juvenile justice

Feminist Criminology 2016, Vol. 11(2) 135 –162 © The Author(s) 2014 Reprints and permissions:

sagepub.com/journalsPermissions.nav DOI: 10.1177/1557085114559514 fcx.sagepub.com Article An Ecological Model for Intervention for Juvenile Justice-Involved Girls:

Development and Preliminary Prospective Evaluation Shabnam Javdani 1 and Nicole E. Allen 2 Abstract Despite increased attention on girls’ delinquency, evidence-based int\ erventions have been largely lacking. We aim to (a) describe the design and implementa\ tion of a gender- specific program for juvenile justice-involved girls and (b) present p\ rospective data on three cohorts of participants. Fifty-two girls were enrolled in the prog\ ram called the Girls Advocacy Project (GAP). Key risk and protective factors were ide\ ntified based on existing literature and assessed. Over time, youth reported greater r\ esilience and self-efficacy; fewer risk behaviors, including violence, crime, and \ substance use; and decreased distress (e.g., depression, anxiety, and anger). Hypothe\ sized program mechanisms and recommendations for future research are discussed.

Keywords intervention/treatment, advocacy, juvenile justice/delinquency, youth/ad\ olescence Adolescent girls exhibiting disruptive behaviors and externalizing psychopathology are at increased risk of physical and mental health problems, including \ violence and substance use, compared with their male counterparts (Moffitt, Caspi, Rutter, & Silva, 2001; Zahn, Hawkins, Chiancone, & Whitworth, 2008; Zahn et al., 2010). Girls’ arrests for serious offenses (e.g., assault) have also increased more or decreased less than those for males over the course of the last two decades (Snyder & \ Sickmund, 1New York University, New York City, USA2University of Illinois at Urbana–Champaign, USA Corresponding Author:

Shabnam Javdani, Department of Applied Psychology, New York University, \ 246 Greene St., Room 706W, New York, NY 10003, USA.

Email: [email protected] 559514 FCX XX X 10.1177/1557085114559514Feminist CriminologyJavdani and Allen research-article 2014 136 Feminist Criminology 11(2) 2006). These patterns represent a significant public health concern with well-k\ nown social sequelae including mortality, infertility, neighborhood disorganization, infant death, and intergenerational poverty (e.g., Bush-Baskette, 2004; Sharp & Marcus- Mendoza, 2001). Despite these patterns and the continued rise in girls’ arrests, the response to girls’ antisocial behaviors and delinquency has been critiqued and characterized as largely inadequate (Belknap, 2010). Indeed, prominent among them has been a cr\ itique that youth programs neglect important gender-specific considerations and are tailored (if at all) toward the needs of boys (Bloom, Owen, Deschenes, & Rosenbaum, 20\ 02; Chesney-Lind & Shelden, 2004;; Zahn et al., 2008). This pattern suggests an overall paucity of gender-responsive programming for the largest growing segment of the juvenile justice population—young women. Attention to these patterns has led to the emergence of greater research regarding system-involved girls’ outcomes and needs (Belknap & Holsinger, 2006; Chesney- Lind & Pasko, 2004; Chesney-Lind & Shelden, 2004; Javdani, Sadeh, & Verona, 2011b; Zahn et al., 2010). Indeed, the Office of Juvenile Justice and Delinquency Prevention convened an interdisciplinary group to examine the evidence b\ ase with respect to the causes and correlates of girls’ delinquency (e.g., Zahn et al., 2008). This report (Zahn et al., 2008) underscored two critical points: (a) Girl\ s have distinct risk and protective factors that predict system involvement and (b) few gen\ der-responsive programs target girls’ needs. The Girls Advocacy Project (GAP), a strengths-based, youth-driven, advocacy intervention, was developed by building on the Co\ mmunity Advocacy Project (Sullivan & Bybee, 1999) and Adolescent Diversion (Davidson & Rapp, 1976) to address these disparities.

