Diminishing Returns? Threshold Effects of Dispositions and Recidivism Among Court-Involved Girls

Abstract

This study examines juvenile court responses among justice-involved girls. We analyze 10 years of court records on girls (N = 1,102) from a Midwestern juvenile court to assess the impact of various aspects of placements and dispositions on recidivism outcomes. We explore how the number of dispositions girls receive, the type of disposition, and type of placement affect 2-year recidivism. Our findings indicate there may be a threshold effect to receiving dispositions—receiving three or more dispositions was significantly related to increased recidivism. Furthermore, the combination of receiving both treatment and sanction dispositions was significantly related to an increased likelihood of recidivism. Girls who received only community-based placements were more likely to recidivate than those who did not receive any dispositions. This study advances our understanding of court responses to girls and how these responses influence girls’ experiences and outcomes while involved with the juvenile justice system.

Keywords

juvenile justice girls gender sanctions recidivism dispositions

Introduction

The goal of the juvenile justice system is to reduce recidivism by intervening in the lives of delinquent youth early and steering them toward the “right path.” The juvenile justice system aims to take a rehabilitative approach compared with the adult criminal justice system to reduce the likelihood of future juvenile offending (Steinberg & Scott, 2003; Sullivan, 2019). While many juvenile dispositions and interventions may be treatment oriented, such as counseling or job training programs, supervision and surveillance are often paired with treatments. Goals of the system may also differ based on gender. For example, the historic paternalism of courts has led to practices that use the juvenile justice system to obtain services for girls they could not otherwise access, to protect girls from sexual victimization, and to prevent teen pregnancy (Sherman, 2013). These responses arise out of fears of expressions of sexuality, and the intolerance of girls who are not compliant has impacted arrest, charge, and detainment decisions of girls who come in contact with the system (Chesney-Lind & Pasko, 2012; Gaarder et al., 2004; Sherman, 2013). Multiple competing goals may undermine the system’s ability to actually be rehabilitative—especially for girls.

There are multiple responses courts use for juveniles including community service, individual and family therapy, mentoring, probation, detention, residential treatment, and more (Feld, 2009; Ryan et al., 2014). Often times, youth involved in the system are required to participate in more than one service. While it may be beneficial to receive multiple dispositions, there may be a point in which multiple dispositions become “too much of a good thing,” resulting in potential over-programming of youth. As girls usually present to juvenile courts with lower criminogenic risks (as measured by standardized risk assessments) and with less severe charge types (Snyder & Sickmund, 2006), it is likely that less court intervention can lead to positive effects for girls (Creaney, 2012). Prior research on dosage has been measured using the number of hours of services, or supervision contact (Bourgon & Armstrong, 2005; Lipsey, 1995). This study broadens the scope of “dosage” to encompass the number of dispositions. It is anticipated that while it may be beneficial for girls to receive some dispositions, there may be a “threshold effect” in which the amount or type of court intervention becomes iatrogenic. The number of dispositions, types of placement, and types of dispositions are important aspects of court responses to explore, especially taking gender into account.

It is common for youth to receive multiple types of dispositions while court-involved, making it difficult to tease out individual program effects. While each individual program evaluation is helpful, categorization of multiple dispositions can help us evaluate broader trends in which types of programs are successful for girls. This study examines various programming outcomes of girls in a Midwestern court. This study explore how juvenile court responses impact girls’ recidivism via the number of dispositions girls are required to participate in, the type of dispositions they are sentenced to (treatments versus sanctions), and the type of placement they must be in (community-based or out-of-home placement). Various goals of the juvenile justice system (e.g., social control and social welfare) are used as the theoretical framework to guide this study. The current project begins with an overview of the goals of the juvenile justice system and detail the various methods courts use to respond to youth delinquency. A series of logistic regression models were used to explore the relationship between court responses to girls and 2-year recidivism rates. To conclude, we provide a discussion of our findings and the implications of our study, focusing on the ways in which our study contributes to advancing knowledge around girls’ experiences and outcomes in juvenile courts.

Literature Review

Competing Goals of the Juvenile Justice System

The juvenile justice system was developed to act as a “kind and just parent,” guiding youth in the right direction (Ward & Kupchik, 2010). A separate system for juveniles was created because youth are immature and therefore, less criminally responsible (Scott & Steinberg, 2008). Youth have heightened reward sensitivity, slower development, and self-regulation issues compared to adults (Sullivan, 2019). Therefore, Steinberg and Scott (2003) explain that the primary goal of the juvenile justice system may be to rehabilitate and identify youth who show early signs of needing intervention.

Morris and McIsaac (1978) proposed two approaches of the juvenile justice system: social control and social welfare. The social control perspective holds youth accountable for their actions through surveillance and punishment. Social welfare perspectives focus more on what is in the “best interest” of the youth (Ward & Kupchik, 2010). While these approaches may be viewed as two extremes on one continuum, the juvenile court has always embraced both views (Ward & Kupchik, 2010). Unfortunately, courts often default to punishment when responding to youth in the system (Ward & Kupchik, 2010). Sherman (2013) notes that some responses can appear to be social welfare responses, but in fact, are social control approaches. As an example, probation may be a method of mentoring, supervising, and helping youth develop. However, it also uses enforcement and surveillance, combining multiple goals of the juvenile justice system. Courts often combine these goals by providing several dispositions to youth—some that function as social control mechanisms and others that provide treatment from a social welfare perspective.

There is evidence demonstrating that the goals of the court and therefore the responses to youth may be different for boys than for girls. Girls represent an increasing proportion of cases in the juvenile justice system (De La Rue & Ortega, 2019). Girls make up 30% of juvenile arrests and 25% of all referred delinquency cases, with adjudications of girls tripling over the last 20 years (Pasko & Lopez, 2018). Due to an increased criminalization of domestic disputes (family violence), sexual vulnerability, and zero-tolerance policies for low-level offending, there has also been an increase in punitive responses to girls early on (Steffensmeier et al., 2005; Stevens et al., 2011). This “up-criming” of responses to girls has led to a diminished tolerance for girls who act out. Assaults are viewed as breaking gender norms for girls, and those who violate those norms will be punished harshly by the justice system (Stevens et al., 2011).

Creaney (2012) explains girls are also drawn into the system for welfare reasons more so than boys. There is a long history of paternalism in which courts try to control girls in the interest of moral welfare, especially following expressions of sexuality (Chesney-Lind & Pasko, 2012; Feld, 2009). Punitive decisions are often disguised as a concern to keep girls safe (e.g., Chesney-Lind & Pasko, 2012; see also Feld, 2009). For example, some courts diverted girls with status offenses to restrictive mental health and substance abuse facilities to avoid restrictions on their ability to place them in secure settings (Morash et al., 2014). These decisions are reflective of the general belief that girls and women cannot care for themselves or make good choices the same way boys and men do (De La Rue & Ortega, 2019). As a result, rather than providing treatment or addressing the root causes of delinquency, girls are more likely to be arrested or detained (Epstein & Edelman, 2014). The differing perspectives on girls’ and boys’ delinquency may contribute to different court responses to youth as well as youth outcomes.

