Abstract
This study explores how typologies of adversity and mental health/substance use disorders impact rule violations during incarceration. Data come from the 2004 cross-sectional Survey of Inmates in State and Federal Correctional Facilities (SI-SFCF). Logistic regression and structural equation modeling were used for analysis. Results link history of adverse experiences to rule violations during incarceration and demonstrate how mental health and substance use disorders mediate this relationship. Incarcerated people with severe histories of adverse experiences had the highest odds of rule violations, relative to people with low adversity, for all typologies. More severe adversity typologies predicted mental health and substance use disorders. Alcohol and substance use disorders predicted drug violations, whereas substance use and mental health disorders predicted major violations. Serious mental illness did not predict rule violations when accounting for adversity. Findings suggest that addressing adverse experiences, mental health, and substance use disorders may prevent rule violations.
Keywords: prison misconduct, prison, mental health, substance use, alcohol, trauma, abuse, quantitative methods
INTRODUCTION
Breaking prison rules during incarceration is common. About half of people incarcerated in U.S. prisons have broken at least one rule during their incarceration (Celinska & Sung, 2014). Rule violations can be benign, such as having more than the allowed number of blankets, which could be considered possession of contraband. Rule violations can also be more serious, potentially qualifying as a new criminal charge, such as assault; in such serious cases, there are new harms and costs (French & Gendreau, 2006). Breaking prison rules also predicts recidivism (Cochran et al., 2014; French & Gendreau, 2006; Mooney & Daffern, 2015). Because prisons are charged with identifying factors that contribute to recidivism, there is a vested interest in potential proactive efforts to alleviate rule violations. Also, as nearly all of incarcerated people are eventually released (T. Hughes & Wilson, 2019), the impact of these approaches is felt by communities. Traditional approaches address rule violations during incarceration punitively and rely on imposing sanctions. Punishment could include loss of phone use or other “privileges.” Even though prohibiting people from using the phone may seem like serious punishment, there is little evidence to demonstrate the effectiveness of this or any other loss of privileges in reducing rule violations during incarceration (French & Gendreau, 2006). The lack of an effective method to address rule violations may stem from the lack of a clear understanding of their cause.
THEORETICAL RATIONALE
Historically, there have been two primary theories to explain prison rule violations—importation theory and deprivation theory. Importation theory proposes that incarcerated people bring individual characteristics with them to prison, which cause them to break rules. Deprivation theory describes rule violations as stemming from elements of the prison environment. However, neither can fully explain prison rule violations (Gover et al., 2000; Steiner et al., 2014; Worrall & Morris, 2011). This is likely because there are useful elements of each. In addition, there is variation among incarcerated people in the types of experiences that they might “import” and the levels of deprivation that they experience during incarceration, which would reduce the correlation. People do not arrive in prison as blank slates. As importation theory describes, people carry their previous characteristics, beliefs, and experiences, including those acquired during their current and previous incarcerations. These histories then shape people’s reactions to the institutional environment (Irwin & Cressey, 1962). Individual characteristics relating to mental health would be imported into the prison and have been demonstrated to contribute to rule violations. However, as noted by deprivation theory, environmental factors can impact rule violations. When faced with deprivation, such as limited access to resources, and a lack of prosocial alternatives to meet their needs, incarcerated people may break institutional rules to satisfy those needs (Sykes & Messinger, 1960).
Given the shortcomings of previous theoretical frameworks, a more useful approach may be to borrow from public health to employ ecosocial theory as a framework to understand rule violations driven by mental health and substance use disorders. Ecosocial theory explains how environmental factors can impact individual health, and how health then influences behavior, such as rule violations. The main proposition of ecosocial theory is that a person exists in the context of their environment and that experiences with their environment literally change their biology through a process called “embodiment” (Krieger, 2001). An example of embodiment would be how exposure to trauma could cause a mental health disorder.
Ecosocial theory also describes how environments exist in multiple levels (Krieger, 2001). As this theory includes multiple levels, it accounts for portions of both deprivation and importation theory. Like deprivation theory, ecosocial theory describes how environmental factors can influence individuals. However, in ecosocial theory, environments primarily influence individuals by impacting their health. In ecosocial theory, environmental factors may include economic and cultural systems. These factors then go on to influence people through their experiences with such things as diet, education, employment, community infrastructure, and housing. Adverse experiences relating to neglect and exposure to community violence occur at this level. Ecosocial theory also describes how environments exist on an interpersonal level (Krieger, 2001). It is on this level that adverse experiences relating to abuse occur. The theory also explains how environments can impact people interpersonally, for example, through cultural beliefs about spanking, which might result in higher levels of physical abuse.
Ecosocial theory’s individual level functions similarly to importation theory, in that it details how individual characteristics impact environmental interactions. In ecosocial theory, individual factors include things such as genetic makeup, biological conditions, gender, sex, and age—which are described as socially constructed (Krieger, 2001). Finally, ecosocial theory explains how elements outside of the individual literally influence their biology through embodiment (Krieger, 2001), for instance, through changes in brain development due to child abuse. Changes may later manifest as mental health or substance use disorders, which would impact behavior through symptoms. In sum, ecosocial theory more fully explains how past adverse experiences would lead to mental health, and substance use disorders, which would impact rule violations during incarceration. Use of ecosocial theory in this scenario is ideal, as it explains the recursive pathways by which environments impact individuals to change how individuals interact with their environments.
LITERATURE REVIEW
A 2014 literature review on rule violations during incarceration found evidence for 46 different contributing factors. Despite the volume of studies, few examined past experiences of abuse (Steiner et al., 2014) or “adverse experiences.” One study that did examine abuse history as a predictor of rule violations found a significant indirect effect at the prison level. This study reported a lower proportion of rule violations related to physical assault in prisons that held men with a low proportion of physical/sexual abuse history (Wooldredge & Steiner, 2015). However, physical/sexual abuse history was part of a latent variable that obscured the role that abuse played in the relationship. Another study examined past exposure to violence as a single construct and found that it predicted rule violations (Meade & Steiner, 2013). However, neither of these studies included nonviolent adverse experiences such as neglect.
A greater proportion of studies have documented how mental health and substance use disorders influence rule violations. Co-occurring mental health and substance use disorders predict assault during incarceration (Wood & Buttaro, 2013), whereas psychosis and major depressive disorder have been associated with rule violations in general (Felson et al., 2012; Stoliker, 2016). Other research on mental health and substance use disorders has identified child abuse and neglect, or “adverse childhood experiences,” as leading to these disorders (Dong et al., 2004; Felitti et al., 1998). Given these relationships, adverse experiences, mental health, and substance use disorders likely collectively associate with rule violations.