Girls’ Pathways to Delinquency A number of characteristics and experiences seem to distinguish girls’\ pathways to delinquency from that of boys. These include early pubertal development, early child- hood sexual victimization, association with older romantic partners, and\ comorbid experience of internalizing symptoms, such as depression and anxiety (B\ elknap & Holsinger, 2006; Javdani, Sadeh, & Verona, 2011a; Zahn et al., 2008; Zahn et al., 2010). Beyond the experience of particular risk factors, it has been ar\ gued that girls’ delinquency emerges in a gendered social context characterized as particularly oppres- sive for young women, including through lower access to social resources and expo- sure to victimization across the life span (Bloom, Owen, Rosenbaum, & P\ iper, 2003; Chesney-Lind, 1989; Javdani, 2013; Javdani et al., 2011a, 2011b; see also J. Miller, 2008). In tandem, these factors can promote a high-risk environment for delinquency, including through commission of status (e.g., running away, truancy) and violent (e.g., assault) offenses. In addition to having a unique constellation of risk, girls may also be \ characterized by unique factors that can protect against delinquency. Among the most notable include support from a caring adult, academic success, and school connec\ tedness (Zahn et al., 2008; Zahn et al., 2010). While these can also serve as protective factors Javdani and Allen 137 for boys’ delinquency, they are particularly protective for girls (Zahn et al., 2010) and underscore the benefits of developing individual (e.g., with a caring a\ dult) and com- munity-based (e.g., within school) relationships to protect against de\ linquency. The importance of girls’ connections with individuals and key proximal contexts (e.g., peers, schools) are also reflected in research suggesting the critical \ role that relation- ship formation plays for girls’ engagement in disruptive behaviors, especially during mid adolescence (Bardone, Moffitt, Caspi, Dickson, & Silva, 1996; Hinshaw & Kranz, 2010; Werner & Crick, 2004); a link demonstrated prospectively (e.g., Bardone\ et al., 1996; Javdani, Rodriguez, Nichols, Emerson, & Donenberg, 2014). Finally, girls’ pathways to delinquency are also influenced by the response of the juvenile justice system itself, including through implementation of poli\ cies and prac- tices that may affect girls differently than boys (Chesney-Lind, 1989; Chesney-Lind & Pasko, 2004; Chesney-Lind & Shelden, 2004; Javdani, 2013; Javdani et al.\ , 2011b).

Evidence indicates that particular types of behaviors, including status \ offenses, drug use, and perpetration of intimate partner violence result in girls’ formal system involve- ment at disproportionately greater rates as compared with boys (Javdani\ et al., 2011b).

The formal response levied by the justice system may thus be characteris\ tically differ - ent for girls, especially as this system was historically designed to se\ rve a predomi- nately male population (Chesney-Lind & Shelden, 2004). These patterns have been described by criminologists studying women and girls’ offending for the last two decades (e.g., Bloom, 1998; Chesney-Lind, 1989; Kruttschnitt, 1996) an\ d serve to underscore the difficulty that girls have in navigating their involvement in the justice system. Indeed, delinquent girls become extensive users of the juvenile \ and criminal justice systems into adulthood, sometimes without the commission of othe\ r crimes (e.g., through technical violations of probation; Bloom, 1998).