Court Responses to Youth

There are many types of dispositions and supervision levels for youth involved with the system including but not limited to dismissal; continuance without finding; restitution fine; probation (with or without conditions such as house arrest, electronic monitoring, curfew, etc.); restorative justice; psychological and medical evaluations; community service; out-of-home placement (residential facility or group home); confinement in training school; and other secure facilities (short- or long-term detention; Feld, 2009; Ryan et al., 2014). Youth may also be required to participate in a number of types of such as cognitive behavioral therapy, substance abuse treatment, sex offender therapy, anger management/aggression replacement therapy, mental health treatment, general skills training, mentoring, family therapy, mediation, and gang interventions (Sullivan, 2019). Effective dispositions should address youth risk factors and criminogenic needs (Andrews & Bonta, 2010; Lipsey, 2009). For example, among high-risk and violent individuals, correctional placement tends to work well to reduce recidivism rather than community-based treatment (Lipsey, 2009; Lowenkamp & Latessa, 2005). However, many of these findings are based on only male samples, or mixed gender samples that do not include separate analyses for girls (Linder, 2018).

It is important to examine how courts respond to girls and their outcomes. Previous research indicates that girls have different trajectories into the juvenile justice system and the types of delinquency they engage in tends to be less serious (Morash et al., 2014; Snyder & Sickmund, 2006). However, courts often respond to all youth without taking gender into account. That is, they provide the same dispositions to all youth regardless of gender-specific risk factors and needs—often referred to as “gender-neutral” responses. For example, interpersonal violence in the home is more common for girls to experience compared to boys and is a driving reason behind removing girls from their homes (Davidson et al., 2011; Leve et al., 2015; Morash, 2016). Because of this, there may be a different impact of community placements on recidivism for girls than for boys based on risky home environments. It may also be the case that courts recognize how home environments can be dangerous for girls and choose to remove girls from their homes “for their own protection” (Chesney-Lind & Pasko, 2012; Feld, 2009).

Does the Number of Dispositions Matter?

Youth who come into the system often have multiple risks and needs, providing numerous options for intervention (Feld, 2009; Ryan et al., 2014). This may result in courts assigning youth multiple dispositions. The dosage literature has focused mainly on the number of hours of treatment related to recidivism rates. Most of the extant literature focuses on classifying the amount of dispositions rather than types of treatments and sanctions. In general, 100 to 300 hours of treatment is recommended to have an appreciable effect on reducing recidivism (Bourgon & Armstrong, 2005; Lipsey, 1995). However, “too much” treatment or supervision may result in iatrogenic effects (Lowenkamp & Latessa, 2005). The RNR model demonstrates that less intervention is needed for low-risk individuals, while more intervention is required for high-risk individuals (Andrews & Bonta, 2010; Lipsey, 2009; Lowenkamp & Latessa, 2005). Therefore, for some youth, intervention may actually be harmful. Given that girls typically commit fewer offenses and less-severe offenses compared to boys (Anderson et al., 2016), it is possible that girls require less court supervision compared to boys. The implications of this are crucial, as providing any sort of intervention may be harmful to girls involved in the system (Schwalbe & Maschi, 2009). This is especially true for low-risk individuals (which girls are typically categorized as), as diversion has shown better outcomes than court involvement for these youth (Schwalbe, 2008).

Whether through a social welfare or social control lens, the goal of the juvenile justice system is to reduce juvenile recidivism. If the number of dispositions does affect recidivism, courts may actually be harming youth by placing them in multiple dispositions. Thus, it is important to explore how receiving multiple dispositions affects recidivism. Specifically, is there a threshold in which receiving multiple dispositions changes from beneficial to detrimental? This study examines dosage as number of dispositions, which has yet to be explored to assess threshold effects for various sanctions and dispositions girls receive while court involved.

Out-of-Home Versus Community Placements

In the juvenile system, courts assume custody of youth and make decisions about placement in residential settings or to remain in the community with their parents (Morash et al., 2014). Youth who receive community-based sanctions are required to follow rules or be subject to additional regulations and sanctions (Morash et al., 2014). In general, effective responses to youth have favored community supervision over detaining youth (Parhar et al., 2008). Early findings indicated that probation with community-based treatment is related to lower recidivism rates compared to youth who receive out-of-home placement, probation without treatment, and diversion without treatment (Wooldredge, 1988). This is partially because youth may actually have worse outcomes due to confinement, especially among low-risk individuals (Ryan et al., 2014). Out-of-home placements may increase the likelihood of youth experiencing poor living conditions, the absence of rehabilitative programs, and exposure to other delinquent youth (Holman & Ziedenberg, 2006). However, other research has indicated no difference in recidivism rates between those placed in the community and those in an out-of-home placement (Lipsey & Wilson, 1998). Bontrager Ryon and colleagues (2013) found that referral and arrest rates were not different for those on probation compared to those in residential placement, but adjudication and conviction rates were significantly higher for youth placed in residential programs compared to similar juveniles on probation (Bontrager Ryon et al., 2013). Another study by Ryan and associates (2014) compared three dispositions for juveniles, finding group homes and camp-based probation were disruptive as youth compared to at-home probation, supporting the research favoring community-based dispositions for youth.

Most of these results are not based on samples of girls and often when girls are included in the sample, results are not disaggregated by gender. It can be argued that out-of-home placements are more serious than community-based placements as out-of-home placements are recommended for higher risk individuals (Lowenkamp & Latessa, 2005). Girls may be placed out of home due to the mentality of “protecting” girls from their environments, or it may be an increased punishment for breaking gender stereotypes by engaging in deviant behavior (Gaarder et al., 2004; Sherman, 2013). Morash and colleagues (2014) explored multiple types of court responses to girls via in-depth interviews with girls and staff in the juvenile justice system. Court-ordered home confinement (house arrest, home detention, and tethering) was ineffective because girls did not stay home, or they were forced to stay in destructive homes. The girls interviewed said restrictive housing was not successful for them due to neglect, violence, drug use, and chaos in the home. When courts were aware of the family chaos, girls were placed in residential placements. Clearly, knowledge of girls’ home lives and exposure to violence can guide court decision-making for in or out-of-home placements. It is also evident that gender-specific risk factors can influence the “success” of an intervention.

