Substance Use Pathways Among Female Adolescent Offenders

Abstract

This study tested a general strain theory (GST) model across two times with an ethnically diverse sample of 184 female adolescent offenders. We tested whether exposure to violence was associated with girls’ emotions and substance use. We found that (a) hostility mediated the relationship between experiencing violence and alcohol use at Time 1; (b) hostility mediated the relationship between witnessing violence and alcohol use at Time 1; and (c) depressive symptoms mediated the relationship between experiencing violence and marijuana use at Time 2. These mediated effects were minimal in magnitude given that that the bivariate relationships between violence exposure and substance use were small or nonsignificant. Exposure to violence was associated with negative emotions across both models.

Keywords

general strain theory pathways theory female adolescent offenders substance use exposure to violence

Stressful and traumatic events in childhood, collectively known as adverse child experiences (ACEs), have been linked to future substance abuse problems (Dube et al., 2006; Felitti et al., 1998; Rothman, Edwards, Heeren, & Hingson, 2008; Sharp, Peck, & Hartsfield, 2012). In a landmark study on the effects of adverse child experiences on later health and well-being, researchers surveyed 9,508 adults about seven different types of childhood trauma and household dysfunction exposures: physical abuse, sexual abuse, emotional abuse, mother treated violently, substance misuse within the household, household mental illness, and having an incarcerated household member. They found that ACEs were strongly related to adverse health problems in adulthood, including those associated with substance abuse (Felitti et al., 1998). Certain adverse child experiences, most notably childhood victimization and witnessing violence, appear to be even stronger predictors of later substance use and abuse than other ACEs (Dube et al., 2006; Rothman et al., 2008; Sharp et al., 2012).

Understanding how childhood victimization and witnessing violence, known collectively as exposure to violence, is linked to substance use problems among high risk individuals such as juvenile offenders is critical given that rates of victimization and witnessing violence are higher among juvenile offenders relative to members of the general population (Baglivio & Epps, 2016; Baglivio et al., 2014). In a population of 64,329 juvenile offenders, 29.5% reported physical abuse, 11.9% reported sexual abuse, and 81.9% reported family violence (Baglivio & Epps, 2016). Among this same population, female adolescent offenders reported higher prevalence rates than males of physical abuse (males: 26%, females: 41%), sexual abuse (males: 7%, females: 31%), and family violence (males: 81%, females: 84%; Baglivio et al., 2014).

Despite the high rates of exposure to violence (Abram et al., 2004; Baglivio et al., 2014; Foy, Ritchie, & Conway, 2012) and substance use (Gehring, 2018; McClelland, Teplin, & Abram, 2004) among female adolescent offenders, scant attention has focused on how experiencing direct and indirect forms of victimization can lead to substance use among this high-risk population. This study seeks to address this void by testing a model based on general strain theory (GST) to determine whether experiencing and witnessing violence are associated with an increased risk of alcohol and marijuana use among female adolescent offenders; and second, to test whether these relationships are mediated by negative emotions (i.e., depressive symptoms, hostility).

Conceptual Framework

GST provides the framework for the current study (Agnew, 1992, 2006; Broidy & Agnew, 1997). This social psychological theory posits that strains or stressors such as ACEs result in negative emotions such as anger, hostility, sadness, and anxiety. To cope with these negative emotions, individuals engage in illicit behaviors if other coping strategies have either not been successful or are unavailable. GST also predicts that girls and women (compared with boys and men) experience different types of, generate different responses to, and develop different coping mechanisms to deal with strain (Broidy & Agnew, 1997). Indeed, several studies have examined these propositions on gender, strain, and crime (Jang, 2007; Piquero & Sealock, 2004). Jang (2007), for example, found empirical support that males and females experience strain differently, with women in the study more likely to generate self-directed (depression, anxiety) rather than other-directed (anger) emotions. Although GST suggests that girls and women may be less likely than their male counterparts to respond to strain with violence, delinquency, and crime and more likely to respond to strain with either legitimate (i.e., exercise) or self-directed coping mechanisms (i.e., substance use; Broidy & Agnew, 1997), research supporting this proposition remains mixed (see Piquero & Sealock, 2004, for review).

According to feminist pathways perspectives, personal victimization is a major strain in the lives of young women who engage in crime (to name a few, Belknap, 2014; Brennan, Breitenbach, Dieterich, Salisbury, & Van Voorhis, 2012; Chesney-Lind & Shelden, 2013; Harlow, 1999). Indeed, research indicates that females are more likely than males to experience certain types of victimization such as sexual abuse, sexual assaults, and rape (sexual abuse: Putnam, 2003; rape and sexual assaults: Tjaden & Thoennes, 2006). Because girls are frequently socialized and expected to be relationship-oriented and put others’ needs before their own (Miller, Winn, Taylor, & Wiki, 2012), they are also more likely than boys to demonstrate significantly higher rates of internalizing symptoms (i.e., depression, anxiety; Hanson et al., 2008; Merikangas et al., 2010).

In particular, sexual abuse is a gendered strain that affects girls’ ability to form attachment bonds needed to deter delinquency and anti-social behavior; it increases the likelihood of depression, suicide attempts, self-injury, aggression/violence, eating disorders, association with delinquent peer groups, and risky lifestyles (Bailey & McCloskey, 2005; Grant et al., 2004; McKnight & Loper, 2002; Riehman, Blutenthal, Juvonen, & Morral, 2003; Simkins & Katz, 2002; Ullman, 2004). Sexually abused girls are 2 to 3 times more likely to develop drug problems than their nonabused counterparts (Bergen, Martin, Richardson, Allison, & Roeger, 2004). While girls also experience anger in response to strains such as sexual victimization, their anger is “more likely than that of boys to be accompanied by depression as well as guilt and anxiety, largely because of how girls are socialized to internalize rather than externalize blame” (Goodkind, Ruffolo, Bybee, & Sarri, 2009, p. 102).

Witnessing violence also represents a salient strain for adolescents, but most of these studies have examined the relationship between strain and delinquency (Agnew, 2002, 2006; Craig, Cardwell, & Piquero, 2017; Lin, Cochran, & Mieczkowski, 2011; Pasko & Chesney-Lind, 2012). Few studies have examined how witnessing violence is related to emotional coping responses and substance use, particularly among girls. The current study adds to this limited research by examining this relationship among girls involved in the juvenile justice system. Understanding the impact of witnessing violence on system-involved girls is critical given these girls not only experience high rates of personal victimization, but often witness violence in their homes, schools, and communities (Abram et al., 2004; Baglivio et al., 2014; Berenson, Wiemann, & McCombs, 2001; Foy et al., 2012).