Gender-Responsive Programming: Current Disparities and Design Recommendations Despite mounting evidence suggesting that girls’ disruptive behavior problems and subsequent delinquency are areas of growing concern, there are few gende\ r- responsive intervention programs for girls (Bloom & Covington, 2001; Ch\ esney-Lind & Okamoto, 2001). According to data from the Office of Juvenile Justice and Delinquency Prevention’s Model Program Guide, only 13 programs exist that are designed for female offenders, representing about 6% of total programs. Moreover, of these programs, all but two are characterized as exclusively prevention \ programs, underscoring a need for the development of interventions for girls at hi\ gh risk of dis- ruptive behavior disorders. Data from a national research collaborative,\ called the Girls Study Group, suggest a similar landscape—nationally, there are no more than 26 programs designed specifically for female offenders, most are characterized as pre- vention efforts only, and none meet the criteria for effectiveness. A recent review of gender-responsive programs for girls with disruptive behavior problems describes nine gender-specific interventions (Zahn, 2009; also see National Girls Institute). These programs vary in the specific population of girls they target 138 Feminist Criminology 11(2) (e.g., Girl’s Educational and Mentoring Services [GEMS], sex trafficking; Reaffirming Young Sister’s Excellence, adjudicated African American girls) and the settings in which programs are implemented (e.g., Working to Insure and Nurture Girls’ Success [WINGS], intensive home and family context; Girls Circle, support group)\ . These studies, as well as other research-based and government-led study groups\ , have sug- gested several program elements that characterize a promising and gender\ -responsive approach for girls (Acoca & Dedel, 1998; Cauffman, 2008; Chesney-Lind & Shelden, 2004; Maniglia, 2003; Zahn, 2009; Zahn et al., 2010). Chief among such recommendations are that female-responsive programs be \ rela- tional, including by building and supporting healthy romantic and non-romantic\ rela- tionships; restorative, through helping girls navigate the justice system with attention to their trauma and victimization histories; socio-culturally anchored through attend- ing to girls’ multiple marginalities that arise by virtue of their gender, age, race, and class; individualized and tailored to meet the needs of a heterogeneous group of young women; and multi-level by design, such that girls’ individual, peer, family, community, and multiple system involvements (e.g., child welfare, school, juvenile\ justice) are targeted. In addition to these intervention characteristics, several innova\ tive program models for youth delinquency across genders have highlighted the importa\ nce of designing programs to be community-based and strengths-based (Larson, 2000; Lerner, Phelps, Forman, & Bowers, 2009). That is, programs characterized as most effective are those that “think outside the box” through providing f\ lexible services to youth in their natural communities (community-based), moving away from\ deficit- oriented approaches, and focusing on promoting developmental assets (po\ sitive development). These intervention elements are supported by two theoretical frames. The first is ecological theory (Bronfenbrenner, 1992), which asserts that multiple levels of analy- sis shape human behavior; from individual biology to proximal social env\ ironments (family, schools, communities) as well as distal environments (socio-cultural realities, policies, economic realities). Bronfenbrenner’s theory has been buoyed by decades of empirical support including as it relates to the many factors affecting girls’ risk of juvenile delinquency (Caspi, Lynam, Moffitt, & Silva, 1993; Javdani, 2013). As an extension of this theoretical framework, the intervention was developed \ to actively intervene with girls’ proximal social environments to change the conditions of their lives with the aim of decreasing risk and increasing protective factors.\ The second core framework supporting this approach to intervention is em\ power - ment theory. Empowerment theory posits that individuals, groups, and communities experience greater well-being when they have more control over their lives (Rappaport, 1981). Disenfranchised groups, youth among them, often have less decisi\ on-making control over facets of their lives that directly affect them. In the intervention, decision- making authority regarding the focus of the intervention and the specifi\ c direction(s) it will take is entirely driven by the girls being served. This is essential given that girls have clearly articulated what they would like to see in an intervention (Belknap, Gaarder, Holsinger, McDaniels Wilson, & Cady, 2011). This follows the tradition of feminist studies that have amplified the voices of justice-involved girl\ s and women Javdani and Allen 139 and incorporate the explicit orientation that they are the experts in their own lives.

Indeed, qualitative studies with female offenders have elucidated the ways in which help-seeking behaviors by girls trying to escape abuse are often criminalized, creating social binds that systematically limit girls’ opportunities for escaping victimization and criminal justice involvement (Belknap 2001). Using a model that em\ phasizes girls’ choice has also gained traction in the context of gender-responsive approaches to probation, where gender-responsive service delivery theoretically has the explicit aim of challenging gender stereotyping and oppression (Belknap & Holsinger, 2006; Morash, 2010).

Program Design: The GAP In response to program disparities for juvenile justice-involved girls, \ and their increas- ing arrest and incarceration rates, the GAP was developed, manualized (Javdani & Allen, 2012), and prospectively evaluated for three cohorts of girls wi\ th disruptive behavior problems. The authors developed this intervention by adapting existing, empirically supported models (described below) and were responsible fo\ r the develop- ment of a manual tailored to the adapted intervention with an explicit f\ ocus on girls.

The central goal of the program, and a distinguishing feature compared w\ ith other programs reviewed in the current literature (e.g., Zahn, 2009), is its\ focus on strength- ening girls’ contexts. Specifically, a key feature of the GAP is to systematically and comprehensively identify and attain community resources in a way that is\ driven by girls’ needs and rights, that targets multiple contexts and systems, and that is based on their individual and environmental strengths and assets. Several extant \ research and intervention efforts guided the design of the GAP. Key among them is ecologically based approach from the field of community psychology.