Another reason why examining these factors is important is that incarcerated people have disproportionately high rates of adverse experiences, mental health, and substance use disorders (Bowen et al., 2018). In one U.S. state, 50% of incarcerated young adults had more than four adverse childhood experiences (Baglivio et al., 2014), as compared with 5% of adults living in the community of a high-income country (Kessler et al., 2010). Having four or more adverse childhood experiences is considered high risk for developing mental health and substance use disorders (K. Hughes et al., 2017). The rate of psychosis among people incarcerated in high-income countries is 3.5%, whereas the rate of major depression is 10% (Fazel & Seewald, 2012). Among adults incarcerated in U.S. state prisons, 53% have a substance use disorder, whereas 26% had an alcohol use disorder (Marotta, 2017).
Previous research using a national sample of people incarcerated in the United States demonstrated that adverse experiences can be grouped into discrete latent classes (Henry, 2019). This study also found that, as compared with people in the low adversity typology, members of all other typologies were at significantly higher odds of having a mental health or substance use disorder (odds ratio [OR] = 1.57–3.73). In that study, 10 indicators of adverse experiences were used in latent class analysis to generate adversity typologies. Five indicators were associated with deprivation, whereas another five were associated with violence. Four typologies of adversity were determined: Class 1 = low exposure; Class 2 = moderate deprivation, high violence exposure; Class 3 = high deprivation, low violence exposure; and Class 4 = high exposure.
STUDY AIMS
I build on previous research by connecting the relationship between classes of prior adverse experiences to mental health and substance use disorders, as would be explained by the embodiment process from ecosocial theory. I use ecosocial theory to explain how mental health and substance use disorders influence rule violations by mediating the impact of adversity as a predictor of rule violations. To test this, I examine how individual-level diagnoses predict rule violations during incarceration using logistic regression and structural equation modeling and (a) estimate how classes or typologies of adversity predict rule violations, (b) estimate how mental health and substance use disorders predict rule violations, (c) identify mediating effects that mental health and substance use disorders have on the relationship between adverse experiences and rule violations, and (d) determine whether relationships differ by type of rule violation.
I tease apart the impact that adverse experiences and mental health/substance use disorders have on rule violations by accounting for the relationship between these two factors. Such testing examines whether adverse experiences impact rule violations through the disorders. This allows for a fuller understanding of which predictor is dominant, therefore, guiding an appropriate response. Understanding rule violations as a traumatic response stemming from adverse experiences may help correctional systems target trauma-informed programs or management strategies to prevent rule violations and promote rehabilitation. Answering these questions may inform policies that direct responses to rule violations to address the root cause, rather than respond in a way that exacerbates the underlying cause.
METHOD
DATA
Data came from the 2004 cross-sectional Survey of (adult) Inmates in State and Federal Correctional Facilities (SI-SFCF; U.S. Department of Justice, 2016). A combined state and federal sample was used. The survey used computer-assisted personal interviews conducted via telephone and two-stage clustered sampling. In Stage 1, prisons were randomly selected; in Stage 2, people were randomly selected from within prisons. The survey had a high response rate (89.1% for state prisons and 84.6% for federal prisons), a large sample (N = 18,185), and a low rate of missing data (<5% for included variables). If missing, survey administrators imputed Hispanic ethnicity, race, and age using hot decking. Hot decking imputes missing values based from a “donor” participant with similar covariates, using the smallest possible distance to indicate the best match (Myers, 2011). In this case, values of imputed variables were taken from the “nearest neighbor” within the same prison and sampling stratum (U.S. Department of Justice, 2016). Race was only imputed for people who did not identify as Hispanic to prevent the overestimation of people identifying as non-white Hispanic. In cases where both ethnicity and race were missing, Hispanic ethnicity was imputed first to determine whether race would also be imputed. If gender was missing, it was assigned based on facility of residence. Percent missing of imputed variables was negligible (<1%). No other variables were imputed. Transgender identity was not collected (U.S. Department of Justice, 2016).
INDEPENDENT VARIABLES
Four typologies of adverse experiences, together with categories for type and level of mental health and substance use disorders, were used as independent variables. Latent class analysis was used to identify typologies of adversity and started with a two-class model and systematically added additional classes while comparing fit with a bootstrap likelihood ratio test (Nylund et al., 2007) until the addition of more classes did not improve the model (Cloitre et al., 2014). See Table 1 for fit indices. Categories of mental health and substance use disorders included (a) alcohol use disorder, (b) substance use disorder, (c) mental health disorder, and (d) serious mental illness. Alcohol use disorder and substance use disorder pertained to having a diagnosis in the year prior to the current incarceration and were constructed via an index made from affirmative responses to questions related directly to the diagnostic criteria of alcohol and substance use disorders. Mental health disorder and serious mental illness pertained to ever being diagnosed. Mental health disorder encompassed all mental health diagnoses except schizophrenia and bipolar disorder, which were categorized as serious mental illnesses. Survey questions used to construct these variables included seven questions that asked, “Were you ever diagnosed with 1) bipolar disorder, 2) schizophrenia, 3) a depressive disorder, 4) posttraumatic stress disorder, 5) another anxiety disorder, 6) a personality disorder, or 7) any other mental condition?” Categories of mental health and substance use disorders were not mutually exclusive. Therefore, people categorized as having an alcohol use disorder may also have a substance use disorder. For more details, see Henry (2019).
TABLE 1:
Fit Indices for Latent Class Solutions
| Model | df | LL | BIC | AIC |
|---|---|---|---|---|
| One class | 10 | 62,589.7 | 125,277.5 | 125,199.5 |
| Two classes | 21 | 54,238.1 | 108,680.1 | 108,518.2 |
| Three classes | 32 | 53,791.1 | 107,892.8 | 107,646.1 |
| Four classes | 43 | 53,622.5 | 107,662.5 | 107,331.0 |
| Five classes | 54 | 53,589.9 | 107,704.2 | 107,287.9 |
| Six classes | 62 | 53,592.5 | 107,787.1 | 107,309.1 |
Note. A solution with four classes demonstrated the best fit according to the BIC index and did not conceptually or significantly differ from the five-class solution, which had a lower AIC index. LL = log-likelihood; BIC = Bayesian information criterion; AIC = Akaike information criterion.
DEPENDENT VARIABLES
Dependent variables used in analysis indicated if people were ever found guilty of, or written up, for institutional rule violations. Two types of rule violation variables were constructed for analysis. Each was based on self-report and indicated (a) “drug violation” or (b) “major violation.” The drug violation indicator was positive for incarcerated people responding “yes” to survey items regarding drug and alcohol violations. The major violation indicator was positive for people responding “yes” to survey items regarding possession of a weapon, verbal assault on staff, physical assault on staff, verbal assault on another incarcerated person, physical assault on another incarcerated person, escape/attempted escape, or “any other major violation.” Covariates included gender (male/female), race (white/non-white), age, sentence length, and time served to date. Covariates were selected based on significant prediction of rule violations during incarceration in previous studies (Steiner et al., 2014).