Exposure to Violence and Substance Use

Studies have generally found a significant association between experiencing violence and substance use, and this is true whether the violence is experienced in the home (child abuse, sexual abuse: Simpson & Miller, 2002), within intimate partner relationships (Carbone-López, Kruttschnitt, & Macmillan, 2006), or at school and in the community (bullying: Radliff, Wheaton, Robinson, & Morris, 2012; school and community violence: Wright, Fagan, & Pinchevsky, 2013). Dube et al. (2006) found that adults who reported physical abuse were 2.2 times more likely to have initiated alcohol use before age 14; those who reported sexual abuse were 2.8 times more likely to initiate alcohol use before age 18; and those who had a battered mother were 2 times more likely to initiate alcohol use before the age of 18. Based on a national probability sample of current and former drinkers, Rothman et al. (2008) found that five ACEs—including sexual and physical abuse—were associated with early initiation of alcohol use before the age of 14 even after controlling for confounds. In another study comparing the cumulative effects of ACEs with the effects of individual ACEs, Sharp et al. (2012) found that certain ACEs, including physical abuse and having a battered mother, were both directly and indirectly associated with increased substance use among a sample of incarcerated women. The relationship between experiencing violence and substance use appears to be especially pronounced for girls as indicated by a meta-analysis, which found that victimization (physical and sexual abuse) was associated with increased substance use among girls, but not boys (Simpson & Miller, 2002). In another study of high-risk youth, adolescent girls who experienced violence were 2 to 3 times as likely as those who had neither witnessed nor experienced violence to report alcohol and substance use (Berenson et al., 2001).

Witnessing violence also appears to be associated with increased substance use among adolescents (Berenson et al., 2001; Lin et al., 2011; Pinchevsky, Fagan, & Wright, 2014; Vermeiren, Schwab-Stone, Deboutte, Leckman, & Ruchkin, 2003; Wright et al., 2013; Zinzow et al., 2009). Although we were only able to identify one study specific to high-risk girls, this study found that girls who witnessed but did not experience violence reported marijuana use in the past 12 months more often than girls who had neither witnessed nor experienced violence (Berenson et al., 2001). Clearly, more research is needed to further investigate the relationship between witnessing violence and substance use among girls.

Exposure to Violence, Negative Emotions, and Substance Use

Only a handful of studies have examined the relationships among exposure to violence, negative emotions, and substance use. In their study of 139 women who had been exposed to multiple forms of gender-based violence, Rees et al. (2011) found that nine out of 10 had mental health disorders, with nearly half reporting substance abuse. In addition, Francis (2014) found that depression/anxiety influenced substance use among youth, but the type of effect varied by gender. For girls, depression/anxiety was directly associated with increased substance use, but only acted as a moderator of the exposure to violence–substance use relationship for boys. Kaufman (2009), using the Add Health dataset, found that depressive symptoms mediated the relationship between violent victimization and weekly drinking for girls, but not boys. Only one study to our knowledge has examined the relationship between witnessing violence, depressive symptoms, and substance use among adolescents. This study, which was based on a national probability sample of adolescents, found that “vicarious victimization” was directly related to drug use, and this relationship was mediated by depression (Lin et al., 2011).

Moreover, very few studies have examined the relationship between anger and substance use, especially among female adolescent offenders. One study with incarcerated youth found a positive association between anger expression and marijuana and alcohol use, but this study was specific to boys (Eftekhari, Turner, & Larimer, 2004). In a study of multiethnic youth, Nichols, Mahadeo, Bryant, and Botvin (2008) found that anger was associated with drinking alcohol, smoking cigarettes, and using marijuana, with affective factors playing a greater role in girls’ drug use. Espelage, Low, Rao, Hong, and Little (2014) examined the relationships among family violence, bullying, fighting, and substance use among 1,232 middle school students from the Midwest. For boys, fighting and bullying perpetration mediated the relationship between family violence and substance use. For girls, there was a direct relationship between exposure to family violence and substance use over time, but fighting and bullying did not mediate this relationship. Comparing incarcerated women who were addicted to drugs prior to incarceration to those who were not, Sharp et al. (2012) found that strain was indirectly related to substance abuse through anger expression. Physical abuse and having lived with someone who used alcohol or drugs were specific strains most likely predicting daily substance abuse. In addition, sexual abuse was directly related to daily drug use but not alcohol use.

Study Purpose

The current study expands upon the extant research in several ways. First, this study examined the direct relationships between two different types of exposure to violence (personally experienced and witnessing violence) and substance use among a sample of serious female adolescent offenders (see Figure 1). To date, most GST studies have not distinguished between different types of violence exposure nor have they considered substance use as an outcome. Second, this study examined both depressive symptoms and hostility as possible mediators of the exposure to violence—substance use relationship. We hypothesized that these negative emotions (depressive symptoms and hostility) would mediate the relationships between exposure to violence (witnessing violence and experiencing violence) and alcohol and marijuana use. We focused on alcohol and marijuana use outcomes because both result in decreased anxiety and increased feelings of euphoria, which can ameliorate negative feelings such as anger and sadness that are often associated with exposure to violence and both marijuana and alcohol are more socially accepted than other drugs, and therefore, more readily available to adolescents. Finally, most studies testing GST among adolescents have focused on school or community samples and have been cross-sectional. This study is longitudinal in nature and tested major tenets of GST among a sample of serious female offenders.

Figure 1.

Conceptual model predicting substance use (control variables not pictured).

Method

Data

Data for the current study were drawn from the Pathways to Desistance (PTD) project, a multisite effort to examine changes in criminal activity among serious adolescent offenders (Mulvey, 2013). This study followed over 1,000 adolescent offenders in Maricopa County, Arizona, and Philadelphia County, Pennsylvania, from 2000 to 2010. The PTD project was created to examine the relative impact of juvenile justice practices and life changes among serious adolescent offenders. The project recruited 1,345 adjudicated adolescents between the ages of 14 and 17 at the time of committing their offense. Youth were deemed eligible through a process involving review of court files, which determine they had been found guilty of a serious criminal offense. Adolescents who committed any felony offense were eligible, but those who were adjudicated for less serious property crimes, misdemeanor weapons offenses, and misdemeanor sexual assault were excluded. In addition, male offenders adjudicated for drug offenses were limited to 15% of the sample to maintain heterogeneity, while all female offenders meeting the age and offense requirements remained eligible, even if they were charged with a drug offense. See Mulvey (2004) and Schubert et al. (2004) for detailed information about the methodology and data collection procedures used by the researchers who supervised the data collection process. The data were deposited for public access as part of the National Addiction & HIV Data Archive Program (NAHDAP), a collaborative project sponsored by the University of Michigan’s Inter-University Consortium for Police and Social Science Research (ICPSR), the National Institute on Drug Abuse (NIDA), and the National Institutes of Health Office of Behavioral and Social Sciences Research.

The longitudinal research design included a series of interviews. The baseline interview covered six domains, including (a) demographic background characteristics; (b) individual factors such as substance use, mental health, and school performance; (c) psychosocial development and attitudes; (d) family context; (e) personal relationships; and (f) community context. Interviews were separated into multiple 2-hr sessions based on the extensive nature of the information collected. Follow-up interviews were conducted in 6-month intervals from baseline through 18 months. This interval was stretched to 12 months beginning at the 24-month follow-up and concluded with a 7-year follow-up interview.

