Youth Pathways: Evaluating the Influence of Gender,Involvement With the Public Mental Health System,Perceived Mental Health Need,and Traumatic Experiences on Juvenile Justice System Processing

Abstract

This study examined the influence of gender, participation in public mental health services, perceived mental health need, and traumatic experiences on juvenile justice system processing. Data were collected on youth formally processed for delinquency or status offenses by a large statewide juvenile justice system and youth served by the state’s public mental health system during a 7-year period (n = 271,427). The results indicated that accessing mental health services in the community or being viewed as having a mental health need by a juvenile probation officer increased the likelihood of youth being removed from their home by the juvenile justice system, especially for girls. While girls generally had decreased odds of out-of-home placement compared to boys, their odds significantly increased when combined with involvement in public mental health services and/or receiving a violation of probation. Once placed out of the home, boys discharged much more quickly than girls, with the rate of discharge being related primarily to the severity of their offenses. When controlling for relevant variables using propensity score matching, girls who experienced trauma or involvement with the public mental health system had a significantly lower rate of discharge from out-of-home placement than similarly matched boys.

Keywords

mental health gender out-of-home placement disposition trauma

The number of juvenile offenders in out-of-home placement across the United States has decreased by 42% since 1997. However, the rate of girls’ involvement in the juvenile justice system has steadily increased (Hockenberry & Sickmund, 2014). Once juvenile justice system–involved girls are removed from the home, their rate of release from out-of-home placement is slower than for similarly situated boys (Espinosa & Sorensen, 2015). Girls who have had traumatic life experiences are prone to involvement in status offenses and other forms of minor delinquency, which are likely to bring them into contact with the juvenile justice system. Judges have been shown to place female offenders in out-of-home placement or confinement for low-level offenses such as contempt of court, violation of probation (VOP), or violations of court orders. A prior study conducted in three metropolitan counties in Texas during 2007–2008 found that exposure to traumatic experiences increased the likelihood of out-of-home placement or confinement for boys and girls but had a stronger influence on all levels of placement—from detention through state correctional commitment—for girls. Similarly, while a VOP increased the odds of secure out-of-home placement for both, girls with a VOP experienced 7 times greater risk of confinement than boys (Espinosa, Sorensen, & Lopez, 2013).

This study examined pathways girls take to and through the juvenile justice system by evaluating data on youth served by a large statewide juvenile justice system from 2007 to 2014. While much attention has been paid to gender disparities in juvenile justice case processing, most studies have limited their focus to pretrial detention decisions, adjudication, or dispositional outcomes. Given girls’ pathways to and through the juvenile justice system, which often includes trauma, mental health needs, and the use of violation of a court order or contempt of court as a means of confining girls, it is reasonable to hypothesize that girls who have had contact with the mental health system will be subjected to many of these same forces in the processing of their cases through the juvenile justice system. This study examined differences in out-of-home placements between boys and girls. Specifically, the study sought to determine whether girls—particularly those who have experienced trauma—who were involved with the public mental health service system or perceived to have a mental health need by their supervising juvenile probation officer experienced more severe and/or longer out-of-home placements than similarly situated boys.

Gendered Pathways

Girls experience a different pathway toward and through the juvenile justice system than boys. Although female offenders occasionally engage in conduct that is stereotypically male, such as aggression and assaultive behavior (Cauffman, Feldman, Waterman, & Steiner, 1998), they more often suppress their aggression and struggle with the difficulty of managing their emotions, especially those associated with depression and anxiety (Ford, Chapman, Mack, & Pearson, 2006). Compared to boys, female offenders have a higher risk of suicidality (Wannan & Fombonne, 1998; Wasserman, McReynolds, Ko, Katz, & Carpenter, 2005) and greater conflict with school and family (Zoccolillo, Tremblay, & Vitaro, 1996).

Expanding on the research on the impact of adverse childhood experiences (ACEs) on negative adult health and criminal justice outcomes, researchers have begun to examine the influence or impact of ACEs on juvenile justice involvement. One such study found that while composite ACE scores were generally similar across genders for youth involved in the juvenile justice system, girls reported higher rates of sexual abuse than boys (Baglivio, Epps, Swartz, Huq, & Hardt, 2014). In addition, the study indicated girls had a higher rate of endorsement on any single ACE indicator than boys. In a follow-up study, Baglivio and Epps (2016) found the number of justice-involved girls who had six or more ACE endorsements (29%) was more than 2 times greater than those of their male counterparts (14%).

The lives of girls involved with the juvenile justice system are often complicated by chaos and interpersonal conflict, and although virtually absent from formal theories of girls’ delinquency, research has identified a correlation between girls’ sexual abuse and subsequent delinquency. Chamberlain and Moore (2002) found that 72% of girls involved in the juvenile justice system had been involved in relational aggression. Sedlak and McPherson (2010) in their analysis of data collected by the Survey of Youth in Residential Placement (SYRP) found girls were almost twice as likely to report prior physical abuse (42% of girls vs. 22% of boys) and had a greater likelihood of the perpetrator of the physical abuse being a sibling or parent than boys (69% vs. 40%). The same study found that girls reported prior sexual abuse at a rate of almost 4 times higher than boys (35% vs. 8%) and reported higher rates of physical injuries due to the abuse (Sedlak & McPherson, 2010).

Girls who reside in chaotic and violence-prone homes have heightened risk factors for delinquency such as truancy, sexual promiscuity, running away, and substance abuse (Osofsky, 1999; Thornberry, Huizinga, & Loeber, 2004). Girls who are arrested as runaways often experience family violence, including emotional, physical, and sexual abuse (Chesney-Lind & Shelden, 1998; Koroki & Chesney-Lind, 1985). This has led some to theorize that victimization may affect girls’ juvenile justice system involvement (Gavazzi, Yarcheck, & Chesney-Lind, 2006; Joe & Chesney-Lind, 1995; Medina, Ralphs, & Aldridge, 2012).

Gender Disparity in System Processing

Studies of juvenile justice system responses to delinquent behavior have consistently found both legal and extralegal factors contribute to the detention and dispositional decisions. However, findings on the influence of gender on disposition decisions have been inconsistent. Some studies have indicated that girls are the recipients of more severe sanctions than boys, especially related to the commission of status offenses (Bishop & Frazier, 1992; Chesney-Lind, 1977; Odem, 1995). Other studies indicate that girls receive more lenient outcomes for delinquent behavior than boys (Bishop & Frazier, 1996) or fail to find any differences in the handling of delinquency cases based on gender (Kempf-Leonard & Sontheimer, 1995).

There is also evidence suggesting that girls receive both more severe and more lenient outcomes than boys depending on the stage of case processing. In their analysis of the petition, adjudication, and disposition stages of juvenile court processing, MacDonald and Chesney-Lind (2001) reported no difference between boys and girls in the decision to petition an offense. However, during adjudication, “charge seriousness” was more important for girls than boys in court determinations to process the case forward to formal disposition. During the disposition stage, the reverse was discovered. Thus, once adjudicated delinquent, female juvenile offenders were “more likely than boys to be given a restrictive sanction for a less serious offense” (p. 187).

It has been suggested that courts detain girls through findings of contempt of court, VOPs, or violations of court orders for underlying status offenses or minor delinquent behavior (Bishop & Frazier, 1992; Sherman, 2005). As a result, some have postulated there has been an increase in the number of girls being detained preadjudication for offenses that are less threatening to the community than those of boys. In some circumstances, more than 70% of youth detained for status offenses were girls (Sickmund, Sladky, & Kang, 2004). A survey of four study sites conducted by the American and National Bar Associations (2001) found 30% of girls with an initial detention were detained again within one year. For those redetained, 53% of girls were detained for probation or technical violations compared to 41% of boys. For those who returned to detention twice in one year, 60% of the girls compared to 47% of the boys were detained for probation/technical violations. Even more striking, 72% of the girls compared to 49% of the boys who were detained three or more times within 1 year were detained for VOPs.