ANALYSIS
Analyses were conducted in Stata 15 (StataCorp, 2017). Bivariate (t test and correlation) analyses were used to obtain descriptive statistics. Logistic regression was used to predict (a) rule violations by adversity typology and mental health/substance use disorders and (b) mental health and substance use disorders by adversity typology. To limit a potential alternative source of variation, all models used the same set of covariates (gender, race, age, sentence length, and time served to date). Finally, generalized structural equation modeling (GSEM) was performed to simultaneously examine factors that contribute to rule violations and test for mediation. This procedure also checks for multicollinearity and drops collinear variables from the model. Maximum likelihood estimation was used, which employs equation-wise deletion instead of listwise deletion. This technique uses all available data rather than dropping cases with missing data. Indirect effects were calculated by multiplying the direct path of coefficients using Stata’s nonlinear combination of estimators’ function, which tests for significance and produces standard test statistics along with the coefficients (StataCorp, 2013).
RESULTS
DESCRIPTION OF THE SAMPLE
The sample included 18,185 individuals, who were predominantly men (78.6%), and largely identified as white (49.7%) or black (42.9%; Table 2). The most common adversity typology was low adversity (Class 1, 52.1%), followed by high deprivation/low violence (Class 3, 24.7%). Much smaller proportions of incarcerated people were classified as moderate deprivation/high violence (Class 2, 12.7%), or high exposure (Class 4, 10.5%). Self-reports of mental health and substance use disorders were common. Nearly half of incarcerated people had a substance use disorder in the year prior to their arrest (43.6%), while about a third had an alcohol use disorder (32.6%). About a quarter of incarcerated people had a mental health disorder (25.5%), and 12.6% had a serious mental illness. Nearly a quarter were found guilty or written up for a drug violation, major violation, or both during their current incarceration (22.5%). About a fifth of the sample (19.2%) had a major violation (including people who also had a drug violation), whereas 6.9% had a drug violation (including people who also had a major violation). Only a small proportion had both drug and major rule violations (3.6%).
TABLE 2:
Description of the Sample
| N | Measure of central tendancy | |
|---|---|---|
| Total | 18,185 | 100.0 (%) |
| Demographics | ||
| Gender | ||
| Male | 14,297 | 78.6 (%) |
| Female | 3,888 | 21.4 (%) |
| Race | ||
| White | 9,002 | 49.7 (%) |
| Black | 7,770 | 42.9 (%) |
| Native American/Alaskan | 995 | 5.5 (%) |
| Asian | 196 | 1.1 (%) |
| Hawaiian/Pacific Islander | 143 | 0.8 (%) |
| Other | 721 | 4.0 (%) |
| Ethnicity | ||
| Hispanic | 3,438 | 18.6 (%) |
| Age | 18,185 | 35.8 (mean) |
| Adversity typology | ||
| Class 1—low exposure/highest combat veteran | 9,474 | 52.1 (%) |
| Class 2—moderate deprivation/high violence | 2,309 | 12.7 (%) |
| Class 3—high deprivation/low violence | 4,492 | 24.7 (%) |
| Class 4—high exposure | 1,909 | 10.5 (%) |
| Behavioral health disorders | ||
| Alcohol use disorder | 4,837 | 32.6 (%) |
| Substance use disorder | 7,736 | 43.6 (%) |
| Mental health disorder | 4,548 | 25.5 (%) |
| Serious mental illness | 2,252 | 12.6 (%) |
| Covariates | ||
| Time served (years) | 16,682 | 4.5 (mean) |
| Sentence length (years)a | 17,627 | 36.3 (mean) |
Note. Missing <5%; missing not included in calculations.
People serving life sentences were coded as having the next longest sentence in years.
BIVARIATE STATISTICS
There were variations in the rate of rule violations by mental health and substance use disorders. People with an alcohol or substance use disorder had both types of violations at significantly (p < .001) higher rates than those without these disorders. People with a mental health disorder, or serious mental illness, had significantly higher rates of only major violations (Table 3). All mental health and substance use disorders were positively and significantly correlated. However, correlations were mostly weak (r ≤ .30), except for serious mental illness and other mental illness, which were moderately correlated (r = .52; data not shown in tables). However, during GSEM, no multicollinearity was identified.
TABLE 3:
Rule Violations by Type and Mental Health/Substance Use Disorders
| Mental Health/Substance Use Disorders | Drug (%) | Major (%) | ||
|---|---|---|---|---|
| All | 6.9 | 19.2 | ||
| Alcohol use disorder | 9.4 | * | 21.6 | * |
| No alcohol use disorder | 6.2 | 18.0 | ||
| Substance use disorder | 9.1 | * | 22.0 | * |
| No substance use disorder | 5.1 | 17.0 | ||
| Mental health disorder | 7.5 | 24.6 | * | |
| No mental health disorder | 6.6 | 17.3 | ||
| Serious mental illness | 7.1 | 25.3 | * | |
| No serious mental illness | 6.8 | 18.3 | ||
Note. N = 18,185; missing <5%.
p < .001.
MULTIVARIATE LOGISTIC MODELS
Multivariate logistic modeling established the relationship between adversity typologies and rule violation outcomes without the inclusion of potential mediating factors. While membership in all adversity classes predicted all types of rule violations, relative to the low adversity typology (Class 1), the magnitude of this relationship was strongest for Class 4 (high exposure), which had twice the odds of ever having a drug (OR = 2.00) or major violation (OR = 2.16), as compared with members of Class 1 (low adversity; Table 4).