The sample for the current analyses included all female adolescent offenders who participated in the baseline (T₁) and 6-month (T₂) follow-up interviews. The 184 female participants ranged in age from 14 to 18 years, with a mean of 15.99 (SD = 1.03). The largest (37%) racial group was comprised of Black participants, followed in sequence by Latina (30%) participants, White (27%) participants, and participants (6%) who identified as a different race or ethnicity. Participants came from a wide variety of family structures, with the largest proportion reportedly having a single biological mother who had never been married (22%). The group with a single biological mother who had divorced or was separated included the same number of female participants (n = 34, 19%) as the group with a biological mother and stepfather. The group with two biological parents was comprised of 13% of the sample while the group with another adult relative included 12% of the sample. The remaining 15% of the sample included those who reported another type of family structure.

Measures

Exogenous Variables

Exposure to violence

A modified version of the Exposure to Violence Inventory (Selner-O’Hagan, Kindlon, Buka, Raudenbush, & Earls, 1998) was utilized to assess participants’ experience with violent situations. The overall scale consisted of two subscales. The violence victimization scale, which was comprised of seven items, asked youth whether they have ever been the victims of various forms of violence (e.g., Have you ever been raped or the victim of attempted rape or other sexual attack? Have you ever been chased where you thought you might be seriously hurt?). Responses to these seven items were summed to create an index, with higher numbers indicative of exposure to a wider array of forms of violence.

The observed violence scale was also comprised of seven items designed to assess the different types of violence witnessed by participants (e.g., Have you ever seen someone else being raped, an attempt made to rape someone, or any other type of sexual attack? Have you ever witnessed someone else being chased and thought they might be seriously hurt?). Responses to these items were summed to create an index representing the number of forms of violence witnessed by youth. Higher counts on this scale indicated observations of a greater variety of violence.

Endogenous Variables¹

Depressive symptoms

The Brief Symptom Inventory was used to assess participants’ recent experiences with an array of symptoms (Derogatis & Melisaratos, 1983). The Inventory has nine subscales, and the depression subscale was utilized for the current study. This subscale was comprised of six items (e.g., recent feelings of loneliness, feeling blue, lacking interest in things, and feeling hopeless about the future), which were measured with items ranging from (0) not at all, if participants had not experienced the symptom, to (4) extremely, if the participant was severely bothered. These six items were used to compute a mean depressive symptom score, which ranged from zero to four with higher values indicative of more recent experiences of a variety of symptoms.

Hostility

This measure was also collected as part of the Brief Symptom Inventory. Assessed in a similar fashion, this scale assessed the extent to which participants recently experienced several indicators of hostility (e.g., feeling easily annoyed or irritated, having urges to break or smash things, and getting into frequent arguments). Responses ranged from (0) not at all,” if participants had not experienced the symptom, to (4) extremely, if the participant was severely bothered. These five items were used to compute a mean score, which ranged from zero to four, with higher values associated with heightened recent hostile experiences.

Control Variables

Race and ethnicity

Research has documented the disproportionality of criminal justice system contact and detention of racial and ethnic minority youth (Rodriguez, 2013). In addition, youth of color have been found to experience violence at greater rates relative to their White counterparts (López et al., 2017). Furthermore, previous research indicates that the effect of experiencing violence on substance use varies by race/ethnicity with the effect being stronger for White adolescents than Black and Latino adolescents (Schilling, Aseltine, & Gore, 2007). Considering the racial and ethnic diversity of the sample, and the primary issue under investigation, a control variable was created with White non-Hispanic female offenders coded “0” and racial or ethnic minority youth coded “1.”²

Age of first prior offense

One of the most influential factors related to offending among youth is the age at which the first offense or the first incarceration occurred (Myner, Santman, Cappelletty, & Perlmutter, 1998). A continuous measure of age of first prior offense was collected from official records to control for this potentially confounding relationship. This measure ranged from 12 to 18 years, with female offenders in the sample being slightly older than 15 years (M = 15.37, SD = 1.39) at the time of their first recorded offense.

Adults in the household

Research has also shown the composition of adults in the household can include the likelihood children are exposed to violence (Waldfogel, Craigie, & Brooks-Gunn, 2010). A measure of the number of adults in the home was included to control this influence. This item ranged from 0 to 4, with zero representative of no adults in the home and four representative of four or more adults in the home. Most (45%) female offenders in the sample reported having two adults in the home, followed by those who reported having one adult (42%) in the home.

Substance Use

Marijuana use

Participants’ marijuana use was measured with a modified version of the Substance Use Inventory (Chassin, Rogosch, & Barrera, 1991). The substance use subscale asked adolescents to report how frequently they had used marijuana in the past 6 months. Responses ranged from (1) not at all to (9) every day. Higher values were indicative of more frequent marijuana use. The current study included a baseline (T₁) measure of marijuana use and the identical measure collected during the 6-month (T₂) follow-up interview.

Alcohol use

Participants’ alcohol use was also measured with a subscale of the same inventory. Adolescents were asked to report how often they had used alcohol in the past 6-month period. Responses ranged from (1) not at all to (9) every day with higher values representative of more frequent alcohol use. The current study included a baseline (T₁) measure of alcohol use and the identical measure collected during the 6-month (T₂) follow-up interview.

Data Analyses

Analyses for the current study were conducted in several stages. Data management, descriptive, and bivariate statistics were conducted with Stata 15 (StataCorp, 2017). Based on the hypothesized relationships between variables, it was determined path analysis would provide the optimal approach. Path analysis is ideally suited as a formal method to assess the direct and indirect effects among multiple exogenous and endogenous variables. In addition to testing hypothesized relationships between multiple variables, this method also offers the ability to neatly summarize results in a compact and straightforward diagram (Klem, 1995). All path analyses for the current study were conducted with MPlus 8.1 (Muthén & Muthén, 1998-2017). This software performs analyses using a Full-Information Maximum Likelihood estimator with standard errors and a chi-square test statistic that are robust to nonnormality (MLR). This statistical estimator has been found to be an efficient estimation method, which effectively addresses missing data (Enders & Bandalos, 2001).

The proportion of missing data in the study sample was minimal. Information was missing from baseline (T₁) interviews from nine participants, which amounted to a missing data rate of 4%. These missing data were also limited to two T₁ measures: depressive symptoms and hostility. All other T₁ measures had complete data. A slightly higher (7%) missing data rate was observed among the T₂ substance use measures.

Results

Descriptive Statistics

The female adolescent offenders in the sample observed a fair amount of violence. Descriptive information for the aggregate sample as well as female offenders who experienced violence is provided in Table 1. On average, female offenders reported witnessing approximately three (M = 3.26, SD = 2.02) different types of violence. Slightly more than half (55%) of female offenders in the sample reported being victimized and on average, those who were victims experienced one type of violence (M = 1.14, SD = 1.31). In terms of negative emotions, female offenders in the sample reported few (M = 0.72, SD = 0.80) depressive symptoms and slightly higher indications of hostility (M = 1.03, SD = 0.90).