Sanction Versus Treatment

In line with the social control versus social welfare perspective, dispositions can be viewed as more punishment-oriented (sanctions) or treatment-oriented, which may result in different outcomes for youth receiving them. Morash and colleagues (2019) examined outcomes of treatment and punitive responses based on violation types among adult women. They used the following pairings and analyzed recidivism rates: drug violation and substance use treatment; drug violation and punitive response; nondrug violation and substance use treatment; and nondrug violations and punitive responses (Morash et al., 2019). Treatments were categorized as dispositions that encouraged, required, changed, or increased substance abuse treatment. Punitive responses included jail, community service, increased or lengthy supervision, increased drug testing, and revocation of supervision leading to incarceration. Among high-risk women, there were greater reductions (68.5% decrease) in recidivism when treatment responses increased. In contrast, increases in punitive responses were related to a 222% increase in recidivism. Overall, treatment responses to nondrug offenses was linked to reductions in recidivism while punitive responses increased recidivism (Morash et al., 2019). Among low-risk women, when treatment was paired with drug violations, or nondrug offenses were paired with punishment, recidivism was lower. When nondrug offenses were paired with treatment or increased supervision, recidivism increased (Morash et al., 2019). Overall, their findings indicate that punitive responses may create new burdens for women who are already dealing with other issues such as mental illness, single parenting stress, and unemployment.

Another important factor to consider is how dispositions may be perceived by youth and their families. Sullivan (2019) explains that requiring treatment may be viewed as punishment simply because it is mandated. Youth may feel as if they are “forced” to participate in treatment. This perception may result in a lack of motivation to participate in treatment and hinder treatment effects (Parhar et al., 2008). Ward and Kupchik (2010) explain there are considerable overlap in the distribution of treatment and punishment. Court-ordered treatment may be paired with threats of consequences of failure to comply (Parhar et al., 2008). This is especially likely to happen when youth are ordered to treatment as a condition of their release. As these conditions are imposed prior to conviction, youth may view these as orders of punishment. As discussed earlier, there can be multiple components to a disposition, which complicates labeling a court response as treatment or sanction (Morash et al., 2019). Therefore, it is important to explain the purpose of treatments and sanctions so youth can understand them without compromising motivation (Sullivan, 2019).

Current Study

Most juvenile court responses have not been examined with a focus on the types and amount of dispositions in a large sample of girls. This study explores multiple aspects of court responses to girls including both dosage (defined as number of dispositions) and types of dispositions and placements. We use juvenile court records from one Midwestern court to analyze the relationship between service and placement types as well as dosage of dispositions while court involved. In this study, we seek to answer three interrelated research questions:

How does the number of dispositions affect recidivism? Is there a threshold of the number of dispositions that becomes criminogenic?

How do types of dispositions (sanction vs. treatment) impact recidivism?

How do the types of placements (community vs. out of home) impact recidivism?

Based on previous research, we hypothesized that as the number of dispositions increases, the likelihood of recidivism will increase. In addition, given research on girls’ pathways to delinquency, we expect increased recidivism to be related to community-based placements and with out-of-home placements related to decreased recidivism. Based on research on punitive and treatment responses, we expect sanctions to be related to increased recidivism and treatment to be related to decreased recidivism. Related to our hypotheses regarding multiple dispositions, we anticipate a combination of disposition types (both treatment and sanction) and placement types (both community-based and out-of-home placement) to be related to an increased odds of recidivism, regardless of risk level, charge type, race/ethnicity, or age.

Material and Methods

Sample and Design

This study uses secondary data in a cross-sectional design. Participants came from a Midwestern family court’s truancy and delinquency divisions. The sample included 1,102 girls referred to the court from 2004 to 2015. Risk assessment and demographic information was provided for each girl. Risk assessments were completed postadjudication, but prior to predisposition to aid in decision-making and case planning. This study examined recidivism (any delinquency or status petition) 2 years after the girls’ initial referral to court. Summary statistics for each demographics and variables used in the models are provided in Table 1.

Table 1:

Summary Statistics for Demographics, Services, Risk, and Recidivism (N = 1,102)

Variable	Frequency (%)	M (SD)	(Min, Max)
Age	—	14.33 (1.32)	(11, 18)
Race/ethnicity
White American	394 (34.84%)	—	—
Hispanic/Latinx American	124 (10.95%)	—	—
Black/African American	378 (33.42%)	—	—
Other	15 (1.33%)	—	—
Missing	74 (6.14%)	—	—
Charge type
Person	248 (22.30%)	—	—
Property	164 (14.75%)	—	—
Weapon	7 (.63%)	—	—
Drug	12 (1.08%)	—	—
Status	651 (58.54%)	—	—
Sex	2 (0.18%)	—	—
Public ordinance	22 (1.98%)	—	—
Other	6 (0.54%)	—	—
YLS/CMI total		14.20 (6.6)	(0, 38)
2-year recidivism	—	0.36 (0.48)	(0, 1)
Number of services
No services	697 (57.84%)	—	—
One service	162 (13.44%)	—	—
Two services	104 (8.63%)	—	—
Three services	87 (7.22%)	—	—
Four or more services	155 (12.86%)	—	—
Types of services
Only treatment	58 (4.81%)	—	—
Only sanction	176 (14.61%)	—	—
Both treatment and sanction	274 (22.74%)	—	—
Types of placements
Only out-of-home placement	22 (1.83%)	—	—
Only community	238 (19.75%)	—	—
Both community and out of home	248 (20.58%)	—	—

Note. YLS/CMI = Youth Level of Service/Case Management Inventory.

Data Collection and Measures

Data were retrieved through the juvenile and family court’s data management system across both their truancy and delinquency divisions. Girls could be sentenced to any of the court placements (in the community vs. out-of-home) and interventions (sanction vs. treatment) regardless of the original unit they were referred from (e.g., delinquency or truancy). The court implemented the Youth Level of Service/Case Management Inventory (YLS/CMI) risk assessment in 2003, which was scored by the assigned juvenile court officer via face-to-face interviews. This sample includes data on only the girls referred to the court between 2004 and 2015. These data included all girls with data on dispositions and sanctions received, YLS/CMI scores, most serious charge type, offense severity, supervision intensity, and demographic information (e.g., race/ethnicity, age). On average, girls in our sample were 14 years old (SD = 1.31). About 35% of the sample identified as white American, 33% as black/African American, 11% as Hispanic/Latinx American, 20% as Multiracial, and just over 1% were categorized as “Other.”

Dependent Variable

Recidivism was measured 2 years following the girls’ initial YLS/CMI assessment. Recidivism included any new petitions and excluded technical violations of probation. We used this measure to accurately capture all new petitions, each of which further entrenches girls in the system, regardless of whether they are still under court supervision. If girls aged out of the juvenile system during the 2-year window following their assessment, adult records were checked to measure any recidivism recorded by the criminal justice system during that timeframe. Recidivism was coded dichotomously, with “0” indicating no new petitions and “1” indicating at least one new petition. In this sample, 35% of the girls recidivated within 2 years of receiving their YLS/CMI assessment.