In their evaluation of the risk and needs of youth in a detention center, Gavazzi, Yarcheck, and Chesney-Lind (2006) noted girls were significantly more likely to be detained for incorrigibility and domestic violence, and parents were more likely to be the complainants. Their findings indicated boys were significantly more likely to have been arrested for property offenses, with complainants being citizens in the community rather than the boys’ parents. Providing a strong statement as to the difference in gender-based risk factors for detention, the authors concluded: “…boys are detained as a response to public safety issues, whereas girls are detained because of problems at home” (p. 608).

Gender differences have been found to exist for VOPs as well, wherein VOP resulted in out-of-home placement more often for girls than boys. A study conducted with a sample of over 30,000 juveniles found VOP was a positive predictor of the level of out-of-home placement for both, yet girls with a VOP experienced 7 times greater risk of secure out-of-home placement than boys (Espinosa et al., 2013). The findings suggest that the use of VOP may serve as a gateway to out-of-home placement.

Trauma

Studies of juvenile justice–involved youth have found disproportionately high rates of traumatic experiences, generally between 70% and 90% (Ford, Chapman, Conner & Cruise, 2012; Teplin, Welty, Abram, Dulcan, & Washburn, 2012). Although not all youth involved with the juvenile justice system have experienced a traumatic event, (Abram et al., 2004; Espinosa et al., 2013; Ford, Hartman, Hawke, & Chapman, 2008) research have indicate as much as three out of every four youth involved with the juvenile justice system have been exposed to trauma (Ford et al., 2006). In their analysis of the SYRP, Yoder, Whitaker, and Quinn (2017) found that youth who had experienced trauma had greater likelihood of arrest and incarceration. A study of detained youth found over 90% had at least one traumatic experience, with 84% having more than one, and 55% having been exposed to six or more (Abram et al., 2007).

Sexual victimization, in particular, is a common form of trauma experienced by girls. In a study of chronically delinquent offenders, Sherman (2005) found only 3% of boys had a history of sexual abuse in comparison to 77% of girls, suggesting a relationship between trauma and chronic delinquency in girls. Sedlak et al. (2010) found that among a sample of juveniles in out-of-home placements, girls reported a rate of prior sexual abuse that was over 4 times higher than boys (35% vs. 8%) and reported higher rates of physical injuries as a result of sexual abuse than boys. Goodkind, Ng, and Sarri (2006) reported girls who had a sexual abuse history were more likely to be placed in secure residential care than girls with no abuse history. Espinosa, Sorensen, and Lopez (2013) found a history of traumatic experiences was a significant influential factor in placement decisions regardless of gender, yet separate multinomial logistic regression models showed past traumatic experiences to be an especially influential predictor of placement level for girls. The disparity between boys and girls for the influence of trauma on placement level was most pronounced for local nonsecure and county-operated secure placements. Expanding on their initial analysis, another study found that even after controlling for other influential variable such as offense severity, prior record, age at referral, and facility type, girls spent longer periods in postadjudicatory confinement than boys, particularly when the girls reported prior traumatic experiences (Espinosa & Sorensen, 2015). Those findings are consistent with the Conrad et al. (2014) conclusion that a history of sexual abuse is a predictor of recidivism for female offenders but not for males.

Gender Disparity in Mental Health Treatment

Offenders with mental health needs are much less successful under supervision in the community than those without (Monahan et al., 2005; Skeem, Emke-Frances, & Louden, 2006; Solomon, Draine, & Marcus, 2002). This seems to be particularly true for girls, who have been found to have higher rates of mental health needs than boys and to respond less positively to system involvement (Teplin et al., 2002; Wasserman et al., 2005). Espinosa et al. (2013) found mental health need was a significant predictor of severity of placement.

Research suggests boys and girls are treated differentially within public mental health systems. Several studies have identified a correlation between severe mental health disorders and justice system involvement. Davis, Fisher, Gershenson, Grudzinskas, and Banks (2009) found girls involved with the public mental health system were arrested at younger ages and more frequently than girls without mental health system involvement. Copeland, Miller-Johnson, Keeler, Angold, and Costello (2007) identified a trend toward arresting girls with depressive and anxiety disorders. They found that “20.6% of female crime and 15.3% of male crime” was attributed to mental health disorders, with anxiety and depressive disorders being the strongest indicators of involvement in delinquent behavior by female offenders (p. 1673). This has led to the conclusion that girls are typically undertreated for their mental health needs (Cuffe, Waller, Cuccaro, Pumariega, & Garrison, 1995) and that this lack of treatment may increase their involvement in the juvenile justice system (Wasserman et al., 2005).

Current Study

This study aimed to expand on previous research examining the pathway youth take into and through the juvenile justice system by expanding the scope of data collection. The specific research questions were as follows: (a) Are juvenile justice involved girls placed in more severe out-of-home placements than boys for similar offenses? (b) Are juvenile justice system–involved girls with perceived or identified mental health needs placed in more severe out-of-home placements than boys with mental health needs? (c) Are girls who have accessed public mental health services placed in more severe out-of-home juvenile justice placements than boys who have accessed public mental health services? (d) When placed in out-of-home placements, do girls who are in need of, or have accessed mental health services, serve longer sentences than boys? and (e) when matched on similar demographic, offense, and mental health variables, do girls stay longer in out-of-home placement than boys?

Administrative data were collected from both the state agencies overseeing the juvenile justice (Department of Juvenile Justice [DJJ]) and public mental health systems (MHS) in a large state covering an 8-year time period. The study examined the following: decision-making at disposition points impacting girls, including mechanisms (trauma history, mental health need, and contact with the mental health system) leading to deeper involvement in the juvenile justice system (i.e., severity of out-of-home placement), and length of stay (i.e., length of confinement) in local correctional facilities.

Participants

The sample was extracted from a data set inclusive of all youth who were referred to the 165 local juvenile probation departments across the state (LJPD, N = 337,022) from January 1, 2007, to December 31, 2014. The sample for this study was restricted to youth who met the statutory age for delinquency referrals (10 through 17 years of age). Youth who were certified as an adult and transferred to the criminal justice system were also excluded. Remaining in the sample were youth referred to LJPDs and formally processed for delinquency or status offenses (n = 271,427).

Girls constituted 31.4% (n = 85,279) of the sample, while boys made up 68.6% (n = 186,148). A sample restricted to youth who were disposed to local placement facilities was also constructed. This “placement” sample (n = 25,976) allowed for the analysis of length of stay in out-of-home placement. Girls constituted 19.1% (n = 4,953) of the placement sample, while boys made up the remaining 80.9% (n = 21,023). Finally, Propensity Score Matching was used to match 4,830 girls to 4,830 of the boys in the placement sample to control for covariates in isolating the effects of gender on length of stay.

Study Measures

The main predictor variables included referral offense seriousness, gender, and mental health need. For this study, gender was a dichotomous static variable coded as male (0) or female (1) defined based on information recorded in the state’s data collection system by juvenile probation officers. However, because the variables of offense seriousness and level of mental health need could be interpreted with a broad array of values, specific operational definitions and categorical values for those variables were developed prior to conducting analyses.

Offense Seriousness

Youth can be referred to LJPDs for multiple offenses on multiple referral events. This study targeted offenses associated with the referral event that resulted in the most severe disposition during the sample period. The categorical coding guidelines identified within the DJJ data codebook were used as a guideline for establishing the operational definitions and assigning values for offense seriousness. The DJJ data codebook is used by juvenile probation officers collecting and entering data into the state’s data collection system. Using this coding strategy reduced some potential threats to reliability in the study. The first set of recodes categorized the 4,019 types of offenses contained in the DJJ codebook into a continuous classification variable ranging from the least serious, 1 (status offenses), to the most severe, 8 (capital felony), referred to hereafter as the offense composite measure.