Table 1.

Descriptive Statistics According to Exposure to Violence.

Variable	Aggregate sample(N = 184)	Witnessed violence(n = 67)	Violently victimized(n = 5)	Witnessed & victimized(n = 97)
Variable	M (SD)/%	M (SD)/%	M (SD)/%	M (SD)/%
Racial or ethnic minority	73.4	77.6	80.0	69.1
Age of first prior offense	15.37 (1.39)	15.27 (1.25)	16.60 (1.14)	15.40 (1.46)
Adults in household	1.51 (0.75)	1.67 (0.77)	1.80 (0.45)	1.38 (0.74)
Depressive symptoms	0.72 (0.80)	0.45 (0.62)	0.42 (0.61)	0.99 (0.87)
Hostility	1.03 (0.90)	0.77 (0.81)	0.60 (0.69)	1.31 (0.92)
T₁ Marijuana use	4.15 (3.19)	3.42 (2.79)	3.00 (2.83)	4.84 (3.39)
T₁ Alcohol use	2.68 (2.24)	2.25 (1.92)	2.20 (1.64)	3.13 (2.49)
T₂ Marijuana use	2.50 (2.59)	2.03 (2.03)	1.00 (0.00)	3.10 (3.02)
T₂ Alcohol use	2.04 (1.71)	1.89 (1.51)	1.20 (0.45)	2.25 (1.88)

In terms of substance use, a larger proportion of participants reported marijuana use (61%) at T₁ relative to the proportion that reported alcohol use (54%) during the same time period. Specifically, female offenders reported frequency of marijuana use to be around once per month (M = 4.15, SD = 3.19), on average. Alcohol use was less frequent, with average reports of use falling between 1 and 2 times in the past 6 months (M = 2.68, SD = 2.24). Frequency of substance use was lower during the T₂ interview, with alcohol (M = 2.04, SD = 1.71) and marijuana (M = 2.50, SD = 2.59) reportedly used between one and two times per month, on average.

Bivariate Analyses

Control variables

Initial examination of the control variables demonstrated several significant bivariate relationships. Racial and ethnic minority female offenders (M = 1.47, SD = 1.49) reported higher levels of violence victimization relative to their White (M = 1.02, SD = 1.22) counterparts, t(184) = 2.06, p =.04. In addition, depressive symptom scores varied according to female offenders’ age at first prior offense, F(6, 168) = 2.61, p = .02. Specifically, girls whose first offense occurred at 18 years of age reported the greatest (M = 1.71, SD = 0.35) mean depressive symptom scores, followed by girls whose first offense occurred at 13 years of age (M = 1.21, SD = 0.83). The lowest (M = 0.52, SD = 0.64) depressive symptom scores were reported by female offenders whose first prior offense occurred at 14 years of age. Finally, there was an association between the number of adults in the household and the different types of violence witnessed by female offenders, F(4, 179) = 2.54, p =.04. Female offenders with four or more adults in the home reported witnessing the widest variety of violence (M = 4.50, SD = 3.54) and those with two adults in the home reported witnessing the fewest (M = 2.88, SD = 1.98) types of violence.

Associations with exposure to violence

When comparing female offenders who witnessed violence to offenders who witnessed and experienced violence, some interesting patterns emerged. Table 1 contains descriptive information according to female offenders’ experiences with violence. There were several significant differences with a clear pattern demonstrating greater mental health risk and substance use among female offenders who witnessed violence and also experienced violent victimization. One-way analysis of variance (ANOVA) results indicated female offenders who witnessed violence and also experienced violent victimization had greater depressive symptoms, F(3, 171) = 8.48, p < .001, and greater hostility, F(3, 171) = 7.84, p < .001. Female offenders who witnessed violence and also experienced violent victimization also reported more frequent marijuana, F(3, 180) = 3.27, p = .023, and alcohol, F(3, 180) = 2.96, p = .034, use at baseline. Finally, girls who witnessed violence and also experienced violent victimization reported more frequent marijuana, F(3, 180) = 3.85, p = .011, use at the 6-month follow-up interview.

Correlations

A correlation matrix was analyzed to examine the bivariate relationships among the key study variables (Table 2). There were a number of significant associations between measures of exposure to violence, negative emotions, and substance use. Beginning with the correlations between different exposures to violence, as female offenders observed more forms of violence, reports of violent victimization also increased (r = .45, p < .001). Although there was a significant amount of overlap between these two measures with only five girls reporting experiencing violence only, they were only moderately correlated suggesting distinct constructs were observed. This evidence supports the use of a more precise approach accounting for these two forms of violence exposure can yield important observations with regard to the ways in which witnessing violence and direct victimization influence negative emotions and substance use. In addition, this multidimensional approach is supported with initial analyses conducted by the scale creators and other PTD researchers (Center for Research on Health Care (CRHC), 2018; Selner-O’Hagan et al., 1998).

Table 2.

Matrix of Correlations Among Measures of Exposure to Violence and Substance Use.

Variable	1	2	3	4	5	6	7	8
1. Witness violence	—
2. Violence victim	.45*	—
3. Depressive symptoms	.23^a	.31*	—
4. Hostility	.33*	.30*	.60*	—
5. Alcohol use (T₁)	.08	.15^b	.19^a	.26*	—
6. Marijuana use (T₁)	.17^b	.18^a	.13	.14	.50*	—
7. Alcohol use (T₂)	−.06	<.01	.17^b	.17^b	.54*	.24*	—
8. Marijuana use (T₂)	.09	.12	.21^a	.09	.32*	.32*	.35*	—

Statistical significance at .01.

Statistical significance at .05.

Statistical significance at .001.

Observed violence was also significantly associated with both mental health measures. As the types of violence observed by female offenders increased, depressive symptoms and hostility also increased. Witnessing violence was also positively associated with T₁ marijuana use, although this relationship was weak.

Similar relationships were observed between violent victimization, depressive symptoms, hostility, and substance use. Violent victimization was significantly associated with a rise in depressive symptoms, as well as a rise in hostility. A moderate positive correlation was also observed between depressive symptoms and hostility, which indicated greater depressive symptoms in the presence of higher levels of hostility. Violent victimization was positively associated with both T₁ substance use measures, but these relationships were also fairly weak. The strongest association was observed between hostility and T₁ alcohol use.

With regard to violence exposure and T₂ substance use, the associations did not meet the generally accepted level of significance. However, the relationship observed between violent victimization and T₂ marijuana use approached significance (r = 0.12, p = .113) and the associations between depressive symptoms, hostility, and T₂ substance use measures remained significant. Thus, the mediated effects of depressive symptoms and hostility were explored with the understanding that observed effects would be minimal in their magnitude.

Path analyses

The conceptual model that tested the pathways between exposure to violence, mental health measures, and substance use is presented in Figure 1.