Independent Variables

Number of dispositions

Girls were referred to multiple dispositions in this court. Types of dispositions included residential treatment, detention, family-based interventions, in-home detention, intensive supervision, court-run alternative school, group home, job training program, and aftercare for girls after a residential placement. Number of dispositions was coded as a continuous variable indicating the total number of dispositions girls received while under court supervision. While there were nine types of dispositions, no girl received more than seven dispositions. Over 60% of the sample did not receive any type of formal dispositions while court involved. Of those who did receive dispositions, 11% of the sample only received one disposition, followed by another 8% who received two dispositions and 7% who received three dispositions. About 5% of girls received four dispositions, about 4% with five dispositions, and about 2% received six dispositions. Only seven girls (less than 1%) received seven dispositions.

Types of dispositions: Sanction versus treatment

Dispositions were categorized as treatments or sanctions. Types of dispositions were coded into three independent variables: only treatment, only sanction, and both sanction and treatment. The family-based intervention, court-run school, job training program, and group homes (which included trauma and relationally focused care) were coded as treatment. Residential placement, detention, intensive supervision, in-home detention, and aftercare were coded as sanction. Aftercare was considered a sanction because it was only used in conjunction with sanctions (e.g., postresidential placement) rather than treatments. Although youth referred to residential placement or who received intensive supervision may have received treatment, these interventions were coded as “sanctions” as the primary focus is on punishment rather than rehabilitation. Girls who did not receive any dispositions were coded as “0” for no dispositions. Girls who received only treatment-based dispositions were coded as “1” for treatment only. Those who received only sanction-based dispositions were coded as a “1” for sanction only, while girls who received a combination of the two types of dispositions were coded as a “1” for treatment and sanction. For example, a girl who only received the family-based intervention would fit in the treatment-only category. Those who received only residential and aftercare would be categorized in the sanction-only category. If a girl was placed in a group home and later received in-home detention, she would be coded in the treatment and sanction category. With regard to types of dispositions received, about 4% received only treatment, compared to 13% who received only sanction dispositions. Almost one-fourth (22.74%) of the sample received both sanction and treatment dispositions.

Placements: Community versus out of home

Dispositions were also categorized by the type of placement: out of home, community based, or receiving a combination of both. These placements were coded as independent dichotomous variables. Community-based placements included: family-based intervention, intensive supervision, court-run school, job training programming, in-home detention, and aftercare. Out-of-home placements included residential placements, detention, or group home placement. For example, if a girl only received the family-based intervention and intensive supervision, she would be coded as a “1” for community only placements. If she received only detention and no community-based dispositions, she would be coded as a “1” for out-of-home-only placement. Girls who received a combination of both types of dispositions (e.g., job training and group home placement), were coded as a “1” for the community and out-of-home placement variable. Those who were not placed in out of home or community dispositions were coded a “0” for no placement. When considering community compared to out-of-home placements, less than 2% were given only an out-of-home placement (which may in part be explained by the common use of aftercare in combination with out-of-home placements). In contrast, 18% of the sample received only a community-based placement. Another 19% received a combination of out-of-home and community placements¹.

Control Variables

Most serious charge type was controlled for in our models. The most common charge was for a status offense with almost 60% of girls charged with a status offense (e.g., truancy). The next most common charge was person-related offenses (22%) followed by property offenses (14%). A small percentage (under 2% for each category) of girls were charged with a weapon, drug, sex, or public ordinance crime.

As treatment effects may be impacted by risk level (Morash et al., 2019), we controlled for risk in our study. The YLS/CMI is a widely used risk assessment tool comprised of 42 items divided into eight subscales (Hoge & Andrews, 1996; Schwalbe, 2008). The eight subscales measured with the YLS/CMI include prior offenses, family and parenting, education, peer relationships, substance use, leisure and recreation, personality and behavior, and attitudes and orientation². The mean YLS/CMI score was 14.17 (SD = 6.58) on a scale of 1 to 38, indicating that most girls in the sample were classified as moderate risk.

Supervision intensity was measured using the number of days girls were involved with the court. The number of days under supervision was calculated using the first date girls came under supervision of the court and the date girls were released from court supervision under their final disposition. Initially, the number of days under supervision ranged from 3 days to 2,429 days. However, this variable was highly positively skewed, and thus was truncated at the 90th percentile. Under the transformed variable, the number of days under supervision ranged from 3 to 1,070 days, with a mean of 571.23 days (SD = 414.19) and a median of 466 days.

Offense severity was measured using a variety score of types of offenses (Sweeten, 2012). A variety score measures the number of types of crime an individual engages in, rather than frequency or a binary violent/nonviolent crime measure³. A higher variety score indicates engaging in more types of crime. Crime codes were indicated in the administrative data and coded to represent 14 types of crimes ranging from status offenses to murder. Records for girls’ previous five crimes were used to create the variety score. On average, girls committed between one and two crimes (M = 1.45, SD = .73).

Analyses

To explore our research questions, we ran a series of logistic regression analyses. Model 1 explored the relation between number of dispositions and 2-year recidivism rates. Model 2 focused on the type of dispositions provided while Model 3 examined the type of placement. One important consideration is that youth who commit more serious crimes may receive more dispositions. For example, someone who commits an offense against a person may receive more dispositions than someone who commits a status offense. It is also possible other variables such as risk level, type of placement, and type of dispositions may be strongly correlated. We ran a multicollinearity check using variance inflation factor (VIF) measures. Each VIF factor was below 4 and ranged from 1.03 to 3.87, suggesting multicollinearity is not a cause for concern (Tabachnick et al., 2007).

Results

Table 2 displays the results for the logistic regression analysis of the effect of the number of dispositions girls received on 2-year recidivism rates. Receiving one or two dispositions was not significantly related to 2-year recidivism compared to receiving no dispositions. However, receiving three dispositions was significantly related to a 79% increase in the odds of recidivating within 2 years relative to receiving zero dispositions (OR = 1.79, p = .03). Interestingly, receiving four or more dispositions was not significantly associated an increase in the odds of recidivating within 2 years. However, until girls receive seven dispositions, an increase in number of dispositions is related to an increased likelihood of recidivism compared to receiving no dispositions. Several control variables were also significant in the model. The YLS/CMI score was significantly related to 3% increased odds of 2-year recidivism (OR = 1.03, p = .03). Age was significantly related to 24% decreased odds of recidivating within 2 years (OR = .76, p < .001). Only one charge type was significantly related to 2-year recidivism: status offenses. Being charged with a status offense was associated with 44% decrease in the odds of recidivating within 2 years compared to girls who were charged with person-related offenses (OR = .56, p < .001). Committing more types of crimes was related to a 29% increase in the odds of recidivating (OR = 1.29, p = .01). While supervision intensity was statistically significant, the magnitude of this relationship is essentially nonexistent (p = .01).