Community-Based Advocacy Grounded in the concepts of evidence-based ecological intervention appro\ aches for youth being diverted from formal system involvement (Adolescent Diversi\ on, see Davidson & Rapp, 1976; Smith, Wolf, Cantillon, Thomas, & Davidson, 2004) and female survivors of domestic violence (Community Advocacy Project, see Sullivan, 1997; Sullivan & Bybee, 1999), the GAP is designed to incorporate specific commu- nity-based advocacy tenets and intervention components. These ecological approaches are particularly promising programs given compelling evidence, including\ through randomized control trial (RCT), that suggest they are effective in reducing risk (e.g., depression, experience of violence, future offending) and promoting well-being (e.g., safety, quality of life; see Smith et al., 2004; Sullivan & Bybee, 1999). Ind\ eed, Sullivan and colleagues’ community-based advocacy for survivors of domestic violence has been included in the Substance Abuse and Mental Health Services Administration’s (SAMHSA) evidence-based practice registry, underscoring the utility and effective- ness of the four-phase advocacy model. The GAP was based on the four-phase advocacy model developed by Davidson and colleagues. Specifically, the youth-driven principles from the Adolescent Diversion 140 Feminist Criminology 11(2) Project (Davidson & Rapp, 1976; Smith et al., 2004) were combined with\ an explicit gender-salient framework grounded in an understanding of gender-based oppression from the community-based advocacy model employed with domestic violence \ survi- vors (Sullivan, 1997; Sullivan & Bybee, 1999). As such, the GAP was designed to be both youth-driven and responsive to gender-based power and oppression. In keeping with the ecological advocacy approach that centers on targeting important contexts (e.g., school, family, peer) through promoting access to needed resources, a central goal of the GAP is to “change girls’ contexts” in a way that promotes positive develop- ment and decreases their engagement with risk-enhancing contexts. The intervention is implemented by connecting girls with highly skilled, intensively trained advocates who engage in four overlapping phases of advocacy over the course of a 6\ -month intervention (described in more detail in GAP Intervention Process). Over the course of all phases, the GAP centralizes youth-driven, strengths-based, and community- centered tenets that characterize the ecological advocacy approach. Thus, girls are fully charged with “driving” their own intervention, meaning that they can s\ et, modify, and introduce new goals (and therefore new resources to seek). Further\ more, every effort is made to systematically incorporate attention to girls’ existing and emerging strengths, and to mobilize resources that can enhance and take advantage of these strengths. Finally, every hour of intervention is conducted in girls’ natural community contexts, including homes, schools, and neighborhoods, or within formal systems of which they are a part (e.g., juvenile detention, child welfare, substan\ ce use).

Female-Responsive Services (FRS) Model GAP’s intervention design also incorporates and is highly consistent with se\ veral national recommendations for FRS for adolescent offenders (Maniglia, 2003). Notably, the GAP is an explicitly relationship-based approach to change, with advocates receiv- ing intensive training on non-judgmental, emotionally supportive relatio\ nship build- ing grounded in humanistic principles of empathic, non-directive communi\ cation (relational; Rogers, 1981). The focus of each relationship is also on identifying and promoting girls’ strengths and mobilizing resources to enhance these strengths (strengths-based). The GAP is also an inherently individualized approach, with advo- cates working with their clients for 6 to 8 hr per week and tailoring th\ e intervention process to meet the unique needs of each girl across a diversity of need\ areas (individu- alized). The GAP also explicitly targets girls’ contexts, including informal (e.g., peer, family) and formal (e.g., school, juvenile justice system) ecological\ levels (multi-level; Javdani, 2013). Importantly, the GAP model is also in keeping with feminist studies of girls’ treat- ment wants and needs (e.g., Belknap, Holsinger, & Dunn, 1997), which suggest that girls highly value respect (during and after incarceration) and fairne\ ss in obtaining privileges (e.g., in comparison with boys), report frustrating or degr\ ading experiences when attempting to obtain services, and report a multitude of needs, inc\ luding sex education, medical care, substance abuse, and mental health concerns (B\ elknap et al., 2011; Belknap et al., 1997). These considerations have been incorporated into Javdani and Allen 141 gender-responsive probation models, which demonstrate some success, and differ from probation “as usual” in that they have higher monitoring, incorporation of gradu- ated sanctions, needs assessment, and multiple phases of treatment inter\ vention, par - ticularly for substance abusing women (Morash, 2010). GAP advocates receive extensive training on girls’ unique pathways to delinquency, including the potential for girls to have experienced victimization, rel\ ationship vio- lence, and unique motivations for engaging in behaviors (e.g., running \ away from home). These behaviors are interpreted through a socio-structural lens that loc\ ates girls’ experiences in a gendered social context and simultaneously recognizes \ the psy- chological sequelae of victimization (restorative; socio-culturally anchored).