The first iteration of the model was tested to examine the associations between exposure to violence, negative emotions, and T₁ substance use measures. Fit indices demonstrated a significant improvement over an unconstrained model and suggested an adequate fit to the data, baseline χ² = 226.718, df = 25, p < .001; constrained χ² = 15.653, df = 13, p = .268; comparative fit index (CFI) = 0.987; root mean square error of approximation (RMSEA) = 0.033; 90% confidence interval (CI) = [0.000, 0.084]. Results from the model test are presented in Figure 2, with significant associations represented by solid arrows.

Figure 2.

Significant pathways predicting T₁ substance use, β (SE).

As female offenders reported more instances of experiencing firsthand violence, the number of depressive symptoms and levels of hostility significantly increased. Hostility also rose in tandem with the amount of violence female offenders witnessed, but this form of exposure to violence was not associated with depressive symptoms. In addition, hostility was the sole negative emotion associated with substance use, and it was related to alcohol use, not marijuana use.

The second iteration of the model was tested to examine the same associations but utilized the T₂ substance use measures to assess the lagged associations between violence, negative emotions, and subsequent alcohol and marijuana use. Fit indices for this model also demonstrated a significant improvement over an unconstrained model and suggested an adequate fit to the data (baseline χ² = 197.995, df = 25, p = .000; constrained χ² = 17.002, df = 13, p = .199; CFI = 0.977; RMSEA = 0.041, 90% CI = [0.000, 0.089]). Significant associations are represented with solid arrows in Figure 3.

Figure 3.

Significant pathways predicting T₂ substance use, β (SE).

The associations between exposure to violence, depressive symptoms, and hostility remained the same as the first model. The notable difference, utilizing the T₂ substance use measures, was observed between the negative emotions measures, alcohol, and marijuana use. The association between hostility and alcohol use was no longer significant, but the association between depressive symptoms and subsequent marijuana use achieved significance in this model.

To summarize, four important results emerged from this series of model tests. The first was related to the significant positive associations between firsthand experiences of violence, depressive symptoms, and hostility. The second was related to the significant positive association between observed violence and elevated hostility. The third indicated greater hostility was associated with more frequent and recent alcohol use, but only at Time 1. The fourth highlighted the association between elevated depressive symptoms and more frequent but later marijuana use at Time 2. Together, these results indicate that alcohol and marijuana use have different precursors among young female offenders, and the determinants of these behaviors appear to change over time.

Discussion

The intent of this study was to empirically test a GST model with an ethnically diverse sample of serious female adolescent offenders. We focused on witnessing and experiencing violence as the types of strains or adverse child experiences that girls endured and examined the emotional responses to such strains as well as substance use as the coping mechanism to such negative emotions. In particular, we hypothesized that the relationships between exposure to violence and substance use would be mediated by depressive symptoms and hostility. Unlike prior GST studies, we examined how two different types of exposure to violence—personal victimization and witnessing violence—impacted girls’ emotional coping responses and substance use. Our study also has implications for Pathways Theory, which posits that personal victimization—usually in the form of childhood abuse—leads to mental health issues such as depression and anxiety, which, if left untreated, may result in self-medicating substance use behaviors (Belknap, 2014; Chesney-Lind & Pasko, 2013; Gehring, 2018; Vaske & Gehring, 2010).

Consistent with other research on female adolescent offenders (Abram et al., 2004; Baglivio et al., 2014; Foy et al., 2012), exposure to violence was very high among the study participants: Approximately 92% of the sample had been exposed to violence. Furthermore, there was a large overlap between witnessing violence and being violently victimized. Despite this overlap, experiencing violence and witnessing violence independently contributed to negative emotions despite the fact that they tend to happen simultaneously. More specifically, we found that personally experiencing violence was associated with increased depressive symptoms and hostility and these relationships held steady across both the cross-sectional and longitudinal models. These findings add to our limited understanding of how female adolescent offenders respond to being personally victimized and underscores the importance of further studying anger as a possible coping response among members of this population.

Witnessing violence was associated with increased hostility, a relationship that was maintained at 6-month follow-up. No other research to date has examined this relationship among serious female adolescent offenders. Future research is needed to further understand why witnessing violence is related to hostility among female adolescent offenders. One possibility is that witnessing violence may result in normative beliefs about the acceptability of aggression (Boxer et al., 2008). Context may also play a role with regard to how adolescents emotionally respond to witnessing violence (Mrug, Loosier, & Windle, 2008; Wright et al., 2013). For example, girls growing up in neighborhoods characterized by high rates of violence may be more prone to react with hostility in response to witnessing violence (Ness, 2004).

Contrary to our hypothesis, depressive symptoms did not mediate the relationship between experiencing violence and marijuana use at Time 1; however, support for mediation was found at Time 2. One possibility for these contradictory findings is that female adolescent offenders were introduced to more marijuana-using peers via their involvement in the juvenile justice system, as previous research indicates that peer substance use is associated with more frequent drug use among female adolescent offenders (Neff & Waite, 2007). Of course, this remains conjecture at this point given that we did not include a measure of peer substance use in our model.

Hostility mediated the relationship between experiencing violence and alcohol use, but only in our cross-sectional model. This finding was consistent with Sharp et al. (2012) who also found in their cross-sectional analysis that anger expression mediated the relationships between physical abuse and alcohol use among women prisoners. Unlike Sharp et al. (2012), we did not find support for hostility mediating the relationship between experiencing violence and marijuana use.

Hostility also mediated the relationship between witnessing violence and alcohol use, but again evidence for mediation was only found in our cross-sectional model. This finding is inconsistent with Sharp et al. (2012) who failed to find a mediation effect for anger expression on the relationship between witnessing violence and daily alcohol use among a sample of women prisoners. While Sharp and colleagues (2012) included a one-item measure of witnessing violence specific to the family domain (i.e., having a battered mother), we included a comprehensive measure of witnessing violence across various domains.

Contrary to our expectations, hostility did not mediate the relationship between either exposure to violence measure and either of the substance use measures at Time 2. One possible explanation for the lack of mediation effects at Time 2 is that the participants in our sample were exposed to anger management programming, which is quite common in juvenile detention and correctional facilities (Young, Dembo, & Henderson, 2007). Again, this remains conjecture at this point and additional research, preferably with a longer follow-up, is needed to test this premise.

Finally, it is worth noting that girls in our sample reported higher rates of marijuana use than alcohol use. This finding is consistent with previous research on female adolescent offenders. In a study of 1,829 randomly selected detained youth in Cook County, Teplin, Abram, McClelland, Dulcan, and Mericle (2002) found that approximately 41% of girls met the diagnostic criteria for marijuana use disorder compared with approximately 27% who met the criteria for alcohol use disorder. Although beyond the scope of the current study, understanding why female adolescent offenders use marijuana at higher rates than alcohol is worthy of additional research with a focus on drug and alcohol use motives.