“Bootstrapping” has been defined in the literature as the process of courts detaining girls through findings of contempt of court, VOPs, or violations of court orders for underlying status offenses or minor delinquent behavior (Sherman, 2005). To capture the potential impact of bootstrapping, a second group of offense recodes were undertaken, resulting in the creation of two dichotomous indicator variables. A traditional bootstrap variable (Status) included offenses typically categorized as status offenses (otherwise known as “Conduct Indicating Need for Supervision Offenses” or CINS offenses), Class C misdemeanors, and contempt of court referrals. Class C misdemeanors are typically violations of city or county ordinances and are processed similarly to status offenses. The Status indicator includes activities such as runaway, truancy, and curfew violations. A second bootstrap variable (VOP) was developed to allow for independent analysis of a VOP as a potential predictor of the outcomes under study.

Mental Health Need

A survey conducted in 2015 indicated the Massachusetts Youth Screening Inventory: Version 2 (MAYSI-2; Grisso, 2004) was the most widely used mental health screening instrument used by juvenile justice systems in the United States (Wachter, 2015) with 42 states mandating its use in at least one juvenile justice setting (probation, detention, or corrections). The MAYSI-2 is a 52-item, self-report screening instrument completed by youth ages 12–17 upon intake into the juvenile justice system. The MAYSI-2 contains seven factor-analytically derived subscales: Alcohol and Drug Use, Angry-Irritable, Depressed-Anxious, Somatic Complaints, Suicidal Ideation, Thought Disturbance, and Traumatic Experiences. Studies have demonstrated good concurrent validity when comparing MAYSI-2 scales with scores on other mental health scales. With variances on specific items and scales, the instrument has shown evidence of reliability, clinical utility, and validity across multiple juvenile justice settings (Cauffman, 2004; Goldstein et al., 2003; Grisso & Barnum, 2006; Wasserman et al., 2005). More specifically, using different self-administered diagnostic interviews and questionnaires, several studies have determined the MAYSI-2 subscales to have moderate diagnostic utility (Kerig, Moeddel, & Becker, 2011; Kuo, Stoep, & Stewart, 2005; Wasserman et al., 2005). Test-retest reliability when administering the MAYSI-2 up to 8 days later was moderate to good, ranging from 0.53 to 0.89 (Grisso & Barnum, 2006).

Cut-off scores for each of the MAYSI-2 subscales have been developed, identifying youth scoring higher than 90% of the normative sample on the subscales (Grisso & Barnum, 2006). Each subscale was classified separately as elevated/warning (1) or not elevated (0). Overall mental health need was defined as the total number of subscales reaching this “warning” cutoff for five of the six scales and thus ranged from 0 to 5. Thought Disturbance was excluded as it was valid only for boys; the Traumatic Experiences subscale does not have established warning cutoffs, so the subscale was kept in the original format, with a scoring range of 0–5 representing the number of traumatic experiences endorsed by the youth.

Two additional mental health indicators were included in the analyses. The first indicator, behavioral health referral, was extracted from the DJJ database. This indicator is scored by trained juvenile probation officers and indicates they perceived that the youth had a mental health need warranting further assessment by a behavioral health specialist. The second, involvement with the public mental health system, was extracted from the state’s mental health-care system (MHS) and included youth who were matched in both the MHS and DJJ data sets. Each of the variables were coded as dichotomous indicators where 0 = no and 1 = yes.

Outcome Measures

Out-of-home placement was divided into several categories based on the severity of placement. These categorizations reflect not only a determination of whether the facility was secure or nonsecure but also consideration of what intercept point in the juvenile justice system (pre- or postdisposition) the youth was placed within the facility. Although often a secure setting, juvenile detention facilities are the first type of facility (pre and/or adjudication) within the juvenile justice system into which a juvenile may be placed and therefore constituted the first outcome measure. Detention is a dichotomous outcome measure, with “1” indicating youth had been placed and “0” indicating no placement. The outcome measure consisted of a postadjudication facility placement composite, with county nonsecure (1), county secure (2), and state correctional facility (3). Those who were not placed in a postadjudication facility (0) served as the reference category in the analyses. In addition to facility placement, number of days spent in confinement was calculated for those in the local, county-level placement sample.

Control Measures

While their most recent contact with the juvenile justice system constituted the basis for previously described measures, numerous control measures were included from the current referral as well as prior referrals. Control measures included age at first referral, age at target referral, race, ethnicity, and variables related to prior contact with DJJ. Variables extracted from prior involvement with the juvenile justice system include the number of prior referrals, prior offense history, and prior number of probated dispositions. Prior offense history is based on the same 8-point scale relied on to measure the severity of current referral offense and refers to the most serious prior offense on record.

Analysis

Initial differences in the variables of interest and control variables by gender were examined through contingency analyses. Due to the large size of the samples, effect sizes (η for continuous variables and φ for categorical comparisons) are presented rather than p values. The standard rule for the correlation coefficients presented is a threshold of .10 for observed relationships to be considered a “small” as opposed to “negligible” effect size (Cohen, 1992). Multivariate analyses are used to examine the severity of out-of-home placement by gender, mental health need, and trauma history. Analyses examined the influence of the predictor variables by specific facility type and gender at pre- and postdisposition intercepts. Separate analyses were completed to allow for an examination of any interaction effects by gender. Binary logistic regression was used to model the binary detention decision outcome. Multinomial logistic regression was used to model postadjudicatory placement, including separate panels to describe the influence of predictors on placement in nonsecure facilities, county-operated secure facilities, and state correctional facilities. Alternatives to placement served as the reference category for the multinomial comparisons. Detention was included as a predictor variable in the postadjudication placement model to account for potential indirect effects of predictors on postadjudication placement type.

Additional analyses examined the influence of the predictor variables by specific facility type and gender at pre- and postadjudication intercepts, requiring separate models for boys and girls. Separate models allowed for an examination of any interaction effects by gender. Binary logistic regression was used to model the binary detention decision outcome. Multinomial logistic regression was used to model postadjudicatory placement, including separate panels to describe the influence of predictors on placement in nonsecure facilities, county-operated secure facilities, and state correctional facilities. Alternatives to placement served as the reference category for the multinomial comparisons. Detention was included as a predictor variable in the postadjudication placement model to account for potential indirect effects of predictors on postadjudication placement type.

Additional analysis was conducted to test for differences between regression coefficients of the gendered models. The formula suggested by Brame, Paternoster, Mazerolle, and Piquero (1998) was used to test for differences between the model coefficients:

Z = \frac{b_{1} - b_{2}}{\sqrt{S E b_{1}^{2} + S E b_{2}^{2}}} Z = b_{1} .

Survival analysis was used to predict the time to discharge from custody for youth adjudicated to out-of-home placement in the restricted placement sample. Used to model time to an event, survival analysis is typically employed on a sample that includes some “survivors,” that is, those for whom the outcome event had not yet occurred by the end of the observation period (Luke & Homan, 1998). Observations for the survivors are then “censored” at the end of the observation period to indicate survival up to the point that the observation period closed. This study included only youth who had discharged from placement, leaving no need for censoring. Its relaxed assumptions about the shape of the outcome distribution and straightforward interpretation made survival analysis the most appropriate procedure for this analysis. As with the other analyses, proportional hazard models were estimated using Cox regression for the entire placement sample and then separately for girls and boys. Finally, Propensity Score Matching (PSM) was utilized to isolate the effect of gender on time served in local correctional facilities.