Table 4:
Logistic Regression Models Predicting Rule Violations by Adversity and Mental Health/Substance Use Disorders
| Drug violations | Major violations | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| (N = 15,847) | (N = 13,008) | (N = 15,850) | (N = 13,008) | |||||||||
| Predictors | OR (SE) | Beta (SE) | p | OR (SE) | Beta (SE) | p | OR (SE) | Beta (SE) | p | OR (SE) | Beta (SE) | p |
| Adversity Typologies | ||||||||||||
| Class 2 | 1.41 (0.19) | 0.35 (0.14) | * | 1.27 (0.19) | 0.23 (0.15) | 1.43 (0.12) | 0.36 (0.08) | * | 1.20 (0.11) | 0.18 (0.09) | ||
| Class 3 | 1.67 (0.14) | 0.51 (0.08) | * | 1.49 (0.13) | 0.40 (0.09) | * | 1.51 (0.08) | 0.41 (0.05) | * | 1.40 (0.09) | 0.34 (0.06) | * |
| Class 4 | 2.00 (0.18) | 0.70 (0.09) | * | 1.65 (0.17) | 0.50 (0.10) | * | 2.16 (0.13) | 0.77 (0.06) | * | 1.81 (0.12) | 0.59 (0.07) | * |
| Mental Health/Substance Use Disorders | ||||||||||||
| AUD | Omitted | 1.20 (0.09) | 0.18 (0.08) | * | Omitted | 0.97 (0.05) | −0.03 (0.05) | |||||
| SUD | 1.83 (0.15) | 0.60 (0.08) | * | 1.34 (0.07) | 0.27 (0.05) | * | ||||||
| MHD | 1.15 (0.11) | 0.14 (0.10) | 1.66 (0.11) | 0.46 (0.06) | * | |||||||
| SMI | 0.96 (0.12) | −0.04 (0.13) | 1.16 (0.09) | 0.14 (0.07) | ||||||||
| Covariates | ||||||||||||
| Age | 0.96 (<0.01) | −0.04 (<0.01) | * | 0.96 (<0.01) | −0.04 (<0.01) | * | 0.94 (<0.01) | −0.06 (<0.01) | * | 0.94 (<0.01) | −0.06 (<0.01) | * |
| Time served | 1.01 (<0.01) | 0.01 (<0.01) | * | 1.01 (<0.01) | 0.01 (<0.01) | * | 1.01 (<0.01) | 0.01 (<0.01) | * | 1.01 (<0.01) | 0.01 (<0.01) | * |
| Sentence length | 1.00 (<0.01) | <0.01 (<0.01) | * | 1.00 (<0.01) | <0.01 (<0.01) | * | 1.00 (<0.01) | <0.01 (<0.01) | * | 1.00 (<0.01) | 0.01 (<0.01) | * |
| Male | 2.85 (0.35) | 1.05 (0.12) | * | 2.52 (0.33) | 0.93 (0.13) | * | 1.40 (0.09) | 0.34 (0.06) | * | 1.53 (0.11) | 0.43 (0.07) | * |
| White | 0.97 (0.07) | −0.04 (0.07) | 0.92 (0.07) | −0.08 (0.07) | 0.79 (0.03) | −0.24 (0.04) | * | 0.73 (0.04) | −0.32 (0.05) | * | ||
| Intercept | 0.03 (0.01) | −3.40 (0.18) | * | 0.03 (0.01) | −3.67 (0.21) | * | 0.53 (0.06) | −0.63 (0.10) | * | 0.42 (0.05) | −0.86 (0.12) | * |
Note. OR = odds ratio; SE = standard error; Class 2 = moderate deprivation, high violence exposure; Class 3 = high deprivation, low violence exposure; Class 4 = high exposure; AUD = alcohol use disorder; SUD = substance use disorder; MHD = mental health disorder; SMI = serious mental illness; reference group for adversity typologies was Class 1 = low exposure; reference group for male was female; reference group for white was non-white.
p < .05.
Logistic regression established the relationships between the primary predictors (adversity typologies), and the potential mediators (self-reported alcohol use disorder, substance use disorder, mental health disorder, and serious mental illness; Table 5). All mental health and substance use disorders were significantly associated with all adversity classes, as compared with the low adversity typology. However, once again, people with the high exposure typology (Class 4) had the highest odds of having any mental health or substance use disorder, relative to members of Class 1 (low adversity). Class 4 was a particularly strong predictor of mental health disorders, whereby members of this class had about 4 times the odds of having a mental health disorder (OR = 3.57), or serious mental illness (OR = 4.03), as compared with members of Class 1 (low exposure). Relative to Class 1, Classes 2 and 3 (moderate deprivation/low violence and high deprivation/low violence) also had higher odds of all disorders. However, for substance use disorders, Class 3 (high deprivation/low violence) had the second highest magnitude of impact, whereas for mental health disorders (including serious mental illness), Class 2 (moderate deprivation/high violence) had the second highest magnitude of impact.
Table 5:
Logistic Regression Models Predicting Mental Health/Substance Use Disorders by Adversity Typology
| Mental health and substance use disorders | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Alcohol use disorder (N = 13,275) |
Substance use disorder (N = 15,811) |
Mental health disorder (N = 15,895) |
Serious mental illness (N = 15,872) |
|||||||||
| Predictors | OR (SE) | Beta (SE) | p | OR (SE) | Beta (SE) | p | OR (SE) | Beta (SE) | p | OR (SE) | Beta (SE) | p |
| Class 2 | 1.60 (0.11) | 0.47 (0.07) | * | 1.73 (0.10) | 0.55 (0.06) | * | 3.40 (0.21) | 1.22 (0.06) | * | 3.34 (0.26) | 1.21 (0.08) | * |
| Class 3 | 1.66 (0.08) | 0.51 (0.05) | * | 1.84 (0.08) | 0.61 (0.04) | * | 1.61 (1.61) | 0.47 (0.05) | * | 1.70 (0.12) | 0.53 (0.07) | * |
| Class 4 | 1.85 (0.10) | 0.61 (0.05) | * | 2.18 (0.10) | 0.78 (0.05) | * | 3.57 (0.18) | 1.27 (0.05) | * | 4.03 (0.26) | 1.39 (0.06) | * |
| Age | 0.99 (<0.01) | −0.01 (<0.01) | * | 0.98 (<0.01) | −0.02 (<0.01) | * | 1.00 (<0.01) | <0.01 (<0.01) | 1.00 (<0.01) | <−0.01 (<0.01) | ||
| Time served | 1.00 (<0.01) | <0.01 (<0.01) | * | 1.00 (<0.01) | <−0.01 (<0.01) | * | 1.00 (<0.01) | <−0.01 (<0.01) | 1.00 (<0.01) | <−0.01 (<0.01) | ||
| Sentence length | 1.00 (<0.01) | <−0.01 (<0.01) | 1.00 (<0.01) | <−0.01 (<0.01) | * | 1.00 (<0.01) | <0.01 (<0.01) | 1.00 (<0.01) | <−0.01 (<0.01) | |||
| Male | 1.50 (0.08) | 0.41 (0.05) | * | 0.9 (0.04) | −0.11 (0.04) | * | 0.53 (0.02) | −0.64 (0.05) | * | 0.54 (0.03) | −0.62 (0.06) | * |
| White | 1.20 (0.05) | 0.18 (0.04) | * | 1.39 (0.05) | 0.33 (0.03) | * | 1.77 (0.07) | 0.57 (0.04) | * | 1.59 (0.08) | 0.46 (0.05) | * |
| Intercept | 0.32 (0.03) | −1.14 (0.09) | * | 1.20 (0.09) | 0.19 (0.07) | * | 0.25 (0.02) | −1.4 (0.09) | * | 0.11 (0.01) | −2.22 (0.11) | * |
Note. OR = odds ratio; SE = standard error; Class 2 = moderate deprivation, high violence exposure; Class 3 = high deprivation, low violence exposure; Class 4 = high exposure; reference group for adversity typologies was Class 1 = low exposure; reference group for male was female; reference group for white was non-white.
p < .05.