Table 2:

Logistic Regression for Number of Dispositions Compared to No Dispositions on 2-Year Recidivism

Variable	b	SE	Odds ratio	SE	p value
Number of dispositions
One disposition	0.38	0.23	1.47	0.32	.08
Two dispositions	0.12	0.25	1.13	0.29	.63
Three dispositions	0.58*	0.27	1.79*	0.48	.03
Four dispositions	0.20	0.28	1.23	0.34	.47
Five dispositions	0.49	0.35	1.63	0.57	.17
Six dispositions	0.77	0.45	2.16	0.96	.09
Seven dispositions	−0.30	0.81	0.74	0.60	.70
YLS total	0.03*	0.01	1.03*	0.01	.03
Age	−0.27*	0.06	0.76*	0.04	<.001
Race/ethnicity
Hispanic/Latinx	0.25	0.23	1.29	0.30	.35
African American	0.06	0.16	1.06	0.17	.57
Multiracial	0.12	0.19	1.13	0.21	.33
Other	−0.65	0.64	0.52	0.33	.38
Charge type
Property	0.08	0.22	1.08	0.24	.89
Weapon	−0.12	0.89	0.89	0.79	.69
Drug	−0.30	0.66	0.74	0.49	.59
Status	−0.57*	0.18	0.56*	0.10	<.001
Sex	1.08	1.56	2.94	4.60	.60
Public ordinance	0.08	0.47	1.08	0.51	.82
Other	−0.11	0.92	0.90	0.82	.82
Supervision intensity	0.00*	0.00	1.00*	0.00	.01
Offense severity	0.26*	0.10	1.29*	0.13	.01

Note. Number of dispositions compared to no dispositions; race/ethnicity compared to white youth; charge type compared to person charges. YLS = Youth Level of Service.

p < .05.

To address our second research question, we explored the differences between sanction and treatment dispositions on 2-year recidivism. Our findings from the logistic regression are displayed in Table 3. While the treatment-only and sanction-only groups were not significantly related to recidivism, the combination of treatment and sanction was statistically significant (p = .04). Girls who received at least one treatment and one sanction had 44% increased likelihood of recidivating compared to girls who did not receive any dispositions (OR = 1.44)⁴. Again, the YLS/CMI score was significantly related to 3% increase in recidivism (OR = 1.03, p = .02) while increased age was associated with 24% decreased odds in recidivism (OR = .76, p < .001). Only status offenses were significantly associated with 2-year recidivism. Girls charged with a status offense were 43% less likely to recidivate compared to girls charged with a person offense (OR = .50, p = .002). Again we see a negligible, but statistically significant relationship between supervision intensity and recidivism (OR = 1.00, p = .01). Being charged with more types of crimes was again significantly associated with a 28% increase in the odds of recidivating.

Table 3:

Logistic Regression for Types of Dispositions Compared to No Dispositions on 2-Year Recidivism

Variable	b	SE	Odds ratio	SE	p value
Types of services
Only treatment	0.57	0.18	1.77	0.58	.09
Only sanction	0.27	0.33	1.31	0.28	.21
Treatment and sanction	0.36*	0.21	1.44*	0.25	.04
YLS total	0.03*	0.01	1.03*	0.01	.02
Age	−0.27*	0.06	0.76*	0.04	<.001
Race/ethnicity
Hispanic/Latinx	0.26	0.23	1.30	0.20	.26
African American	0.05	0.16	1.05	0.17	.75
Multiracial	0.12	0.19	1.13	0.21	.52
Other	−0.61	0.63	0.55	0.35	.34
Charge type
Property	0.09	0.22	1.10	0.24	.67
Weapon	−0.19	0.90	0.83	0.75	.84
Drug	−0.23	0.67	0.79	0.53	.72
Status	−0.56*	0.18	0.57*	0.10	.002
Sex	1.00	1.57	2.72	4.28	.53
Public ordinance	0.06	0.47	1.06	0.50	.90
Other	−0.11	0.93	0.89	0.83	.90
Supervision intensity	0.00*	0.00	1.00*	0.00	.01
Offense severity	−0.25*	0.10	1.28*	0.13	.01

Note. Disposition categories were compared to receiving no dispositions; race/ethnicity compared to white girls; charge type compared to person-related offenses. YLS = Youth Level of Service.

p < .05.

To answer research question three, we examined the effect of type of placement (community versus out of home) on 2-year recidivism. Table 4 provides the logistic regression results for the effect of various types of placements (i.e., out-of-home and community-based placement) on 2-year recidivism. Receiving an out-of-home placement was not significantly related to recidivism. In addition, the combination of at least one out-of-home placement and one community-based placement was not significantly related to recidivism. Receiving only community placements was related to a 50% increase in the odds of recidivating within 2 years (OR = 1.50, p = .03). The null combination effect may be driven by the out-of-home placement effect. It is also important to note that the “only out of home” placement group was small (n = 17). It would be beneficial to further explore this relationship with a larger sample size. Similar to the first two regression models, YLS/CMI score (OR = 1.03, p = .02), age (OR = .76, p < .001), status offenses (OR = .58, p = .002), supervision intensity (OR = 1.00, p = .004), and offense severity (OR = 1.29, p = .01) are significantly related to recidivism in the same directions as the first two models.

Table 4:

Logistic Regression for Types of Placement on 2-Year Recidivism

Variable	b	SE	Odds Ratio	SE	p value
Types of placements
Only out of home	0.45	0.51	1.57	0.81	.38
Only community	0.40*	0.18	1.50*	0.27	.03
Out of home and community	0.30	0.19	1.35	0.25	.12
YLS total	0.03*	0.01	1.03*	0.01	.02
Age	−0.27*	0.06	0.76*	0.04	<.001
Race/ethnicity
Hispanic/Latinx	0.26	0.23	1.30	0.30	.26
African American	0.05	0.16	1.05	0.17	.75
Multiracial	0.12	0.19	1.13	0.21	.52
Other	−0.60	0.63	0.55	0.35	.34
Charge type
Property	0.10	0.22	1.10	0.24	.85
Weapon	−0.17	0.90	0.85	0.75	.69
Drug	−0.26	0.66	0.77	0.51	.554
Status	−0.55*	0.18	0.58*	0.10	.002
Sex	0.94	1.59	2.56	4.07	.55
Public ordinance	0.07	0.47	1.07	0.50	.89
Other	−0.12	0.92	0.89	0.82	.90
Supervision intensity	0.00*	0.00	1.00*	0.00	.004
Offense severity	0.25*	0.10	1.29*	0.013	.01

Note. Placement categories were compared to receiving no placements; race/ethnicity compared to white girls; charge type compared to person-related offenses. YLS = Youth Level of Service.

p < .05.