Results

Table 1 presents placement outcomes, offense seriousness, mental health need, and control variables by gender. Regarding placement outcomes, the correlation coefficients indicate girls were less likely than boys to be placed in detention and secure county placements. Overall, the findings show that boys generally received more severe out-of-home placement dispositions than girls. This is consistent with the findings that girls, on average, commit less serious offenses. The moderate correlation between the overall measure of offense seriousness and gender supports this supposition. Further, mean scores indicate that girls were more than twice as likely as boys to have been referred to juvenile court for status offenses. Measures of prior criminal behavior buttress the argument that girls referred to juvenile courts had a less serious prior record overall, and just over half of the average number of prior referrals and probated dispositions in relation to boys.

Table 1.

Level of Placement, Offense Seriousness, Mental Health Need, and Control Variables by Gender.

Variables	FemaleMean (SD)(n = 85,279)	MaleMean (SD) (n = 186,148)	Effect Size
Level of placement
Detention	0.32	0.44	φ = −.12*
Nonsecure placement	0.03	0.04	φ = −.01*
County secure placement	0.03	0.08	φ = −.10*
State correctional commitment	0.01	0.04	φ = −.08*
Offense seriousness
Severity of offense	2.59 (1.23)	3.25 (1.56)	η = −.21*
Status offense (Boot 1)	0.20	0.08	φ = .17*
VOP (Boot 2)	0.06	0.09	φ = −.05*
Mental health variables
MH need (total # Warnings)	0.47 (0.90)	0.22 (0.61)	η = .16*
Drug	0.02	0.02	φ = .01*
Angry	0.11	0.06	φ = .08*
Depressed	0.09	0.04	φ = .11*
Somatic	0.08	0.04	φ = .10*
Suicide	0.17	0.07	φ = .16*
Trauma score	1.21 (1.40)	1.05 (1.25)	η = .06*
Behavioral health referral	0.16	0.20	φ = −.05*
Contact with MHS	0.15	0.15	φ = −.00
Control variables
Race—Black	0.22	0.22	φ = .00
Ethnicity—Hispanic	0.46	0.47	φ = −.01*
Age at referral	15.34 (1.29)	15.39 (1.37)	η = −.02*
Age of onset	14.97 (1.40)	14.78 (1.55)	η = .06*
Prior referrals	0.59 (1.43)	1.00 (1.99)	η = −.10*
Prior offense history	0.76 (1.46)	1.36 (2.04)	η = −.15*
Prior probated dispositions	0.16 (0.56)	0.33 (0.84)	η = −.11*

Note. t tests were computed to test mean differences, and χ ² was used to test percentage differences. VOP = violation of probation; MHS = mental health system.

*p < .001.

Generally, girls had a higher level of mental health need, as evidenced by more than twice the total number of warnings on the MAYSI-2 and a similar disparity between the percentage of girls meeting the criteria for elevated levels of depression and suicidal thoughts. There was no difference, however, between genders for youth involved in the juvenile justice system who also had some involvement with the public mental health system, with 15% of both boys and girls represented as dual-system involved. The following sections focus on determining whether gender differences persisted, while simultaneously controlling for alternate predictors, during the processing of cases through the juvenile justice system.

Detention Decision

The logistic regression analyses in Table 2 show the coefficients for predictors were larger for girls than boys on average, indicating greater discernment overall in the detention decision for girls relative to boys. For criminogenic variables, the test of difference indicated that offense severity, offense history and prior probation increased the odds of detention more for girls than for boys, as well as the bootstrap variables, VOP and status offense. Girls with trauma histories and behavioral health referrals were more likely to be detained than boys. However, there was very little difference in the impact of being engaged in the public mental health system by gender. In fact, being involved in the public mental health system was a significant predictor of detention for both boys and girls, second only to having a VOP in terms of the strength of its relationship with placement in detention facilities.

Table 2.

Predictor Variables Regressed on Detention Decision by Gender.

Variables	Female		Male		Test of Difference
Variables	OR	SE	OR	SE	Test of Difference
Severity of offense	1.71*	.01	1.53*	.00	11.76*
Status offense (Boot 1)	0.59*	.03	0.39*	.03	9.67*
VOP (Boot 2)	9.10*	.06	7.41*	.03	2.89*
Mental health need	1.11*	.01	1.15*	.01	−2.79*
Trauma score	1.19*	.01	1.07*	.01	11.00*
Behavioral health referral	1.76*	.02	1.42*	.01	8.35*
Contact with MHS	1.77*	.02	1.72*	.02	0.80
Race—Black	1.10*	.02	1.13*	.02	−1.07
Ethnicity—Hispanic	0.90*	.02	1.05*	.01	−7.25*
Age at target referral	0.94*	.02	1.05*	.01	−5.40*
Age of first referral	1.06	.01	1.02**	.01	2.10
Prior offense history	1.48*	.01	1.34*	.00	10.50*
Prior probated dispositions	1.50*	.03	1.31*	.01	4.47*
Constant	.05*	.11	.03*	.07
Model Pseudo R²	.38*		.36*

Note. VOP = violation of probation; MHS = mental health system.

*p < .001.

Postadjudication placement

The multinomial logistic regression model in Table 3 illustrates the influence of predictors on specific types of postadjudication placement (relative to no placement, the reference category) for girls and boys. The indicator for status offenses had to be dropped from this model due to zero cells, as placement in state juvenile correctional facilities for status offenses is prohibited by law. The largest coefficient in any of the models (nonsecure, county-operated secure, or state correctional) was VOP on state correctional facility commitment, suggesting that VOP may have been a way of bootstrapping boys and girls into the state correctional system.

Table 3.

Multinomial Logistic Regression Model Predicting Postadjudication Placement by Gender.

Nonsecure Placement	Female		Male		Test of Difference
Nonsecure Placement	OR	SE	OR	SE	Test of Difference
Severity of offense	1.13*	.02	1.25*	.01	−5.00*
VOP (Boot 2)	2.03*	.06	2.44*	.04	−2.35
Mental health need	1.05*	.02	1.12*	.02	−2.38
Trauma score	1.13*	.02	1.08*	.01	2.76*
Behavioral health referral	1.66*	.04	2.00*	.03	−3.65*
Contact with MHS	2.38*	.04	1.92*	.03	4.08*
Race—Black	1.02	.05	0.76*	.04	4.71*
Ethnicity—Hispanic	0.81*	.05	0.80*	.03	0.08
Age at target referral	0.97	.03	0.96	.02	0.16
Age of first referral	1.06	.03	0.99	.01	2.03
Prior offense history	1.13*	.01	1.18*	.01	−2.75*
Prior probated dispositions	2.16*	.03	1.75*	.02	5.92*
Detention	5.14*	.06	2.82*	.03	8.49*
County-Operated Secure Placement
Severity of offense	1.32*	.02	1.32*	.01	−0.05
VOP (Boot 2)	2.82*	.07	2.00*	.03	4.48*
Mental health need	0.96	.03	1.07	.02	−3.30*
Trauma score	1.13*	.02	1.02	.01	5.21*
Behavioral health referral	3.01*	.05	2.05*	.02	6.67*
Contact with MHS	1.58*	.05	1.98*	.02	−3.85*
Race—Black	0.90	.07	1.49*	.03	−7.23*
Ethnicity—Hispanic	0.92	.06	1.28*	.03	−5.34*
Age at referral	0.97	.03	0.93*	.01	1.27
Age of onset	0.98	.03	1.12*	.01	−4.17*
Prior offense history	1.18*	.02	1.20*	.01	−0.44
Prior probated dispositions	2.38*	.03	2.60*	.01	−2.57*
Detention	4.26*	.07	3.58*	.03	2.29
State Correctional Commitment
Offense composite	2.59*	.05	2.34*	.01	1.63
VOP (Boot 2)	12.49*	.17	8.84*	.05	1.89
Warnings	1.10	.05	1.21*	.02	−1.94
Trauma score	1.23*	.03	1.10*	.01	3.31*
Behavioral health referral	0.97	.10	0.83*	.03	1.42
Contact with MHS	2.52*	.10	1.93*	.03	2.66*
Race—Black	0.89	.12	1.32*	.04	−3.60*
Ethnicity—Hispanic	0.59*	.12	0.98	.04	−4.59*
Age at referral	0.90	.05	0.98	.02	−1.60
Age of onset	0.95	.05	0.99	.01	−0.94
Prior offense history	1.64*	.03	1.48*	.01	3.52*
Prior probated dispositions	3.29*	.04	3.27*	.02	0.08
Detention	2.08*	.15	2.08*	.04	−0.01
Model Pseudo R²	.38*	.42*

Note. The reference category for postadjudication placements is the alternative, “no placement.” VOP = violation of probation; MHS = mental health system.