The current study has several strengths, including an ethnically diverse sample of serious female adolescent offenders and data collected over two waves. Most studies testing GST among youth focus on adolescents from the general population (e.g., Francis, 2014; Kaufman, 2009; Lin et al., 2011) as opposed to system-involved youth who are most at risk for future substance use. Another unique contribution of our study is that we tested whether hostility mediated the relationship between exposure to violence and substance use. Such an emphasis is important given that much of the literature on adolescent girls’ coping focuses exclusively on internalizing symptoms as a response to violence exposure. Finally, we were able to test for both direct and indirect effects of experiencing violence and witnessing violence on substance use, which has implications for both GST and Pathways Theory as well as studies related to the relationship between adverse child experiences and substance use among adolescents.

Despite our study’s strengths, several caveats and limitations should be noted. First, we did not distinguish between personally experiencing and witnessing violence across a number of different contexts (Mrug et al., 2008; Wright et al., 2013). Second, we examined the effects of depression and hostility separately despite research indicating that system-involved girls’ anger often coincides with depression and anxiety (Cauffman, Lexcen, Goldweber, Shulman, & Grisso, 2007). Indeed, our analyses revealed that depressive symptoms and hostility were positively correlated for the young women in our sample. Finally, it should be emphasized that while our results provide partial support for GST and Pathways Theory, not all hypothesized relationships were significant, and significant relationships varied depending on whether the model was cross-sectional or longitudinal. Moreover, the expected relationships between exposure to violence and substance use were not as strong as we had anticipated, which resulted in either nonsignificant or small mediation effects. Additional research is needed to replicate our findings, as well as research that expands upon our findings by examining outcomes other than substance use.

This research should include a more specific measure of alcohol and marijuana use that taps into whether adolescents are drinking and using marijuana as a coping strategy or for other reasons as suggested by Catanzaro and Laurent (2004). More information on adolescent girls’ perceptions, particularly those who are most at risk, such as the young women in the current study, could inform cognitive-behavioral interventions aimed at increasing positive coping skills and decreasing marijuana and alcohol use among serious female offenders. Further research on girls’ legitimate coping mechanisms to strain can enhance GST’s ability to test alternate models explaining desistance from delinquency as well (Jang, 2007). For instance, additional models, which may help elucidate the relationships between coping and substance use, can include more detailed tests examining the timing between negative emotions and substance use.

Finally, policy recommendations from this study point toward the importance of early identification and treatment of female adolescent offenders’ emotional and physical needs in a significant, timely manner. Previous research has also found that when providers are asked what services female offenders need, they assert, “real drug treatment” (Davidson, Pasko, & Chesney-Lind, 2011). The focus on state-sponsored drug services for youth is often on education and prevention for this at-risk population, with any direct services for adjudicated drug-using youth housed in Juvenile Drug Court or a school-based program. Intervention services and sentencing alternatives for those adolescents who are “in risk” of addiction are limited. Currently, there is a dearth of intensive residential drug treatment options for adolescents, with gender-aware programming becoming even more limited for system-involved girls (Wasserman et al., 2003). As this research—as well as other examinations of female offenders (e.g., Chesney-Lind & Pasko, 2013)—has shown, more community-based substance abuse treatment programs that move beyond the prevention and school-based models and address the severity of drug use for girls who are already “in risk” of lifetime trajectories of addiction are needed.

Footnotes

Declaration of Conflicting Interests

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

Funding

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

Notes

Author Biographies

Vera Lopez is a professor of Justice Studies in the School of Social Transformation at Arizona State University (ASU). Lopez’s research areas include delinquency, substance use, and sexual risk taking with a special emphasis on system-involved girls and Latina girls. She is the author of the book “Complicated Lives: Girls, Parents, Drugs, and Juvenile Justice” which was published by Rutgers University Press in 2017.

Albert Kopak holds a position at the rank of Associate Professor in the Department of Criminology and Criminal Justice at Western Carolina University. He received his Ph.D. in Justice Studies at Arizona State University and specializes in research on substance use disorders among criminal justice populations.

Lisa Pasko is associate professor of Sociology and Criminology and affiliated faculty in Gender and Women’s Studies Program at the University of Denver. In addition to numerous articles, book chapters, and technical reports, she is co-author of The Female Offender: Girls, Women, and Crime and Girls, Women, and Crime: Selected Readings (2013, Sage Publications). Professor Pasko’s research examines experiences of sexual and ethnic minority girls in the justice system, paying particular attention to correctional attitudes about sexual behaviors and identities.

References

Abram

K. M.

Teplin

L. A.

Charles

D. R.

Longworth

S. L.

McClelland

G. M.

Dulcan

M. K.

(2004). Posttraumatic stress disorder and trauma in youth in juvenile detention. Archives of General Psychiatry, 61, 403-410.

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-88.

Agnew

(2002). Experienced, vicarious, and anticipated strain: An exploratory study on physical victimization and delinquency. Justice Quarterly, 19, 603-632.

Agnew

(2006). Pressured into crime: An overview of general strain theory. Los Angeles, CA: Roxbury.

Baglivio

M. T.

Epps

(2016). The interrelatedness of adverse childhood experiences among high-risk juvenile offenders. Youth Violence and Juvenile Justice, 14, 179-198.

Baglivio

M. T.

Epps

Swartz

Huq

M. S.

Sheer

Hardt

N. S.

(2014). The prevalence of Adverse Childhood Experiences (ACE) in the lives of juvenile offenders. Journal of Juvenile Justice, 3, 1-17.

Bailey

McCloskey

(2005). Pathways to adolescent substance use among sexually abused girls. Journal of Abnormal Child Psychology, 33, 39-54.

Belknap

(2014). The invisible woman: Gender, crime, and justice (3rd ed.). Belmont, CA: Thomson Wadsworth.

Berenson

A. B.

Wiemann

C. M.

McCombs

(2001). Exposure to violence and associated health-risk behaviors among adolescent girls. Archives of Pediatrics & Adolescent Medicine, 155, 1238-1242.

10.

Bergen

Martin

Richardson

Allison

Roeger

(2004). Sexual abuse, antisocial behaviour and substance use: Gender differences in young community adolescents. Australian and New Zealand Journal of Psychiatry, 38, 34-41.

11.

Boxer

Morris

A. S.

Terranova

A. M.

Kithakye

Savoy

S. C.

McFaul

A. F.

(2008). Coping with exposure to violence: Relations to emotional symptoms and aggression in three urban samples. Journal of Child and Family Studies, 17, 881-893.

12.

Brennan

Breitenbach

Dieterich

Salisbury

E. J.

Van Voorhis

(2012). Women’s pathways to serious and habitual crime: A person-centered analysis incorporating gender responsive factors. Criminal Justice and Behavior, 39, 1481-1508. doi:10.1177/0093854804265179

13.

Broidy

Agnew

(1997). Gender and crime: A general strain theory perspective. Journal of Research in Crime and Delinquency, 34, 275-306.

14.

Carbone-López

Kruttschnitt

Macmillan

(2006). Patterns of intimate partner violence and their associations with physical health, psychological distress, and substance use. Public Health Reports, 121, 382-392.

15.