Discussion

This study explored the effect of various aspects of court responses to girls in the juvenile justice system on recidivism outcomes. We first explored how the number of dispositions impacted 2-year recidivism, particularly to understand whether there was a “threshold effect,” in which an increased number of dispositions may be helpful and then potentially harmful to girls’ recidivism outcomes. Beyond the number of services, we explored categorizations of types of placements and the orientation of dispositions girls received. Service orientation were operationalized as either sanctions or treatments and placements were categorized as community-based or out-of-home placements. We explored these questions with a girl-only sample, given that there is limited empirical research on the effects of juvenile court responses to girls. As well, given the research on gendered pathways to delinquency and recent scholarship advocating for gender-responsive programming, we were interested in which aspects of standard “gender-neutral” juvenile court programming may be important for girls’ success specifically.

Our results provide support for a threshold effect of the number of dispositions on girls’ recidivism. Receiving one or two dispositions was not significantly related to recidivism. However, receiving three or more dispositions was significantly related to increased odds of recidivism, even when controlling for age, race/ethnicity, most serious charge type, and risk level. This finding may be evidence of potential over-programming of girls in the system and is in line with prior research that indicates multiple dispositions, and court requirements may place an added burden on women (Morash et al., 2019). As such, courts should be cautious when assigning multiple dispositions to girls. Our findings also pose difficulties for future research trying to disentangle the effects of sequential versus overlapping dispositions as well as gender-responsive versus gender-neutral treatment modalities. Most girls who are court-involved will also be receiving other types of dispositions that do not have gender-responsive elements (e.g., trauma-informed, relationship-focused), and mixing these types of intervention modalities makes it hard to discern program effects. Therefore, it may be the case that fewer “gender-neutral” dispositions and interventions are beneficial to girls who encounter the system. These suggestions may contradict historic practices in the juvenile justice system which has attempted to “protect” girls by providing them with multiple interventions.

Related to the number of dispositions, only the combination of treatments and sanctions was significantly related to recidivism. This may even further support the over-programming hypothesis. It is also possible that the combination of sanctions and treatments leads individuals to perceive treatments as sanctions (Parhar et al., 2008; Sullivan, 2019). Should this be the case, courts should spend time explaining the purpose of each requirement and disposition to youth to bolster treatment effects.

Our findings related to treatments and sanctions were counter to our original hypothesis. The goals of the juvenile court include both rehabilitation and accountability, which we operationalized as treatments and sanctions in this study. However, neither of these goals were significantly related to recidivism. This result has two potential explanations. One possible explanation is that the juvenile justice system’s intertwined goals are not successful methods of reducing recidivism. However, a more likely explanation, supported by research in feminist criminology over the last few decades, is that the court’s gender-neutral approach to reducing recidivism is not appropriate for girls—and may actually lend itself to increasing girls’ odds of recidivating. In a study by Gaarder and colleagues (2004), a lack of understanding and knowledge of culturally and gender-appropriate dispositions resulted in a disconnect between girls’ needs and dispositions received. This finding is unsurprising as justice system interventions were developed with boys in mind and applied to girls without attention to their specific risks and needs (De La Rue & Ortega, 2019; Leve et al., 2015).

Finally, our findings regarding out-of-home and community placements contradict those found in the broader juvenile justice literature. Community-based interventions have been found to reduce recidivism in male-only and mixed-gender samples (Andrews & Bonta, 2010; Ryan et al., 2014). Research supports community-based interventions to keep important social bonds intact and to reduce disruption in the lives of youth (Bontrager Ryon et al., 2013). Our study challenges these results using a girl-only sample. While we found a null effect of out-of-home placements on recidivism, we did find girls who received only community-based placements were more likely to recidivate than those who did not receive any dispositions. This finding may be indicative of gender-specific risk factors and needs, as community-based placements have found success in reducing recidivism in mixed-gender samples (Holman & Ziedenberg, 2006; Parhar et al., 2008).

Research and Practice Implications

Given girls’ risk factors related to negative home environments, it may be detrimental to keep girls in at-home placements. It is important to explore this finding in future research. For example, qualitative research may allow us to examine potential issues girls face within community-based placements to provide context to this finding. Leve et al. (2015) advocate for community-based intervention but explain that if girls are living at home, courts should consider home-based treatments that include the family and prioritize safety. On the contrary, findings from Morash and colleagues (2014) suggest courts consider out-of-home placement for girls who come from destructive and unstable families. In the Morash et al. (2014) study, girls who had at least one stable parent in the home were less likely to re-offend. However, those who did not have at least one stable parent at home did better in out-of-home placements that empowered them and prepared them to live on their own (Morash et al., 2014). Clearly, focusing on the specific needs and experiences of girls will impact their success in various placements and must be taken into account when courts are considering placement options. In each model, girls who received no dispositions (and controlling for the covariates in our models) fared better in terms of returning to court on new petitions compared to girls who received any sort of disposition from the court.

Overall, our findings provide some pause for use of gender-neutral approaches for court-involved girls. For example, community-based interventions have been more effective at reducing recidivism compared to out-of-home placements for juveniles in general (Andrews & Bonta, 2010; Leve et al., 2015; Parhar et al., 2008); however, due to girls’ chaotic family histories and exposure to violence, community placement may be detrimental rather than helpful in reducing girls’ recidivism. Given research on gender-responsive programming, girls’ may respond better to dispositions with a treatment emphasis compared to those that are more punitive (Davidson et al., 2011; Morash, 2016; Morash et al., 2019). Of note to this study, Davidson and colleagues (2011) reviewed a gender-specific program in Honolulu and concluded when gender-responsive programming is delivered well, it can avoid issues that are typically associated with punitive responses. Courts might consider employing fewer interventions to reduce recidivism among girls. They should also consider the contextual factors such as parental concerns and home conditions, of girls when assigning them dispositions and placements. Our study’s findings align with prior literature that has cautioned against the justice system using a gender-neutral approach in responding to girls. In addition, we advocate for further examination of gender-specific programming and its effect in contrast to gender-neutral court responses examined here.

Limitations

This study is informative in many ways but is limited in a few key aspects. First, while the categorization of placements and dispositions was based on prior research and goals of the justice system, an empirical approach to categorizing dispositions (e.g., latent class analysis) for future studies is ideal. In addition, the number of programs that girls received varies and these programs likely overlap or run sequentially making it difficult to discern the individual or combined program effects on recidivism outcomes. While examining specific program or service effectiveness was not a goal of this article, it would be important to examine in the future when trying to unpack the mechanisms that might be producing increases or decreases in girls’ recidivism.