*p < .001.

Nonsecure placement

The strongest predictor of whether a youth was placed in a nonsecure placement was having previously been placed in detention. Girls with a previous detention were almost 2 times more likely to be placed in nonsecure placement than boys. Involvement with the public mental health system was the second strongest predictor for girls but the fourth strongest predictor for boys. Other mental health variables increased the odds of being placed in nonsecure confinement for both girls and boys, as did severity of offense and VOP.

County-operated secure placement

Similar to the nonsecure placement decision, the strongest predictor of placement in a county-operated secure placement for boys and girls was having been previously detained. Referral for behavioral health services and contact with MHS also were strong predictors for both. Most pronounced for girls was the impact of a behavioral health referral by the youth’s juvenile probation officer, an effect compounded by its influence on the earlier detention decision. VOP also increased the likelihood of placement in a secure county-operated facility. This relationship was more prominent for girls than for boys, having the net effect of increasing the odds of placement by nearly 3 times for girls versus two times for boys.

State correctional commitment

Unlike nonsecure placement and secure county-operated placement, detention was not among the strongest predictors of commitment to a state correctional facility. Further, there was little variation between genders relating to the predictors of state correctional commitment. The strongest predictor was VOP, which increased the odds of out-of-home placement in a state-run facility by 12½ times for girls versus close to 9 times for boys. A history of traumatic experiences and prior involvement in the public mental health system had a stronger influence on girls being committed to state correctional settings relative to boys.

Detention decision, indirect effects, and bootstrapping

While detention was associated with out-of-home placement for boys and girls, its influence was more pronounced for girls than boys sent to nonsecure placements. This suggests that gender-specific indirect effects resulting from detention decisions were amplified in local postadjudication placement decisions. While the direct effect of bootstrapping girls into facilities based on status offenses cannot be estimated due to its exclusion from the multinomial models, the data indicate it had an indirect effect through the detention decision. The other bootstrapping variable, VOP, retained a larger effect for girls relative to boys on placement in county-operated secure and state correctional facilities, indicating both direct and indirect effects acted to increase gender disparities in confinement. Similarly, traumatic experiences, a variable with the second largest differential impact by gender on the detention decision, continued to differentially impact each of the postadjudication placement decisions, thereby amplifying gender disparities throughout the system of case processing. These findings suggest girls with a history of involvement in the public mental health system who commit lesser offenses, particularly status offenses and VOP, are confined more often than boys both prior to, and after, adjudication.

Time Served in Local Postadjudicatory Secure Placements

The next step in the analysis involved estimating proportional hazard (Cox regression) models to determine the extent of gendered relationships between predictors and length of stay in local correctional placements. The hazard rate, or hazard function, predicted by the model is simply the probability that a given youth will discharge from secure county-level placements during the next time interval (day) having been held in placement up to that point in time. Table 4 presents separate Cox regression model for girls and boys in an effort to identify suspected interaction effects.

Table 4.

Cox Regression Model Predicting Weeks to Discharge From Local Postadjudication Placement by Gender.

Predictors	Female		Male		Test of Difference
Predictors	OR	SE	OR	SE	Test of Difference
Severity of offense	1.00	.01	0.89*	.01	7.67*
Status offense (Boot 1)	1.18	.09	0.96	.07	1.80
VOP (Boot 2)	1.05	.04	0.79*	.02	5.92*
Mental health need	0.99	.02	1.00	.01	0.58
Trauma score	0.97*	.01	0.99	.01	−1.33
Behavioral health referral	1.01	.03	0.96	.01	1.72
Contact with MHS	0.92*	.03	0.99	.02	−2.34
Race—Black	1.12	.04	1.26*	.02	−2.51
Ethnicity—Hispanic	1.11	.04	1.20*	.02	−1.97
Age at referral	1.15*	.02	1.20*	.01	−1.73
Age at first referral	1.00	.02	1.03*	.01	−1.63
Prior offense history	0.99	.01	0.99*	.00	0.40
Prior probated dispositions	1.02	.02	1.03*	.01	−0.44
Pretrial detention	0.97	.04	0.94*	.02	0.69
Secure facility placement	0.80*	.03	1.04	.02	−7.34*
	χ² = 222.18*		χ² = 1,877.91*

Note. VOP = violation of probation; MHS = mental health system.

*p < .001.

Most notably, fewer coefficients were significant in the girls’ model. Among the differences, offense seriousness and VOP tended to effectuate the release of boys in a manner to be expected but did not influence the pace of release for girls. Boys were released from confinement at a slower pace than girls as the severity of their offenses increased and for VOP; however, these legally relevant offense variables did not influence the rate of release for girls. Traumatic experiences and contact with MHS, however, were found to significantly decrease the rate of discharge for girls. Finally, boys were discharged from secure facility placement much quicker than girls. Specifically, being held in county-operated secure facilities reduced girls’ odds of being discharged during a given interval decreased by 20%.

PSM was applied to assess the final research question on the differential impact of gender on time to discharge. This method estimates the conditional probability of selection to a particular group based on the observed covariates (Austin, 2014). Once estimated, propensity scores are then used to match individuals to create equal groups. PSM was conducted in R 3.6.1 using the package “MatchIt” 3.0.2 (Ho, King, Stuart, & Imai, 2011; R Core Team, 2017). A “Greedy” matching procedure was used that utilizes a without-replacement method to match girls with boys based on the nearest neighbor score within a caliper of .1. Ultimately, of the 4,945 girls who were confined, 115 girls were not matched, resulting in a final sample of 4,830 girls matched to 4,830 boys to minimize any observed differences between the two groups. T tests and χ² tests shown in Table 5 confirmed the matching procedure was efficient as the groups did not significantly differ on any of the 14 covariates.

Table 5.

Comparison of Covariate Means and Frequencies for Females (n = 4,830) and Males (n = 4,830) Among Unmatched and Matched Samples.

Predictors	Sample	FemalesM / %	MalesM / %	t Test/χ2p Value
Severity of offense	Unmatched	3.51	2.91	<.001
Severity of offense	Matched	2.93	2.95	.478
Status offense (Boot 1)	Unmatched	3.3%	1.0%	<.001
Status offense (Boot 1)	Matched	2.9%	3.0%	.856
VOP (Boot 2)	Unmatched	37.4%	30.0%	<.001
VOP (Boot 2)	Matched	37.7%	37.8%	.933
Mental health need	Unmatched	0.66	0.31	<.001
Mental health need	Matched	0.61	0.60	.862
Trauma score	Unmatched	1.58	1.09	<.001
Trauma score	Matched	1.53	1.51	.601
Behavioral health referral	Unmatched	47.9%	45.8%	.007
Behavioral health referral	Matched	47.6%	48.7%	.290
Contact with MHS	Unmatched	41.9%	33.1%	<.001
Contact with MHS	Matched	41.2%	40.2%	.310
Race—Black	Unmatched	27.5%	27.8%	.707
Race—Black	Matched	27.6%	26.7%	.293
Ethnicity—Hispanic	Unmatched	44.1%	48.5%	<.001
Ethnicity—Hispanic	Matched	44.4%	45.2%	.449
Age at referral	Unmatched	15.51	15.61	<.001
Age at referral	Matched	15.52	15.51	.844
Age at first referral	Unmatched	14.40	14.27	<.001
Age at first referral	Matched	14.40	14.39	.832
Prior offense history	Unmatched	2.51	3.17	<.001
Prior offense history	Matched	2.55	2.56	.807
Prior probated dispositions	Unmatched	1.04	1.17	<.001
Prior probated dispositions	Matched	1.05	1.04	.612
Pretrial detention	Unmatched	85.9%	84.8%	.052
Pretrial detention	Matched	85.9%	85.2%	.297
Secure facility placement	Unmatched	42.8%	66.2%	<.001
Secure facility placement	Matched	43.8%	43.3%	.608

Note. 16,132 boys were not matched; 115 girls were not matched. VOP = violation of probation; MHS = mental health system.