Final logistic regression models collectively analyzed relationships between adversity typologies, along with mediators (self-reported mental health and substance use disorders) to predict rule violation outcomes (Table 4). In this set of analyses, Class 2 (moderate deprivation/high violence) was no longer a significant predictor of either type of rule violation. Having an alcohol or substance use disorder significantly predicted drug violations, although having a mental health disorder or serious mental illness did not. People who had an alcohol or substance use disorder had about 1.5 times the odds of having a drug violation, as compared with those who did not have these disorders (OR = 1.20 and OR = 1.83). Major rule violations were significantly predicted by having a mental health or substance use disorder, but not by having an alcohol use disorder or serious mental illness. People with a mental health disorder had the highest odds of having a major rule violation (OR = 1.66), followed by those who had a substance use disorder (OR = 1.34). Nearly all covariates (age, time served, sentence length, male, and white) were significant predictors of both types of rule violations, although white race did not predict drug violations. However, the magnitude of these relationships was so small that they were not clinically meaningful predictors. One exception to this was gender; compared with women, men had 2.5 times the odds of a drug violation and 1.5 times the odds of a major violation.
Mediation analysis explains why Class 2 was no longer a significant predictor of rule violations, after accounting for potential mediators. Mental health and substance use disorders fully mediated the impact of Class 2 membership (moderate deprivation, high violence exposure) as a predictor of rule violations. In the case of Class 2, it was the adversity typologies that drove rule violations. Mental health and substance use disorders also partially mediated the relationship between Class 3 and 4 (high deprivation/low violence, and high exposure) and drug violations, meaning that some of the relationship between the disorders and rule violations is explained by the impact of adversity (Table 6; Figures 1 and 2).
TABLE 6:
Total, Direct, and Indirect Effects of Rule Violations
| Direct effect | Indirect effect | Total effect | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Paths | Beta | SE | p | Beta | SE | p | Beta | SE | p |
| Drug violation paths | |||||||||
| C2 - > DV | 0.23 | 0.15 | 0.84 | 0.22 | * | ||||
| C3 - > DV | 0.40 | 0.09 | * | 0.96 | 0.12 | * | |||
| C4 - > DV | 0.50 | 0.10 | * | 1.26 | 0.19 | * | |||
| C2 - > AUD - > DV | 0.06 | 0.03 | * | ||||||
| C2 - > SUD - > DV | 0.39 | 0.06 | * | ||||||
| C2 - > MHD - > DV | 0.22 | 0.15 | |||||||
| C2 - > SMI - > DV | −0.06 | 0.19 | |||||||
| C3 - > AUD - > DV | 0.10 | 0.04 | * | ||||||
| C3 - > SUD - > DV | 0.42 | 0.06 | * | ||||||
| C3 - > MHD - > DV | 0.06 | 0.04 | |||||||
| C3 - > SMI - > DV | −0.02 | 0.06 | |||||||
| C4 - > AUD - > DV | 0.11 | 0.05 | * | ||||||
| C4 - > SUD - > DV | 0.52 | 0.07 | * | ||||||
| C4 - > MHD - > DV | 0.19 | 0.14 | |||||||
| C4 - > SMI - > DV | −0.06 | 0.19 | |||||||
| Major violation paths | |||||||||
| C2 - > MV | 0.36 | 0.08 | 1.34 | 0.14 | * | ||||
| C3 - > MV | 0.41 | 0.05 | * | 0.82 | 0.08 | * | |||
| C4 - > MV | 0.77 | 0.06 | * | 1.74 | 0.12 | * | |||
| C2 - > AUD - > MV | −0.01 | 0.02 | |||||||
| C2 - > SUD - > MV | 0.19 | 0.04 | * | ||||||
| C2 - > MHD - > MV | 0.78 | 0.10 | * | ||||||
| C2 - > SMI - > MV | 0.22 | 0.12 | |||||||
| C3 - > AUD - > MV | −0.02 | 0.03 | |||||||
| C3 - > SUD - > MV | 0.21 | 0.04 | * | ||||||
| C3 - > MHD - > MV | 0.22 | 0.04 | * | ||||||
| C3 - > SMI - > MV | 0.07 | 0.04 | |||||||
| C4 - > AUD - > MV | −0.02 | 0.03 | |||||||
| C4 - > SUD - > MV | 0.25 | 0.05 | * | ||||||
| C4 - > MHD - > MV | 0.69 | 0.09 | * | ||||||
| C4 - > SMI - > MV | 0.22 | 0.12 | |||||||
Note. N = 18,182; SE = standard error; C2 = Class 2 (moderate deprivation/high violence); DV = drug violations; C3 = Class 3 (high deprivation/low violence); C4 = Class 4 (high exposure); AUD = alcohol use disorder; SUD = substance use disorder; MHD = mental health disorder; SMI = serious mental illness; MV = major violations.
p < .05.
Figure 1: Generalized Structural Equation Model Predicting Drug Violations.
Note. Numbers along the paths indicate the coefficient of the predictor; numbers inside the boxes indicate the intercept for the direct path to the variable; lines indicate significant paths (p < .05); nonsignificant paths, covariates, and error terms are not shown.
Figure 2: Generalized Structural Equation Model Predicting Major Violations.
Note. Numbers along the paths indicate the coefficient of the predictor; numbers inside the boxes indicate the intercept for the direct path to the variable; lines indicate significant paths (p < .05); nonsignificant paths, covariates, and error terms are not shown.
DISCUSSION
PREDICTORS OF RULE VIOLATIONS AND MENTAL HEALTH/SUBSTANCE USE DISORDERS
As compared with incarcerated people with low exposure to adversity, people with the most severe histories of adversity, marked by high exposure to violence and deprivation (members of Class 4) had the highest odds of having both types of rule violations. However, other typologies of adversity were also strong predictors of violations during incarceration. Incarcerated people with a history of high exposure to deprivation, but low exposure to violence (Class 3), also had higher odds of rule violations, as compared with the low adversity group. These relationships held true across both drug violations and major violations. Results suggest that past experiences of deprivation had a larger magnitude of impact on rule violations than past experiences of violent victimization. Evidence for this relationship comes from comparing the magnitude of Class 3 (high exposure to deprivation/low exposure to violence) as a predictor of rule violations, relative to Class 1, as compared with Class 2 (moderate deprivation/high violence), relative to Class 1 (OR = 1.49 vs. OR = 1.27 for drug violations and OR = 1.40 vs. OR = 1.20 for major violations). However, these experiences also added together (Class 4) to increase risk for mental health and substance use disorders. This pattern held true when examining the relationship between adversity classes and substance use disorders. However, the opposite relationship was observed for mental health disorders. Whereas findings generally built on previous research, which described how children who experienced neglect had relatively worse outcomes than children who were abused (Hildyard & Wolfe, 2002), future research should examine the differences in how abuse and neglect impact mental health versus substance use disorders.