Our measurement of recidivism is a follow-up from the date girls were given the YLS/CMI and accounted for all new charges, regardless of the length of time they were under supervision. Other measures of recidivism may use any new offense after the exit date of court supervision, particularly for program evaluations. We acknowledge the utility of this measure, however over one-third of our sample was committing new offenses even while they were under court supervision. With each new offense, girls are further entrenched into the system, highlighting the importance of measuring all new offenses beginning with their initial entry to the court. We also did attempt to address this limitation by controlling for the number of days youth were under court supervision. Future research should consider the multiple methods for measuring recidivism and explore how these various measurements affect program success and girls’ trajectories within the system.

It is possible that our findings are picking up on a surveillance effect rather than a disposition threshold effect. Youth who are placed on community supervision may have more opportunities to commit new offenses and be more likely to be caught engaging in delinquent activities compared to youth in out-of-home placements. Therefore, these findings may not be indicative of a definitive effect due to dispositions, but rather the combined effect of dispositions and surveillance in which any surveillance effect would be embedded with disposition threshold effects. In other words, it is impossible to separate these two ideas because increased dispositions may also mean increased surveillance. We do attempt to address this limitation by controlling for supervision intensity, but future research would benefit from including additional program characteristics such a number of contacts with probation officers (Morash et al., 2019). While these findings may also be in part explained by increased surveillance, we cannot disentangle these ideas empirically with the current data. Future studies may consider examining diversion programming with minimal surveillance compared to youth who are on probation (and thus under more surveillance).

Related to the placement and dispositions groups, 58 girls received only treatment. The only sanction and the combination groups each had over 150 girls. In addition, there were only 22 girls in the “only out-of-home” placement, compared to over 200 girls in the community and combination groups. The small number of girls in these groups and null findings may have influenced the results in the study due to issues with statistical power. It would also be important for future studies to integrate additional program characteristics to further tease out categories of factors in court dispositions that support girls within the system. These finer disposition components will be especially beneficial to examine in program evaluations.

Finally, it is important to highlight structural differences that may influence girls’ access to adequate resources. Girls involved in the system tend to have family issues at home and often come from lower socioeconomic backgrounds (Lopez, 2017; Morash et al., 2014). Families with lower education levels and lower socioeconomic status may not have the same access to dispositions for delinquent youth. These factors may impact each of the aspects explored in this study. As these families have less access to resources, courts may opt for a sanction rather than treatment responses, especially if treatment requires payment from the family. Under the “protection” hypothesis, courts may be more inclined to remove girls from the home, which may be viewed as disruptive and criminogenic. It is also possible that courts begin with a community-based intervention, but due to barriers these families face, youth are unable to meet expectations and fail to comply. Failure to comply often results in an increase in supervision and may result in out-of-home placement (Bontrager Ryon et al., 2013; Sherman, 2013). This may also increase the number of dispositions youth receive from the court as youth who are on probation may be sentenced to residential placement or a group home as well. Structural factors such as socioeconomic status, parental education, and community variables should be explored in future studies. Examining the structural characteristics to girls’ delinquency is supported by the feminist criminology literature, gender-responsive programming, and trauma-informed care—all of which may lead to better outcomes for justice-involved girls.

Conclusion

The traditional goals of the system tend to place treatment and sanctions on opposite sides of a continuum, when in reality, any court intervention may be viewed as punishment. Our results suggest that less intervention leads to better outcomes for girls. Historically and currently, the justice system has intervened in girls lives to both protect and control them. However, court responses were primarily were designed with boys in mind, leading to courts to respond to girls in gender-neutral ways by applying interventions without girls’ specific needs in mind (De La Rue & Ortega, 2019; Leve et al., 2015; Sherman, 2013). Unfortunately, ineffective gender-neutral programming can have deleterious outcomes among girls. For example, while community-based placements have been found to reduce reoffending in general samples, it was detrimental to girls in our sample and actually increased their likelihood of recidivism. Our findings also point to the importance of developing and implementing gender-responsive dispositions for girls in the juvenile justice system. Policies and programs that emphasize girls’ needs and empower them to have control over their own lives may improve outcomes and reduce their involvement in the system.

Footnotes

ORCID iD

Valerie R. Anderson

Notes

Nicole C. McKenna, MS, is a doctoral student in the School of Criminal Justice at the University of Cincinnati. The primary aim of her research program involves examining juvenile delinquency and traumatic experiences. She primarily focuses on females in the juvenile justice system and effective interventions to address trauma. Her work is featured in Journal of Trauma & Dissociation and Journal of Aggression, Maltreatment, & Trauma.

Valerie R. Anderson, PhD, is an assistant professor in the School of Criminal Justice at the University of Cincinnati. The primary aim of her research program involves examining juvenile corrections and victimization—and the circumstances and contexts in which these areas intersect. She has a particular interest in female juvenile delinquency and human trafficking. Her most recent work is featured in American Journal of Public Health, Crime & Delinquency, Criminal Justice and Behavior, and American Journal of Community Psychology.

References

Anderson

V. R.

Davidson

W. S.

Barnes

A. R.

Campbell

C. A.

Petersen

J. L.

Onifade

(2016). The differential predictive validity of the Youth Level of Service/Case Management Inventory: The role of gender. Psychology, Crime & Law, 22(7), 666–677.

Andrews

D. A.

Bonta

(2010). The psychology of criminal conduct (3rd ed.). Routledge.

Bontrager Ryon

Winokur Early

Hand

Chapman

. (2013). Juvenile justice interventions: System escalation and effective alternatives to residential placement. Journal of Offender Rehabilitation, 52(5), 358–375. https://doi.org/10.1080/10509674.2013.801385

Bourgon

Armstrong

(2005). Transferring the principles of effective treatment into a “real world” prison setting. Criminal Justice & Behavior, 32(1), 3–25.

Chesney-Lind

Pasko

(2012). The female offender: Girls, women, and crime. SAGE.

Creaney

(2012). Risk, prevention and early intervention: Youth justice responses to girls. Safer Communities, 11(2), 111–120. https://doi.org/10.1108/17578041211215348

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(4), 308–327. https://doi.org/10.1080/08974454.2011.609406

De La Rue

Ortega

(2019). Intersectional trauma-responsive care: A framework for humanizing care for justice involved girls and women of color. Journal of Aggression, Maltreatment & Trauma, 28(4), 502–517. https://doi.org/10.1080/10926771.2019.1572403

Epstein

Edelman

(2014). Blueprint: A multidisciplinary approach to the domestic sex trafficking of girls. Center on Poverty and Inequality, Georgetown Law School.

