. N.

Timko

Nash

Owens

M. D.

Harris

A. H. S.

Finlay

A. K.

(2020). Posttraumatic Stress Disorder and justice involvement among military veterans: A systematic review and meta-analysis. Journal of Traumatic Stress, 33(5), 804–812. https://doi.org/10.1002/jts.22526

58.

Timko

Nash

Owens

M. D.

Taylor

Finlay

A. K.

(2020). Systematic review of criminal and legal involvement after substance use and mental health treatment among veterans: Building toward needed research. Substance Abuse: Research and Treatment, 14, 1–13. https://doi.org/10.1177/1178221819901281

59.

Torchalla

Strehlau

Linden

I. A.

Krausz

(2014). Religious participation and substance use behaviors in a Canadian sample of homeless people. Community Mental Health Journal, 50(7), 862–869. https://doi.org/10.1007/s10597-014-9705-z

60.

Tsai

Mares

A. S.

Rosenheck

R. A.

(2012). Does housing chronically homeless adults lead to social integration? Psychiatric Services, 63(5), 427–434. https://doi.org/10.1176/appi.ps.201100047

61.

Tsai

Pietrzak

R. H.

Rosenheck

R. A.

(2013). Homeless veterans who served in Iraq and Afghanistan: Gender differences, combat exposure, and comparisons with previous cohorts of homeless veterans. Administration and Policy in Mental Health and Mental Health Services Research, 40(5), 400–405. https://doi.org/10.1007/s10488-012-0431-y

62.

Tsai

Rosenheck

R. A.

(2011). Religiosity among adults who are chronically homeless: Association with clinical and psychosocial outcomes. Psychiatric Services, 62, 1222–1224. https://doi.org/10.1176/ps.62.10.pss6210_1222

63.

Tsai

Rosenheck

R. A.

Kasprow

W. J.

McGuire

J. F.

(2014). Homelessness in a national sample of incarcerated veterans in state and federal prisons. Administration and Policy in Mental Health, 41(3), 360–367. https://doi.org/10.1007/s10488-013-0483-7

64.

Tsai

Schick

Hernandez

Pietrzak

R. H.

(2020). Is homelessness a traumatic event? Results from the 2019–2020 National Health and Resilience in Veterans Study. Depression and Anxiety, 37(11), 1137–1145. https://doi.org/10.1002/da.23098

65.

Weathers

F. W.

Blake

D. D.

Schnurr

P. P.

Kaloupek

D. G.

Marx

B. P.

Keane

T. M.

(2013). The Life Events Checklist for DSM-5 (LEC-5). National Center for PTSD. www.ptsd.va.gov

66.

White

M. C.

Chafetz

Collins-Bride

Nickens

(2006). History of arrest, incarceration and victimization in community-based severely mentally ill. Journal of Community Health, 31(2), 123–135. https://doi.org/10.1007/s10900-005-9005-1