Based on the Welch’s two sample t test, girls (M = 174.8) are held significantly longer than boys (M = 166.0), t = −3.61, p < .001, Cohen’s d = .08. These results were confirmed via the survival analysis procedure in SPSS, with a Kaplan–Meier (M = 174.8 v. 165.7, Mantel-Cox χ² = 4.30, p = .038) and by bootstrapping in Cox regression (OR = 1.043, Wald = 4.26, p = .039).

Discussion

This study extends previous research on the pathway girls take to, and through, the juvenile justice system. In particular, this study supports prior findings related to the influence of mental health and trauma on out-of-home placement decisions and length of stay in county-operated secure facilities. Interestingly, youth who had accessed the public mental health system only accounted for 15% of the total sample (equal across genders), yet the models indicated contact with the public mental health system (MHS) was a consistent predictor of severity of placement for both boys and girls.

Being identified by their probation officer as needing a behavioral health referral, VOP, and prior probation history were also influential predictors of placement severity. Research has shown that juveniles with mental health needs have higher risk of juvenile justice involvement (Vander Stoep, Evans, & Traub, 1997) and are less successful under probation supervision than those without mental health needs (Espinosa et al., 2013; Monahan et al., 2005; Skeem et al., 2006). These results expand those findings further by suggesting youth who have accessed public mental health services have a higher likelihood of being detained, being removed from the home, and placed in more severe out-of-home placements than youth without mental health needs.

While being placed in detention was the strongest predictor for out-of-home placement in nonsecure and county-operated secure placements, the results suggest that interaction with the public mental health system played a stronger role in placement decisions for girls than boys. For girls, involvement with the public mental health system was the second strongest influence on nonsecure residential placement, while being identified by their probation officer as needing a behavioral health referral was the second strongest predictor of being sent to secure county-operated facilities. Meanwhile, girls had an increased length of stay when placed in county-operated secure placements compared with boys. VOP was the strongest predictor of placement in state correctional institutions. Overall, the findings suggest that when public mental health services are not successful—or available—the juvenile justice system may rely on VOPs as a way to bootstrap youth with mental health needs into out-of-home placements, and once placed, girls stay longer than boys.

Limitations

Although this study represents an important step in the study of the interaction of gender, trauma, mental health, and juvenile justice case processing, several methodological limitations should be noted. One limitation results from the reliance on available measures in the state-level juvenile justice and public mental health databases. Variables that may have influenced court decisions were not available. Specifically, the influence of gender-specific measures, such as evidence of prior dysfunctional relationships and/or lack of family support, which may have differentially influenced decisions made in the cases of girls before the court could not be modeled.

Also, in examining the influence of VOP, the source of the violation resulting in the VOP was not included in the data set. For the purposes of interpretation and discussion, VOP was treated as technical violations as defined by the DJJ data codebook. VOPs in response to subsequent offending behavior or delinquent conduct that was not adjudicated could not be examined. Future study should examine the relationship between VOPs that occur due to technical violations in comparison to VOPs that occur due to nonadjudicated delinquent conduct or offending behavior.

Finally, measures of the relationship between mental health and juvenile justice case processing were limited to a screening instrument, probation officer referrals, and contact with the state mental health system. Specific mental health diagnoses and other follow-up information/outcome measures were not included in the analysis. Relatedly, no attempt was made to unravel the relationship between contact with the mental health and juvenile justice systems over time. Studies that include more refined mental health measures and longitudinal analyses are warranted.

Research and Policy Implications

This study is one of the first to examine the interaction between public mental health system involvement and juvenile justice system processing. The results suggest that accessing mental health services in the community or being viewed as needing a mental health service by a juvenile probation officer increases the likelihood of youth being removed from their homes, especially for girls. While girls generally had decreased odds of out-of-home placement compared with boys, when combined with involvement in public mental health and receiving a VOP, those odds significantly increased.

The link between out-of-home placement and mental health need may occur because those settings are often viewed as treatment options for juveniles (Underwood, Warren, Talbott, Jackson, & Dailey, 2014). However, there is general agreement the juvenile justice system, especially within secure institutional settings, is not the appropriate vehicle for delivering mental health care to youth (Cocozza & Skowyra, 2000; Skowyra & Cocozza, 2007). Investigations by the U.S. Department of Justice have questioned the ability of juvenile justice institutions to adequately respond to and treat the mental health needs of youth in their custody (U.S. Department of Justice, 2005).

Research on outcomes following out-of-home placement for youth with mental health challenges show not only cost increases to the system but also patterns of ongoing system involvement, an added disruption to adolescent development (Hoagwood & Cunningham, 1992; Lyons, Libman-Mintzer, Kisiel, & Shallcross, 1998). Moreover, there is evidence that treating the mental health needs without treating the criminogenic needs is ineffective for impacting later reoffending (Guebert & Olver, 2014). Recidivism studies indicate the rates of rearrest for juvenile offenders who have returned from residential treatment and/or juvenile correctional settings range from 40% (Taylor, Kemper, Loney, & Kistner, 2009) and 65% (Benda, Corwyn, & Toombs, 2001) to as high as 85% (Trulson, Marquart, Mullings, & Caeti, 2005). When youth return to the community from a juvenile justice placement, including placements with mental health treatment, there is a very high likelihood that they will cycle back through the system or become engaged in the adult criminal justice system.

Youth with mental health needs are less likely to successfully complete community supervision and have higher rates of out-of-home placement, longer stays in juvenile detention, and higher rates of involvement in the adult criminal justice system than youth without such conditions (Eno Louden, Skeem, Camp, & Vidal, 2012; Espinosa et al., 2013; Manchak, Skeem, & Rock, 2013, Vaughn, Salas-Wright, DeLisi, Maynard, & Boutwell, 2015). In-home dispositions allow youth access to community-based treatments while remaining in their home and are less likely to inhibit autonomy and psychosocial development (Dmitrieva, Monahan, Cauffman, & Steinberg, 2012; Ryon, Early, Hand, & Chapman, 2013).

Yoder et al. (2017) found that receiving a mental health screening was associated with a reduced likelihood of future arrest and incarceration. However, studies indicate justice involved youth are not accessing mental health services when juvenile justice practitioners identified a possible mental health need. Teplin, Abram, McClelland, Washburn, and Pikus (2005) found less than 30% of the youth identified as needing services received them. In a later study, Teplin, Welty, Abram, Dulcan, and Washburn (2012) showed almost half of boys and one third of girls continued to struggle with mental health needs 5 years after being removed from the home. A survey of youth and caregivers identified the time period immediately after reentry from detention as a particularly critical time to access treatment and supports (Aalsma, Brown, Holloway, & Ott, 2014). Youth and caregivers reported being open to mental health services and supports and emphasized the importance of probation services and courts having consistent information regarding access and participation in those services. This suggests the fragmentation in continuity of care may likely be due to lack of clear communication and referral protocols or processes between juvenile justice and community mental health providers (Colwell, Villarreal, & Espinosa, 2012).