My study also built on previous research that identified history of violence exposure (Meade & Steiner, 2013; Wooldredge & Steiner, 2015), and mental health/substance use disorders (Felson et al., 2012; Stoliker, 2016; Wood & Buttaro, 2013) as predictors of rule violations during incarceration. I extended these results by looking at both adverse violent experiences and experiences related to deprivation. In addition, I examined these adverse experiences as typologies and identified their impact on rule violations as they operated through mental health, and substance use disorders via mediation. My findings are supported by two previous studies that examined adversity typologies as predictors of incarceration. One study included a sample that was representative of the general U.S. population (Roos et al., 2016), whereas the other was focused on U.S. military veterans (Ross et al., 2018). Both studies found that as compared with the “low adversity” typologies, all other typologies significantly predicted incarceration. This research supports my study. However, they did not include mental health, and substance use disorders as mediators. My results add to the literature by examining the role that mental health and substance use disorders play in mediating the relationship between adversity typologies and rule violations during incarceration.
MENTAL HEALTH AND SUBSTANCE USE DISORDERS AS MEDIATORS
To my knowledge, I was the first to describe the mediating effects that mental health and substance use disorders have on the relationship between adverse experiences and rule violations. Although another study of women examined the effects on recidivism (Salisbury & Van Voorhis, 2009), I found that the impact of two adversity typologies (Class 3—high deprivation, low violence exposure and Class 4—high exposure) on rule violations was partially mediated by mental health and substance use disorders. Having a history of many adverse experiences (Class 4), or many deprivation-related experiences (Class 3), predicted both having a mental health or substance use disorder and rule violations. Full mediation was observed for Class 2 (moderate deprivation, high violence exposure) and both types of rule violations. Having moderate exposure to deprivation, but high exposure to violence (Class 2) predicted having mental health and substance use disorders, and on its own predicted rule violations. In the full model, the relationship between Class 2 and rule violations was fully accounted for by the relationship between specific mental health and substance use disorders and each rule violation.
The types of mental health and substance use disorders that predicted rule violations varied by violation type. Having an alcohol or substance use disorder predicted drug violations, likely due to continued use of alcohol and drugs during incarceration. For major rule violations, mental health disorders and substance use disorders were both predictors. These results provide evidence to support the link between substance use, mental health disorders, and rule violations (Felson et al., 2012; Stoliker, 2016; Wood & Buttaro, 2013). I built on this literature by connecting the role of past adversity. My findings are also supported by past research that documented how mental health and substance use disorders mediate the relationship between adversity and recidivism for women (Salisbury & Van Voorhis, 2009). I provided additional evidence that extends this research to rule violations during incarceration, across gender.
Serious mental illness was the only disorder that did not significantly predict any type of rule violation. This is surprising, given that this relationship has been documented in past studies (Felson et al., 2012; Stoliker, 2016). However, bivariate analysis showed that incarcerated people with a serious mental illness were more likely to have both types of rule violations. These unexpected results can be explained as a by-product of the relationship between adverse experiences and serious mental illness. More substantial adversity typologies very strongly predicted having a serious mental illness, even more strongly than they predicted other disorders. These same typologies of adversity also very strongly predicted rule violations. Therefore, it is the adverse experiences that are the driving force behind both the serious mental illness and the rule violations. Given that previous studies have not included this full pathway, it is understandable that the key variable in this sequence was overlooked.
My finding that covariates (age, time served, sentence length, and white race) were either insignificant or not clinically meaningful predictors of rule violations also deviates from previous research (Steiner et al., 2014). This is a potentially important contribution to the literature in that it may suggest that what is driving previously detected meaningful differences in these covariates is actually underlying differences in exposure to adversity. However, additional research is needed to better understand if this is the case. Consistent with other studies (Celinska & Sung, 2014), I found that men were significantly more likely to have rule violations than women, although why this is true is still unclear. Previous studies also struggle to explain this difference (Steiner & Wooldredge, 2014). Future research should explore how the other relationships identified in my study may vary by gender.
THEORETICAL IMPLICATIONS
My results are in line with ecosocial theory, which suggests that adverse experiences would be physically embodied as mental health and substance use disorders, although it is important to highlight that I did not examine physical biomarkers or other physiologic data that would be necessary to verify this relationship. However, other studies have documented this link (Toyokawa et al., 2012), which supports my use of this theory to conceptualize these relationships. Therefore, use of this theory is supported by my findings of full and partial mediation in the case of the relationships between adversity typologies and rule violations. Epigenetic research, which tests ecosocial theory, could help to explain differences between the strength of the associations I found between the different classes and rule violations. Epigenetics explores how social environments impact gene expression. Whereas research in this area has not yet identified the physiological differences in the effects of exposure to violence as compared with deprivation, it is likely that there are differences in the physiological impact of these experiences. Such differences could account for the variations I observed in the strength of mediation between classes where exposure to deprivation appears to have a different magnitude of impact as compared with exposure to violence. A potential exception is the case of serious mental illness, where it was adversity that drove rule violations, and not the disorder. Future research on the relationship between serious mental illness and adversity is needed to better understand how past adversity may explain symptoms and related behavior.
IMPLICATIONS FOR CORRECTIONAL PRACTICE
Results provide evidence for how shifting away from traditionally punitive ways of addressing rule violations could be a useful strategy. Addressing adverse experiences of incarcerated people might also prevent rule violations. Practitioners can apply findings to both develop a better understanding of the people they serve, and to design environments that are enriching and support social connection, rather than further exposure to deprivation and violence. Clinicians can use information about adversity typologies during assessment and treatment planning. In doing this, it may be useful to consider typologies as existing on overall and categorical continuums where both the overall severity of adversity exposure and exposure to categories of deprivation/violence are important.
Regarding environmental design, the use of trauma-informed correctional care could be useful for applying conclusions to correctional practice. Trauma-informed correctional care consists of both trauma-informed correctional management strategies, as well as trauma-informed behavioral health treatments. The high rates of ongoing exposure to trauma within correctional systems have been linked to the correctional culture itself (Penney, 2013). Cultural artifacts such as move teams, strip searches, and restricted movement can trigger traumatic reactions, particularly for people who have histories of experiencing adversity (Owen et al., 2008). Management strategies of trauma-informed correctional care aim to shift the organizational culture of the prison from an inherently traumatizing environment to one that does no harm. This is accomplished through the incorporation of training, screening, and treatment for trauma, in addition to structural design and processes that reduce the potential for traumatization (Miller & Najavits, 2012). If trauma-informed correctional care can prevent rule violations during incarceration, then there is also potential for cost-effectiveness, as rule violations can lead to placements in higher security settings that are more expensive.