Catanzaro

S. J.

Laurent

(2004). Perceived family support, negative mood regulation expectancies, coping, and adolescent alcohol use: Evidence of mediation and moderation effects. Addictive Behaviors, 29, 1779-1797.

16.

Cauffman

Lexcen

F. J.

Goldweber

Shulman

E. P.

Grisso

(2007). Gender differences in mental health symptoms among delinquent and community youth. Youth Violence and Juvenile Justice, 5, 287-307.

17.

Center for Research on Health Care (CRHC). (2018). Exposure to Violence (ETV)—Subject baseline. Retrieved from https://www.pathwaysstudy.pitt.edu/codebook/etv-sb.html

18.

Chassin

Rogosch

Barrera

(1991). Substance use and symptomatology among adolescent children of alcoholics. Journal of Abnormal Psychology, 100, 449-463.

19.

Chesney-Lind

Pasko

(2013). The female offender: Girls, women, and crime. Thousand Oaks, CA: Sage.

20.

Chesney-Lind

Shelden

R. G.

(2013). Girls, delinquency, and juvenile justice. Chichester, West Sussex, UK: John Wiley & Sons, Inc.

21.

Craig

J. M.

Cardwell

S. M.

Piquero

A. R.

(2017). The effects of criminal propensity and strain on later offending. Crime & Delinquency, 63, 1655-1681.

22.

Davidson

J. T.

Pasko

Chesney-Lind

(2011). “She’s way too good to lose”: An evaluation of Honolulu’s girls court. Women & Criminal Justice, 21, 308-327.

23.

Derogatis

L. R.

Melisaratos

(1983). The Brief Symptom Inventory: An introductory report. Psychological Medicine, 13, 595-605.

24.

Dube

S. R.

Miller

J. W.

Brown

D. W.

Giles

W. H.

Felitti

V. J.

Dong

Anda

R. F.

(2006). Adverse childhood experiences and the association with ever using alcohol and initiating alcohol use during adolescence. Journal of Adolescent Health, 38(4), 444.e1-444.e10.

25.

Eftekhari

Turner

A. P.

Larimer

M. E.

(2004). Anger expression, coping, and substance use in adolescent offenders. Addictive Behaviors, 29, 1001-1008.

26.

Enders

C. K.

Bandalos

D. L.

(2001). The relative performance of full information maximum likelihood estimation for missing data in structural equation models. Structural Equation Modeling, 8, 430-457.

27.

Espelage

D. L.

Low

Rao

M. A.

Hong

J. S.

Little

T. D.

(2014). Family violence, bullying, fighting, and substance use among adolescents: A longitudinal mediational model. Journal of Research on Adolescence, 24, 337-349.

28.

Felitti

V. J.

Anda

R. F.

Nordenberg

Williamson

D. F.

Spitz

A. M.

Edwards

. . . Marks

J. S.

(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, 245-258.

29.

Foy

D. W.

Ritchie

I. K.

Conway

A. H.

(2012). Trauma exposure, posttraumatic stress, and comorbidities in female adolescent offenders: Findings and implications from recent studies. European Journal of Psychotraumatology, 3, 17247.

30.

Francis

K. A.

(2014). General strain theory, gender, and the conditioning influence of negative internalizing emotions on youth risk behaviors. Youth Violence and Juvenile Justice, 12, 58-76.

31.

Gehring

K. S.

(2018). A direct test of pathways theory. Feminist Criminology, 13, 115-137.

32.

Goodkind

Ruffolo

M. C.

Bybee

Sarri

(2009). Coping as a mediator of the effects of stressors and supports on depression among girls in juvenile justice. Youth Violence and Juvenile Justice, 7, 100-118.

33.

Grant

K. E.

Katz

B. N.

Thomas

K. J.

O’Koon

J. H.

Meza

C. M.

DiPasquale

A.-M.

. . . Bergen

(2004). Psychological symptoms affecting low-income urban youth. Journal of Adolescent Research, 19, 613-634.

34.

Hanson

R. F.

Borntrager

Self-Brown

Kilpatrick

D. G.

Saunders

B. E.

Resnick

H. S.

Amstadter

(2008). Relations among gender, violence exposure, and mental health: The national survey of adolescents. American Journal of Orthopsychiatry, 78, 313-321.

35.

Harlow

(1999). Prior abuse reported by inmates and probationers. Washington, DC: Bureau of Justice Statistics.

36.

Jang

S. J.

(2007). Gender differences in strain, negative emotions, and coping behaviors: A general strain theory approach. Justice Quarterly, 24, 523-553.

37.

Kaufman

J. M.

(2009). Gendered responses to serious strain: The argument for a general strain theory of deviance. Justice Quarterly, 26, 410-444.

38.

Klem

(1995). Path analysis. In Grimm

L. G.

Yarnold

P. R.

(Eds.), Reading and understanding multivariate statistics (pp. 65-98). Washington, DC: American Psychological Association.

39.

Lin

W. H.

Cochran

J. K.

Mieczkowski

(2011). Direct and vicarious violent victimization and juvenile delinquency: An application of general strain theory. Sociological Inquiry, 81, 195-222.

40.

López

C. M.

Andrews

A. T.

III Chisolm

A. M.

de Arellano

M. A.

Saunders

Kilpatrick

D. G.

(2017). Racial/ethnic differences in trauma exposure and mental health disorders in adolescents. Cultural Diversity and Ethnic Minority Psychology, 23, 382-387.

41.

McClelland

Teplin

Abram

(2004). Detection and prevalence of substance use among juvenile detainees (Juvenile justice bulletin). Washington, DC: Office of Juvenile Justice and Delinquency Prevention, Office of Justice Programs, U.S. Department of Justice.

42.

McKnight

Loper

(2002). The effect of risk and resilience factors on the prediction of delinquency in adolescent girls. School Psychology International, 23, 186-198.

43.

Merikangas

K. R.

J. P.

Burstein

Swanson

S. A.

Avenevoli

Cui

. . . Swendsen

(2010). Lifetime prevalence of mental disorders in U.S. adolescents: Results from the National Comorbidity Survey Replication—Adolescent Supplement (NCS-A). Journal of the American Academy of Child and Adolescent Psychiatry, 49, 980-989.

44.

Miller

Winn

D. M.

Taylor

Wiki

(2012). Girls’ relational orientation and interpersonal dynamics of delinquency. In Miller

Leve

L.D.

Kerig

P.K.

(Eds.), Delinquent girls: Contexts, relationships, and adaptation (pp. 85-101). New York, NY: Springer.

45.

Mrug

Loosier

P. S.

Windle

(2008). Violence exposure across multiple contexts: Individual and joint effects on adjustment. American Journal of Orthopsychiatry, 78, 70-84.

46.

Mulvey

E. P.

(2004). Introduction: Pathways to desistance study. Youth Violence and Juvenile Justice, 2, 211-212.

47.

Mulvey

E. P.

(2013). Research on pathways to desistance, 2000-2010. Ann Arbor, MI: Inter-University Consortium for Political and Social Research [distributor]. doi:10.3886/ICPSR29961.v2

48.