Future research should include longitudinal processs and outcome evaluations to examine how juvenile justice and community providers respond to juveniles with mental health needs within their communities over time. Researchers of youth with deep juvenile justice system involvement have suggested integrating mental health systems of care and juvenile justice systems should be a core strategy of juvenile justice reform (Schubert & Mulvey, 2014). Shufelt, Cocozza, and Skowyra (2010) and others (Skowyra & Cocozza, 2007) indicated a need for a more balanced solution whereby juvenile justice and mental health systems are partners in all efforts to identify and respond to youth mental health needs.

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.

ORCID iD

Erin M. Espinosa

References

Aalsma

Brown

Holloway

Ott

(2014). Connection to mental health care upon community reentry for detained youth: A qualitative study. BMC Public Health, 14, 1–8.

Abram

Teplin

Charles

Longworth

McClelland

Dulcan

M. K.

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

Abram

Washburn

Teplin

Emanuel

Romero

McClelland

(2007). Posttraumatic stress disorder and psychiatric comorbidity among detained youth. Psychiatric Services, 5, 1311–1316.

American Bar Association and the National Bar Association. (2001). Justice by gender: The lack of appropriate prevention, diversion and treatment alternatives for girls in the justice system. Washington, DC: Author.

Austin

P. C.

(2014). The use of propensity score methods with survival or time-to-event outcomes: Reporting measures of effect similar to those used in randomized experiments. Statistics in Medicine, 33, 1242–1258.

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.

Hardt

N. S.

(2014). The prevalence of Adverse Childhood Experiences (ACE) in the lives of juvenile offenders. Journal of Juvenile Justice, 3, 1–23.

Benda

B. B.

Corwyn

R. F.

Toombs

N. J.

(2001). Recidivism among adolescent serious offenders: Prediction of entry into the correctional system for adults. Criminal Justice and Behavior, 28, 588–613.

Bishop

D. M.

Frazier

C. E.

(1992). Gender bias in juvenile justice processing: Implications of the JJDP Act. Journal of Criminal Law and Criminology, 82, 1162–1187.

10.

Bishop

D. M.

Frazier

C. E.

(1996). Race effects in juvenile justice decision-making: Findings of a statewide analysis. The Journal of Criminal Law and Criminology, 86, 392–414.

11.

Brame

Paternoster

Mazerolle

Piquero

(1998). Testing for the equality of maximum-likelihood regression coefficients between two independent equations. Journal of Quantitative Criminology, 14, 245–261.

12.

Cauffman

(2004). A state-wide screening of mental health symptoms among juvenile offenders in detention. Journal of the American Academy of Child and Adolescent Psychiatry, 43, 430–439.

13.

Cauffman

Feldman

Waterman

Steiner

(1998). Posttraumatic stress disorder among female juvenile offenders. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 1209–1216.

14.

Chamberlain

Moore

K. J.

(2002). Chaos and trauma in the lives of adolescent females with antisocial behavior and delinquency. Journal of Aggression, Maltreatment & Trauma, 6, 79–108.

15.

Chesney-Lind

(1977). Judicial paternalism and the female status offender: Training women to know their place. Crime & Delinquency, 23, 121–130.

16.

Chesney-Lind

Shelden

R. G.

(1998). Girls, delinquency, and juvenile justice (2nd ed.). Belmont, CA: West/Wadsworth.

17.

Cocozza

J. J.

Skowyra

(2000). Youth with mental health disorders: Issues and emerging responses. Juvenile Justice Journal, 7, 3–13

18.

Cohen

. (1992). A power primer. Psychological Bulletin, 112, 155–159.

19.

Colwell

Villarreal

Espinosa

. (2012). Preliminary outcomes of a pre-adjudication diversion initiative for juvenile justice involved youth with mental health needs in Texas. Criminal Justice and Behavior, 39, 444–460.

20.

Conrad

Tolou-Smans

Marina

Christie

Placella

Brown

(2014). Gender differences in recidivism rates for juvenile justice youth: The impact of sexual abuse. Law and Human Behavior, 38, 305–314.

21.

Copeland

W. E.

Miller-Johnson

Keeler

Angold

Costello

E. J.

(2007). Childhood psychiatric disorders and young adult crime: A prospective, population-based study. The American Journal of Psychiatry, 164, 1668–1675.

22.

Cuffe

Waller

Cuccaro

Pumariega

Garrison

(1995). Race and gender differences in the treatments of psychiatric disorders in young adolescents. Journal of American Academy of Child and Adolescent Psychiatry, 34, 1536–1543.

23.

Davis

Fisher

Gershenson

Grudzinskas

Banks

(2009). Justice system involvement into young adulthood: Comparison of adolescent girls in the public mental health system and in the general population. American Journal of Public Health, 99, 234–236.

24.

Dmitrieva

Monahan

Cauffman

Steinberg

(2012). Arrested development: The effects of incarceration on development of psychosocial maturity. Development and Psychopathology, 24, 1073–1090.

25.

Eno Louden

Skeem

Camp

Vidal

(2012). Supervision practices in specialty mental health probation: What happens in officer-probationer meetings? Law and Human Behavior, 36, 108–119.

26.

Espinosa

E. M.

Sorensen

J. R.

(2015). The influence of gender and traumatic experiences on length of time served in juvenile justice settings. Criminal Justice & Behavior, 43, 187–203.

27.

Espinosa

E. M.

Sorensen

J. R.

Lopez

M. A.

(2013). Youth pathways to placement: The influence of gender, mental health need and trauma on confinement in the juvenile justice system. Journal of Youth and Adolescence, 42, 1824–1836.

28.

Ford

J. D.

Chapman

Connor

D. F.

Cruise

K. R.

(2012). Complex trauma and aggression in secure juvenile justice settings. Criminal Justice and Behavior, 39, 694–724.

29.

Ford

J. D.

Chapman

Mack

J. M.

Pearson

(2006). Pathways from traumatic child victimization to delinquency: Implications for juvenile and permanency court proceedings and decisions. Juvenile and Family Court Journal, 57, 13–26.

30.

Ford

J. D.

Hartman

J. K.

Hawke

Chapman

J. F.

(2008). Traumatic victimization, posttraumatic stress disorder, suicidal ideation, and substance abuse risk among juvenile justice-involved youths. Journal of Child & Adolescent Trauma, 1, 75–92.

31.

Gavazzi

S. M.

Yarcheck

C. M.

Chesney-Lind

(2006). Global risk indicators and the role of gender in a juvenile detention sample. Criminal Justice and Behavior, 33, 597–612.

32.

Goldstein

Arnold

Weil

Mesiarik

Peuschold

Grisso

… Osman

(2003). Comorbid symptom patters in female juvenile offenders. International Journal of Law and Psychiatry, 26, 565–582.

33.

Goodkind

Sarri

(2006). The impact of sexual abuse in the lives of young women involved or at risk of involvement in the juvenile justice system. Violence Against Women, 12, 456–477.

34.

Grisso

(2004). Double jeopardy: Adolescent offenders with mental disorders. Chicago, IL: University of Chicago Press.

35.

Grisso

Barnum

(2006). Massachusetts Youth Screening Instrument—Version 2: User’s manual and technical report. Sarasota, FL: Professional Resource Press.

36.

Guebert

A. F.

Olver

M. E.

(2014). An examination of criminogenic needs, mental health concerns, and recidivism in a sample of violent young offenders: Implications for risk, need, and responsivity. International Journal of Forensic Mental Health, 13, 295–310.

37.

D. E.

King

Stuart

E. A.

Imai

(2011). MatchIt: Nonparametric preprocessing for parametric causal inference. Journal of Statistical Software, 42, 1–28.

38.

Hoagwood

Cunningham

(1992). Outcomes of children with emotional disturbance in residential treatment for educational purposes. Journal of Child and Family Studies, 1, 129–140.

39.

Hockenberry

Sickmund

(2014). Juveniles in residential placement, 2011. Retrieved from http://www.ojjdp.gov/pubs/246826.pdf

40.