This study highlights the importance of addressing mental health and substance use disorders in tandem with adverse experiences for incarcerated people. As previously discussed, punitive approaches to rule violations during incarceration have little evidence to demonstrate their effectiveness (French & Gendreau, 2006). Evidence-based trauma-informed mental health/substance use disorder treatments have been developed and tested with nonincarcerated populations. However, these interventions have only recently been studied with incarcerated patients (King, 2017). The interventions that have been tested with incarcerated patients were summarized in a 2015 literature review and are currently limited to manualized group therapy based primarily on cognitive behavioral therapy and mindfulness approaches (King, 2017). Continued research is needed to identify and test additional models of evidence-based mental health and substance use disorder treatment for use in prisons.
In addition to expanding the use of evidence-based interventions, an expansion of policies and programs that promote social support among incarcerated people could also be useful. Research has documented how social support can mediate the relationship between past experiences of adversity and quality of life among incarcerated men (Skarupski et al., 2016). Effective in-prison strategies for increasing social support include peer support services (Bagnall et al., 2015) and group religious activities (Kerley & Copes, 2009). Policies that promote social support are also likely linked to reductions in rule violations because frequent contact with family and friends during incarceration is linked to fewer rule violations (Solinas-Saunders & Stacer, 2012). Given these results, adopting such strategies to increase social support within prisons could serve to disrupt the relationship between adverse experiences and mental health/substance use disorders and potentially prevent rule violations.
LIMITATIONS
While my study has many strengths, it is not without limitations. As the public use data files were used in this analysis, information was masked on which prison housed which people. This prevented analysis on the effects of prison characteristics. Another limitation is variables relating to substance use and mental health diagnoses pertained either to the year prior to incarceration, in the case of substance use disorders, or ever, in the case of mental health diagnoses. As timing of these diagnoses might be distant from the rule violation, the relationship between these two experiences is not necessarily direct. However, the impact of this limitation would likely reduce the strength of any identified associations. Therefore, the strength of the identified associations may be understated, although there may also be other important factors that were not identified in analysis. Future studies should seek to replicate these findings.
CONCLUSION
Evidence links history of adverse experiences to rule violations during incarceration and shows that mental health and substance use disorders mediate this relationship. While more severe adversity typologies (Classes 2, 3, and 4) predict mental health and substance use disorders, the primary driver of drug violations is alcohol and substance use disorders. For major violations, the main predictor is mental health and substance use disorders. However, for people with serious mental illnesses, adverse experiences drive rule violations. Evidence can be used to develop strategies that address rule violations through treating adverse experiences, mental health, and substance use disorders during incarceration. Promoting policies and programs that directly counter the underlying prison experiences of deprivation and violence could also be useful in preventing future rule violations.
Acknowledgments
I would like to thank Deborah Garnick, Joanne Nicholson, William Fisher, and Grant Ritter for mentorship and assistance with reviewing early drafts of this manuscript. Research was supported by the National Institute on Drug Abuse (Award Number T32DA037801), and the National Institute on Alcohol Abuse and Alcoholism (Award Number T32AA007567). Content is the author’s sole responsibility and does not necessarily represent official views of the National Institutes of Health. The paper was presented at the Council on Social Work Education’s 65th Annual Program Meeting in Denver, CO, October 2019.
Biography
Brandy F. Henry, PhD, LICSW, is a clinician scientist who uses her years of practice experience to inform her research, which aims to improve the health of criminalized populations. She is currently a postdoctoral research fellow at the Columbia University School of Social Work.
REFERENCES
- Baglivio MT, Epps N, Swartz K, Huq MS, Sheer A, & Hardt NS (2014). The prevalence of adverse childhood experiences (ACE) in the lives of juvenile offenders. Journal of Juvenile Justice, 3(2), 1–23. [Google Scholar]
- Bagnall A-M, South J, Hulme C, Woodall J, Vinall-Collier K, Raine G, … Wright NMJ (2015). A systematic review of the effectiveness and cost-effectiveness of peer education and peer support in prisons. BMC Public Health, 15(1), Article 290. 10.1186/s12889-015-1584-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bowen K, Jarrett M, Stahl D, Forrester A, & Valmaggia L (2018). The relationship between exposure to adverse life events in childhood and adolescent years and subsequent adult psychopathology in 49,163 adult prisoners: A systematic review. Personality and Individual Differences, 131(1), 74–92. [Google Scholar]
- Celinska K, & Sung H-E (2014). Gender differences in the determinants of prison rule violations. The Prison Journal, 94(2), 220–241. [Google Scholar]
- Cloitre M, Garvert DW, Weiss B, Carlson EB, & Bryant RA (2014). Distinguishing PTSD, complex PTSD, and borderline personality disorder: A latent class analysis. European Journal of Psychotraumatology, 5(1), 1–10. 10.3402/ejpt.v5.25097 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cochran JC, Mears DP, Bales WD, & Stewart EA (2014). Does inmate behavior affect post-release offending? Investigating the misconduct-recidivism relationship among youth and adults. Justice Quarterly, 31(6), 1044–1073. [Google Scholar]
- Dong M, Anda RF, Felitti VJ, Dube SR, Williamson DF, Thompson TJ, … Giles WH (2004). The interrelatedness of multiple forms of childhood abuse, neglect, and household dysfunction. Child Abuse & Neglect, 28(7), 771–784. 10.1016/j.chiabu.2004.01.008 [DOI] [PubMed] [Google Scholar]
- Fazel S, & Seewald K (2012). Severe mental illness in 33,588 prisoners worldwide: Systematic review and meta-regression analysis. The British Journal of Psychiatry, 200(5), 364–373. [DOI] [PubMed] [Google Scholar]
- Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, … Marks JS (1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study. American Journal of Preventive Medicine, 14(4), 245–258. [DOI] [PubMed] [Google Scholar]
- Felson RB, Silver E, & Remster B (2012). Mental disorder and offending in prison. Criminal Justice and Behavior, 39(2), 125–143. 10.1177/0093854811428565 [DOI] [Google Scholar]
- French SA, & Gendreau P (2006). Reducing prison misconducts: What works!. Criminal Justice and Behavior, 33(2), 185–218. [Google Scholar]