Muthén

L. K.

Muthén

B. O.

(1998–2017). Mplus user’s guide (8th ed.). Los Angeles, CA: Author.

49.

Myner

Santman

Cappelletty

G. G.

Perlmutter

B. F.

(1998). Variables related to recidivism among juvenile offenders. International Journal of Offender Therapy and Comparative Criminology, 42, 65-80.

50.

Neff

J. L.

Waite

D. E.

(2007). Male versus female substance abuse patterns among incarcerated juvenile offenders: Comparing strain and social learning variables. Justice Quarterly, 24, 106-132.

51.

Ness

C. D.

(2004). Why girls fight: Female youth violence in the inner city. The Annals of the American Academy of Political and Social Science, 595, 32-48.

52.

Nichols

T. R.

Mahadeo

Bryant

Botvin

G. J.

(2008). Examining anger as a predictor of drug use among multiethnic middle school students. Journal of School Health, 78, 480-486. doi:10.1111/j.1746-1561.2008.00333.x

53.

Pasko

Chesney-Lind

(2012). Girls’ violence and juvenile justice: A critical examination. In DeKeseredy

Dragiewicz

(Eds.), Handbook of critical criminology (pp. 252-266). New York, NY: Routledge.

54.

Pinchevsky

G. M.

Fagan

A. A.

Wright

E. M.

(2014). Victimization experiences and adolescent substance use: Does the type and degree of victimization matter? Journal of Interpersonal Violence, 29, 299-319.

55.

Piquero

Sealock

(2004). Gender and general strain theory: A preliminary test of Broidy and Agnew’s gender/GST hypotheses. Justice Quarterly, 21, 125-158.

56.

Putnam

F. W.

(2003). Ten-year research update review: Child sexual abuse. Journal of the American Academy of Child & Adolescent Psychiatry, 42, 269-278.

57.

Radliff

K. M.

Wheaton

J. E.

Robinson

Morris

(2012). Illuminating the relationship between bullying and substance use among middle and high school youth. Addictive Behaviors, 37, 569-572.

58.

Rees

Silove

Chey

Ivancic

Steel

Creamer

. . . Forbes

(2011). Lifetime prevalence of gender-based violence in women and the relationship with mental disorders and psychosocial function. Journal of the American Medical Association, 306, 513-521.

59.

Riehman

Blutenthal

Juvonen

Morral

(2003). Adolescent social relationships and the treatment process: Findings from quantitative and qualitative analyses. Journal of Drug Issues, 33, 865-896.

60.

Rodriguez

(2013). Concentrated disadvantage and the incarceration of youth: Examining how context affects juvenile justice. Journal of Research in Crime and Delinquency, 50, 189-215.

61.

Rothman

E. F.

Edwards

E. M.

Heeren

Hingson

R. W.

(2008). Adverse childhood experiences predict earlier age of drinking onset: Results from a representative US sample of current or former drinkers. Pediatrics, 122, e298-e304.

62.

Schilling

E. A.

Aseltine

R. H.

Gore

(2007). Adverse childhood experiences and mental health in young adults: A longitudinal survey. BMC Public Health, 7, Article 30.

63.

Schubert

C. A.

Mulvey

E. P.

Steinberg

Cauffman

Losoya

S. H.

Hecker

. . . Knight

G. P.

(2004). Operational lessons from the pathways to desistance project. Youth Violence and Juvenile Justice, 2, 237-255.

64.

Selner-O’Hagan

M. B.

Kindlon

D. J.

Buka

S. L.

Raudenbush

S. W.

Earls

F. J.

(1998). Assessing exposure to violence in urban youth. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 39, 215-224.

65.

Sharp

S. F.

Peck

B. M.

Hartsfield

(2012). Childhood adversity and substance use of women prisoners: A general strain theory approach. Journal of Criminal Justice, 40, 202-211.

66.

Simkins

Katz

(2002). Criminalizing abused girls. Violence Against Women, 8, 1474-1500.

67.

Simpson

T. L.

Miller

W. R.

(2002). Concomitance between childhood sexual and physical abuse and substance use problems: A review. Clinical Psychology Review, 22, 27-77.

68.

StataCorp. (2017). Stata statistical software: Release 15. College Station, TX: StataCorp LP.

69.

Teplin

L. A.

Abram

K. M.

McClelland

G. M.

Dulcan

M. K.

Mericle

A. A.

(2002). Psychiatric disorders in youth in juvenile detention. Archives of General Psychiatry, 59, 1133-1143.

70.

Tjaden

P. G.

Thoennes

(2006). Extent, nature, and consequences of rape victimization: Findings from the National Violence Against Women Survey. Washington, DC: National Institute of Justice.

71.

Ullman

(2004). Sexual assault victimization and suicidal behavior in women: A review of the literature. Aggression and Violent Behavior, 9, 331-351.

72.

Vaske

Gehring

(2010). Mechanisms linking depression to delinquency for males and females. Feminist Criminology, 5, 8-28.

73.

Vermeiren

Schwab-Stone

Deboutte

Leckman

P. E.

Ruchkin

(2003). Violence exposure and substance use in adolescents: Findings from three countries. Pediatrics, 111, 535-540.

74.

Waldfogel

Craigie

T.-A.

Brooks-Gunn

(2010). Fragile families and child wellbeing. The Future of Children, 20, 87-112.

75.

Wasserman

G. A.

Keenan

Tremblay

Coie

Herrenkohl

Loeber

Petechuk

(2003). Risk and protective factors of child delinquency. Washington, DC: Office of Juvenile Justice and Delinquency Prevention.

76.

Wright

E. M.

Fagan

A. A.

Pinchevsky

G. M.

(2013). The effects of exposure to violence and victimization across life domains on adolescent substance use. Child Abuse & Neglect, 37, 899-909.

77.

Young

D. W.

Dembo

Henderson

C. E.

(2007). A national survey of substance abuse treatment for juvenile offenders. Journal of Substance Abuse Treatment, 32, 255-266.

78.

Zinzow

H. M.

Ruggiero

K. J.

Resnick

Hanson

Smith

Saunders

Kilpatrick

(2009). Prevalence and mental health correlates of witnessed parental and community violence in a national sample of adolescents. Journal of Child Psychology and Psychiatry, 50, 441-450.

Substance Use Pathways Among Female Adolescent Offenders

Abstract

Keywords

Conceptual Framework

Exposure to Violence and Substance Use

Exposure to Violence, Negative Emotions, and Substance Use

Study Purpose

Method

Data

Measures

Exogenous Variables

Exposure to violence

Endogenous Variables 1

Depressive symptoms

Hostility

Control Variables

Race and ethnicity

Age of first prior offense

Adults in the household

Substance Use

Marijuana use

Alcohol use

Data Analyses

Results

Descriptive Statistics

Bivariate Analyses

Control variables

Associations with exposure to violence

Correlations

Path analyses

Discussion

Footnotes

Declaration of Conflicting Interests

Funding

Notes

Author Biographies

References

Endogenous Variables¹