10.

Feld

B. C.

(2009). Violent girls or relabeled status offenders? An alternative interpretation of the data. Crime & Delinquency, 55(2), 241–265. https://doi.org/10.1177/0011128708330629

11.

Gaarder

Rodriguez

Zatz

M. S.

(2004). Criers, liars, and manipulators: Probation officers’ views of girls. Justice Quarterly, 21(3), 547–578. https://doi.org/10.1080/07418820400095901

12.

Hoge

R. D.

Andrews

D. A.

(1996, August 8–11). The youth level of service/case management inventory: Description and evaluation [Conference session]. APA 1996 Annual Conference, Toronto, Canada.

13.

Holman

Ziedenberg

(2006). The dangers of detention: The impact of incarcerating youth in detention and other secure facilities. Justice Policy Institute. http://www.justicepolicy.org/uploads/justicepolicy/documents/dangers_of_detention.pdf

14.

Leve

L. D.

Chamberlain

Kim

H. K.

(2015). Risks, outcomes, and evidence-based interventions for girls in the US juvenile justice system. Clinical Child & Family Psychology Review, 18(3), 252–279. https://doi.org/10.1007/s10567-015-0186-6

15.

Linder

(2018). An unsupported population: The treatment of women in Texas’ criminal justice system. Texas Criminal Justice Coalition. https://www.texascjc.org/sites/default/files/publications/TCJC-Womens-Part-2.pdf

16.

Lipsey

M. W.

(1995). What do we learn from 400 research studies on the effectiveness of treatment with juvenile delinquents? In McGuire

(Ed.), Wiley series in offender rehabilitation. What works: Reducing reoffending: Guidelines from research and practice (pp. 63–78). John Wiley.

17.

Lipsey

M. W.

(2009). The primary factors that characterize effective interventions with juvenile offenders: A meta-analytic overview. Victims & Offenders, 4(2), 124–147. https://doi.org/10.1080/15564880802612573

18.

Lipsey

M. W.

Wilson

(1998). Effective intervention for serious offenders: A synthesis of research. In Loeber

Farrington

D. F.

(Eds.), Serious and violent juvenile offenders: Risk factors and successful interventions (pp. 313–345). SAGE.

19.

Lopez

(2017). Complicated lives: Girls, parents, drugs, and juvenile justice. Rutgers University Press.

20.

Lowenkamp

C. T.

Latessa

E. J.

(2005). Increasing the effectiveness of correctional programming through the risk principle: Identifying offenders for residential placement. Criminology & Public Policy, 4(2), 263–290.

21.

Morash

(2016). The nature of co-occurring exposure to violence and of court responses to girls in the juvenile justice system. Violence Against Women, 22(8), 923–942.

22.

Morash

Kashy

D. A.

Smith

S. W.

Cobbina

J. E.

(2019). Technical violations, treatment and punishment responses, and recidivism of women on probation and parole. Criminal Justice Policy Review, 30(5), 788–810. https://doi.org/10.1177/1077801215614973

23.

Morash

Stevens

Yingling

(2014). Focus on the family: Juvenile court responses to girls and their caretakers. Feminist Criminology, 9(4), 298–322. https://doi.org/10.1177/1557085114527151

24.

Morris

McIsaac

M. W.

(1978). Juvenile justice? The practice of social welfare. Heinemann.

25.

Parhar

K. K.

Wormith

J. S.

Derkzen

D. M.

Beauregard

A. M.

(2008). Offender coercion in treatment: A meta-analysis of effectiveness. Criminal Justice & Behavior, 35(9), 1109–1135. https://doi.org/10.1177/0093854808320169

26.

Pasko

Lopez

(2018). The Latina penalty: Juvenile correctional attitudes toward the Latina juvenile offender. Journal of Ethnicity in Criminal Justice, 16(4), 272–291. https://doi.org/10.1080/15377938.2015.1015196

27.

Ryan

J. P.

Abrams

L. S.

Huang

(2014). First-time violent juvenile offenders: Probation, placement, and recidivism. Social Work Research, 38(1), 7–18. https://doi.org/10.1093/swr/svu004

28.

Schwalbe

C. S.

(2008). A meta-analysis of juvenile justice risk assessment instruments: Predictive validity by gender. Criminal Justice & Behavior, 35(11), 1367–1381. https://doi.org/10.1177/0093854808324377

29.

Schwalbe

C. S.

Maschi

(2009). Investigating probation strategies with juvenile offenders: The influence of officers’ attitudes and youth characteristics. Law & Human Behavior, 33(5), 357–367. https://doi.org/10.1007/s10979-008-9158-4

30.

Scott

E. S.

Steinberg

(2008). Adolescent development and the regulation of youth crime. The Future of Children, 18(2), 15–33.

31.

Sherman

F. T.

(2013). Justice for girls: Are we making progress. Criminal Justice, 28(2), 9–17.

32.

Snyder

H. N.

Sickmund

(2006). Juvenile offenders and victims: 2006 national report. Office of Juvenile Justice and Delinquency & Prevention.

33.

Steffensmeier

Schwartz

Zhong

Ackerman

(2005). An assessment of recent trends in girls’ violence using diverse longitudinal sources: Is the gender gap closing? Criminology, 43(2), 355–406. https://doi.org/10.1111/j.0011-1348.2005.00011.x

34.

Steinberg

Scott

E. S.

(2003). Less guilty by reason of adolescence: Developmental immaturity, diminished responsibility, and the juvenile death penalty. American Psychologist, 58(12), 1009–1018. https://doi.org/10.1037/0003-066X.58.12.1009

35.

Stevens

Morash

Chesney-Lind

(2011). Are girls getting tougher, or are we tougher on girls? Probability of arrest and juvenile court oversight in 1980 and 2000. Justice Quarterly, 28(5), 719–744. https://doi.org/10.1080/07418825.2010.532146

36.

Sullivan

C. J.

(2019). Taking juvenile justice seriously: Developmental insights and system challenges. Temple University Press.

37.

Sweeten

(2012). Scaling criminal offending. Journal of Quantitative Criminology, 28(3), 533–557.

38.

Tabachnick

B. G.

Fidell

L. S.

Ullman

J. B.

(2007). Using multivariate statistics (Vol. 5). Pearson.

39.

Ward

Kupchik

(2010). What drives juvenile probation officers? Relating organizational contexts, status characteristics, and personal convictions to treatment and punishment orientations. Crime & Delinquency, 56(1), 35–69. https://doi.org/10.1177/0011128707307960

40.

Wooldredge

J. D.

(1988). Differentiating the effects of juvenile court sentences on eliminating recidivism. Journal of Research in Crime & Delinquency, 25(3), 264–300. https://doi.org/10.1177/0022427888025003004