- Gover AR, Mackenzie DL, & Armstrong GS (2000). Importation and deprivation explanations of juveniles’ adjustment to correctional facilities. International Journal of Offender Therapy and Comparative Criminology, 44(4), 450–467. 10.1177/0306624X00444004 [DOI] [Google Scholar]
- Henry BF (2019). Typologies of adversity in childhood & adulthood as determinants of mental health & substance use disorders of adults incarcerated in US prisons. Child Abuse & Neglect, 99, Article 104251. 10.1016/j.chiabu.2019.104251 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hildyard KL, & Wolfe DA (2002). Child neglect: Developmental issues and outcomes. Child Abuse & Neglect, 26(6–7), 679–695. [DOI] [PubMed] [Google Scholar]
- Hughes K, Bellis MA, Hardcastle KA, Sethi D, Butchart A, Mikton C, … Dunne MP (2017). The effect of multiple adverse childhood experiences on health: A systematic review and meta-analysis. Lancet Public Health, 2(8), e356–e366. 10.1016/s2468-2667(17)30118-4 [DOI] [PubMed] [Google Scholar]
- Hughes T, & Wilson DJ (2019). Reentry trends in the United States: Inmates returning to the community after serving time in prison. Bureau of Justice Statistics. https://www.bjs.gov/content/reentry/reentry.cfm [Google Scholar]
- Irwin J, & Cressey DR (1962). Thieves, convicts and the inmate culture. Social Problems, 10(2), 142–155. [Google Scholar]
- Kerley KR, & Copes H (2009). “Keepin’ my mind right”: Identity maintenance and religious social support in the prison context. International Journal of Offender Therapy and Comparative Criminology, 53(2), 228–244. [DOI] [PubMed] [Google Scholar]
- Kessler RC, McLaughlin KA, Green JG, Gruber MJ, Sampson NA, Zaslavsky AM, … Williams DR (2010). Childhood adversities and adult psychopathology in the WHO World Mental Health Surveys. The British Journal of Psychiatry, 197(5), 378–385. 10.1192/bjp.bp.110.080499 [DOI] [PMC free article] [PubMed] [Google Scholar]
- King EA (2017). Outcomes of trauma-informed interventions for incarcerated women: A review. International Journal of Offender Therapy and Comparative Criminology, 61(6), 667–688. [DOI] [PubMed] [Google Scholar]
- Krieger N (2001). Theories for social epidemiology in the 21st century: An ecosocial perspective. International Journal of Epidemiology, 30(4), 668–677. [DOI] [PubMed] [Google Scholar]
- Marotta PL (2017). Childhood adversities and substance misuse among the incarcerated: Implications for treatment and practice in correctional settings. Substance Use & Misuse, 52(6), 717–733. 10.1080/10826084.2016.1261899 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Meade B, & Steiner B (2013). The effects of exposure to violence on inmate maladjustment. Criminal Justice and Behavior, 40(11), 1228–1249. [Google Scholar]
- Miller NA, & Najavits LM (2012). Creating trauma-informed correctional care: A balance of goals and environment. European Journal of Psychotraumatology, 3(1), 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mooney JL, & Daffern M (2015). The relationship between aggressive behaviour in prison and violent offending following release. Psychology, Crime & Law, 21(4), 314–329. 10.1080/1068316X.2014.989163 [DOI] [Google Scholar]
- Myers TA (2011). Goodbye, listwise deletion: Presenting hot deck imputation as an easy and effective tool for handling missing data. Communication Methods and Measures, 5(4), 297–310. [Google Scholar]
- Nylund KL, Asparouhov T, & Muthén BO (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling, 14(4), 535–569. [Google Scholar]
- Owen B, Wells J, Pollock J, Muscat B, & Torres S (2008). Gendered violence and safety: A contextual approach to improving security in women’s facilities. U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. [Google Scholar]
- Penney D (2013). Creating a place of healing and forgiveness: The trauma-informed care initiative at the Women’s Community Correctional Center of Hawaii. National Center for Trauma-Informed Care; https://www.nasmhpd.org/sites/default/files/7014_hawaiian_trauma_brief_2013(1).pdf [Google Scholar]
- Roos LE, Afifi TO, Martin CG, Pietrzak RH, Tsai J, & Sareen J (2016). Linking typologies of childhood adversity to adult incarceration: Findings from a nationally representative sample. American Journal of Orthopsychiatry, 86(5), 584–593. [DOI] [PubMed] [Google Scholar]
- Ross J, Waterhouse-Bradley B, Contractor AA, & Armour C (2018). Typologies of adverse childhood experiences and their relationship to incarceration in U.S. military veterans. Child Abuse & Neglect, 79, 74–84. 10.1016/j.chiabu.2018.01.023 [DOI] [PubMed] [Google Scholar]
- Salisbury EJ, & Van Voorhis P (2009). Gendered pathways: A quantitative investigation of women probationers’ paths to incarceration. Criminal Justice and Behavior, 36(6), 541–566. [Google Scholar]
- Skarupski KA, Parisi JM, Thorpe R, Tanner E, & Gross D (2016). The association of adverse childhood experiences with mid-life depressive symptoms and quality of life among incarcerated males: Exploring multiple mediation. Aging & Mental Health, 20(6), 655–666. [DOI] [PubMed] [Google Scholar]
- Solinas-Saunders M, & Stacer MJ (2012). Prison resources and physical/verbal assault in prison: A comparison of male and female inmates. Victims & Offenders, 7(3), 279–311. [Google Scholar]
- StataCorp. (2013). Stata structural equation modeling reference manual: Release 13. Stata Press. [Google Scholar]
- StataCorp. (2017). Stata Statistical Software: Release 15.
- Steiner B, Butler HD, & Ellison JM (2014). Causes and correlates of prison inmate misconduct: A systematic review of the evidence. Journal of Criminal Justice, 42(6), 462–470. [Google Scholar]
- Steiner B, & Wooldredge J (2014). Sex differences in the predictors of prisoner misconduct. Criminal Justice and Behavior, 41(4), 433–452. [Google Scholar]
- Stoliker BE (2016). Inmate mental health predicting the likelihood of physical and verbal assault on correctional staff. Journal for Social Thought, 1(1), 1–16. [Google Scholar]
- Sykes GM, & Messinger SL (1960). The inmate social system In Cloward RA, Cressey DR, Grosser GH, McCleery R, Ohlin LE, Sykes GM, & Messinger S (Eds.), Theoretical studies in social organization of the prison (pp. 5–19). Social Science Research Council. [Google Scholar]
- Toyokawa S, Uddin M, Koenen KC, & Galea S (2012). How does the social environment “get into the mind”? Epigenetics at the intersection of social and psychiatric epidemiology. Social Science & Medicine, 74(1), 67–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- U.S. Department of Justice. (2016). Survey of Inmates in State and Federal Correctional Facilities (ICPSR04572-v2). Inter-university Consortium for Political and Social Research; 10.3886/ICPSR04572.v2 [DOI] [Google Scholar]
- Wood SR, & Buttaro A (2013). Co-occurring severe mental illnesses and substance abuse disorders as predictors of state prison inmate assaults. Crime & Delinquency, 59(4), 510–535. 10.1177/0011128712470318 [DOI] [Google Scholar]
- Wooldredge J, & Steiner B (2015). A macro-level perspective on prison inmate deviance. Punishment & Society, 17(2), 230–257. [Google Scholar]
- Worrall JL, & Morris RG (2011). Inmate custody levels and prison rule violations. The Prison Journal, 91(2), 131–157. 10.1177/0032885511404380 [DOI] [Google Scholar]