Joe

K. A.

Chesney-Lind

(1995). Just every mother’s angel: An analysis of gender and ethnic variations in youth gang membership. Gender and Society, 9, 408–430.

41.

Kempf-Leonard

Sontheimer

(1995). The role of race in juvenile justice processing in Pennsylvania. In Kempf-Leonard

Pope

Feyerherm

(Eds.), Minorities in juvenile justice (pp. 98–127). Thousand Oaks, CA: Sage.

42.

Kerig

P. K.

Moeddel

M. A.

Becker

S. P.

(2011). Assessing the sensitivity and specificity of the MAYSI-2 for detecting trauma among youth in juvenile detention. Child & Care Forum, 40, 345–362

43.

Koroki

Chesney-Lind

(1985). Everything just going down the drain: Interviews with female delinquents (Report No. 319, ERIC Abstract). Honolulu, HI: Youth Development and Research Center.

44.

Kuo

E. S.

Stoep

A. V.

Stewart

D. G.

(2005). Using the short mood and feelings questionnaire to detect depression in detained adolescents. Assessment, 12, 374–383.

45.

Luke

Homan

(1998). Time and change: Using survival analysis in clinical assessment and treatment evaluation. Psychological Assessment, 10, 360–378.

46.

Lyons

J. S.

Libman-Mintzer

L. N.

Kisiel

C. L.

Shallcross

(1998). Understanding the mental health needs of children and adolescents in residential treatment. Professional Psychology: Research and Practice, 29, 582–587.

47.

MacDonald

J. M.

Chesney-Lind

(2001). Gender bias and juvenile justice revisited: A multiyear analysis. Crime & Delinquency, 47, 173–195.

48.

Manchak

Skeem

Rock

(2013). Care, control, or both? Characterizing major dimensions of the mandated treatment relationship. Law and Human Behavior, 38, 47–57.

49.

Medina

Ralphs

Aldridge

. (2012). Mentoring siblings of gang members: A template for reaching families of gang members. Children and Society, 26, 14–24.

50.

Monahan

Steadman

Robbins

Appelbaum

Banks

Grisso

… Silver

(2005). An actuarial model of violence risk assessment for persons with mental disorders. Psychiatric Services, 56, 810–815.

51.

Odem

M. E.

(1995). Delinquent daughters: Protecting and policing adolescent female sexuality in the United States, 1885-1920. Durham: University of North Carolina Press.

52.

Osofsky

J. D.

(1999). The impact of violence on children. The Future of Children, 9, 33–49.

53.

R Core Team. (2017). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from https://www.R-project.org/

54.

Ryon

S. B.

Early

K. W.

Hand

Chapman

(2013). Juvenile justice interventions: System escalation and effective alternatives to residential placement. Journal of Offender Rehabilitation, 52, 358–375.

55.

Schubert

C. A.

Mulvey

E. P.

(2014). Behavioral health problems, treatment, and outcomes in serious youthful offenders. Juvenile justice bulletin. Washington, DC: Office of Juvenile Justice and Delinquency Prevention. Retrieved from https://www.ojjdp.gov/publications/PubAbstract.asp?pubi=264515

56.

Sedlak

McPherson

K. S.

(2010). Conditions of confinement: Findings from the survey of youth in residential placement. Washington, DC: US Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention.

57.

Sherman

L. W.

(2005). The use and usefulness of criminology, 1751–2005: Enlightened justice and its failures. The ANNALS of the American Academy of Political and Social Science, 600, 115–135.

58.

Shufelt

J. L

Cocozza

J. J.

Skowyra

K. R

. (2010). Successfully collaborating with the juvenile justice system: Benefits, challenges, and key strategies. Washington, DC: National Center for Mental Health and Juvenile Justice. Retrieved from https://www.air.org/sites/default/files/downloads/report/jjresource_collaboration.pdf

59.

Sickmund

Sladky

T. J.

Kang

(2004). Census of juveniles in residential placement databook. Retrieved from http://www.ojjdp.ncjrs.org/ojstatbb/cjrp/

60.

Skeem

Emke-Frances

Louden

(2006). Probation, mental health, and mandated treatment: A national survey. Criminal Justice and Behavior, 33, 158–184.

61.

Skowyra

K. R.

Cocozza

J. J

. (2007). Blueprint for change: A comprehensive model for the identification and treatment of youth with mental health needs in contact with the juvenile justice system. Washington, DC: National Center for Mental Health and Juvenile Justice. Retrieved from http://www.ncmhjj.com/wp-content/uploads/2013/07/2007_Blueprint-for-Change-Full-Report.pdf

62.

Solomon

Draine

Marcus

(2002). Predicting incarceration of clients of a psychiatric probation & parole service. Psychiatric Services, 53, 50–56.

63.

Taylor

Kemper

T. S.

Loney

B. R.

Kistner

J. A.

(2009). Recidivism in subgroups of severe male juvenile offenders. Psychology, Crime and Law, 15, 395–408.

64.

Teplin

Abram

McClelland

Washburn

Pikus

(2005). Detecting mental disorder in juvenile detainees: Who receives services. American Journal of Public Health, 95, 1773–1780.

65.

Teplin

Welty

Abram

Dulcan

Washburn

(2012). Prevalence and persistence of psychiatric disorders in youth in juvenile detention. Archives of General Psychiatry, 69, 1133–1143.

66.

Thornberry

T. P.

Huizinga

Loeber

(2004). Causes and correlates studies: Findings and policy implications. Juvenile Justice, 9, 3–19.

67.

Trulson

C. R.

Marquart

J. W.

Mullings

J. L.

Caeti

T. J.

(2005). In between adolescence and adulthood: Recidivism outcomes of a cohort of state delinquents. Youth Violence and Juvenile Justice, 3, 355–387.

68.

Underwood

Warren

K. M.

Talbott

Jackson

Dailey

F. L.

(2014). Mental health treatment in juvenile secure care facilities: Practice and policy recommendations. Journal of Forensic Psychology Practice, 14, 55–85.

69.

U.S. Department of Justice. (2005). Department of Justice activities under Civil Rights of Institutionalized Persons Act: Fiscal year 2004. Washington, DC: Author.

70.

Vander Stoep

Evans

Traub

(1997). Risk of juvenile justice systems referral among children in a public mental health system. Journal of Mental Health Administration, 24, 428–442.

71.

Vaughn

Salas-Wright

DeLisi

Maynard

Boutwell

(2015). Prevalence and correlates of psychiatric disorders among former juvenile detainees in the United States. Comprehensive Psychiatry, 59, 107–116.

72.

Wachter

(2015). Mental Health Screening in Juvenile Justice. JJGPS StateScan. Pittsburgh, PA : National Center for Juvenile Justice. Retrieved from http://www.ncjj.org/pdf/JJGPSStateScan/JJGPS_Mental_Health_Screening_in_Juvenile_Justice_2015_4.pdf

73.

Wannan

Fombonne

(1998). Gender differences in rates and correlates of suicidal behavior amongst child psychiatric outpatients. Journal of Adolescence, 21, 371–381.

74.

Wasserman

G. A.

McReynolds

L. S.

S. J.

Katz

L. M.

Carpenter

J. R.

(2005). Gender differences in psychiatric disorders at juvenile probation intake. American Journal of Public Health, 95, 131–137.

75.

Yoder

Whittaker

Quinn

(2017). Perceptions of recidivism among incarcerated youth: The relationship between exposure to childhood trauma, mental health status, and the protective effect of mental health services in juvenile justice settings. Advances in Social Work, 18, 250–269.

76.

Zoccolillo

Tremblay

Vitaro

(1996). DSM-III-R and DSM-III criteria for conduct disorder in preadolescent girls: Specific but insensitive. Journal of the American Academy of Child & Adolescent Psychiatry, 35, 461–470.