Furthermore, the GAP model allows for paraprofessionals, in this case college stu- dents, to be trained as advocates and work one-on-one with girls. This model has been underscored as a particularly promising approach for working with juveni\ le justice- involved girls (Holsinger & Ayers, 2004), because it is inherently relationship-based (Holsinger, 2012), has the potential to promote students’ critical consciousness (Holsinger & Ayers, 2004), and serves as a vehicle through which the university set- ting can positively affect the lives of incarcerated girls (Holsinger, 2008). Indeed, based on research suggesting that frontline providers working wi\ th juvenile justice-involved girls often negatively label and stereotype their clien\ ts, advocate training explicitly targets these perceptions through training and weekly ongoing supervision. Indeed, girls have been negatively labeled by providers (B\ elknap et al., 1997) as “criers, liars, and manipulators” (Gaarder, Rodriguez, & Zatz, 2004), a cate- gory directly in contrast to the strengths-based empowerment approach in\ voked in advocacy. Advocate training directly addresses the potential for the presence of t\ hese stereotypes and emphasizes, instead, the promotion of strengths and capa\ cities as fun- damental components of the intervention. The ideas are reinforced weekly during ongoing supervision. In addition, it is important to note that there are\ no particular goals set by the program for an individual girl, and no assumptions abou\ t what girls should and should not do (i.e., there are no parallels to “terms of \ probation”). Thus, the GAP espouses a model, in which the program is accountable to meet the needs\ of each individual girl, rather than one in which individual girls are accountab\ le to meet pro- grammatic mandates. This is a key element of the program that differs from traditional and corrections-centered services, even more effective models such as gender- responsive probation (Morash, 2010). Finally, and though many advocates identify as White, middle-class university students (though there is more variation\ in terms of socio-economic status)—consistent with the undergraduate demographics of the uni- versity population—social justice training explicitly addressed cultural biases, with a focus on oppression and privilege.

Present Study The GAP was implemented in one county in a large Midwestern state over the course of 3 years. Girls were eligible to participate in the GAP, free of charge, if they ever reported contact with the juvenile justice system or law enforcement wit\ h the permis- sion of their parent or guardian. Data were collected prospectively for each participant 142 Feminist Criminology 11(2) at either two or three points in time to examine the extent to which gir\ ls demonstrated shifts in both positive and negative experiences and behaviors. In parti\ cular, key risk behaviors and experiences were examined, including those related to deli\ nquency (i.e., crime), violence, sexual risk–related behaviors, anger, depression, and anxiety.

Each of these risk factors have been related to health disparities for g\ irls with disrup- tive behavior problems, and underscored as necessary targets of intervention (e.g., Zahn et al., 2010). Furthermore, experiences associated with perceived \ resilience and self-efficacy were assessed to examine the extent to which participation in the\ inter - vention was associated with greater perceived problem solving (also see\ data analytic plan for more detail). A primary goal of this article is to describe a new curriculum based on e\ vidence- based practices (e.g., community-based advocacy, Davidson & Rapp, 1976; Sullivan & Bybee, 1999) and guidelines set forth by the Office of Juvenile Justice and Delinquency prevention (e.g., the National Girls Institute) addressing\ an urgent social issue. Although results presented should be considered preliminary, strengths of the study include prospective data collection, examination of three interven\ tion cohorts over the course of 3 years (e.g., to reduce internal validity threats r\ elated to history), and the innovative design of the intervention itself, targeted to an understudied popula- tion for whom disparities are on the rise. We note that though a traditional RCT design is not used in the current study, the intervention (a) is based on the four-phase advo- cacy program demonstrating efficacy in an RCT study (Sullivan & Bybee, 1999; Sullivan, Nguyen, Allen, Bybee & Juras, 2001) and recently included in the SAMHSA evidence-based practice registry, (b) builds on this advocacy model by incorporating key elements of nationally recommended FRS guidelines (Maniglia, 2003)\ , and (c) investigates risk factors based on theoretically grounded and empiricall\ y supported models of female disruptive behaviors (Javdani et al., 2011a, 2011b). Furthermore, while an RCT is a critical next step for future research, evidence from the present \ study is particularly promising given that each key risk factor examined is expected to increase during the age periods examined for this population. Thus, threats to internal validity regarding maturation and regression toward the mean are highly \ unlikely dur - ing this developmental period for the population under study. Finally, repeated testing effects (over time) are likely minimized because of the nature of the intervention itself, which promotes relationship building and trust via trauma-informed principles of care, a process associated with reduced social desirability biases (Fisher, 1993). The lack of RCT design in this study cannot rule out threats to internal validity and w\ e caution readers to refrain from making causal attributions; however, particular elements of the design, implementation, and evaluation of the GAP program lend credibility to the results presented.

Method Setting and Participants The current study involves data collected across the first 3 years of th\ e program.

During that time, 40 students were trained as advocates and worked with girls for an Javdani and Allen 143 average of 6 months. Girls were recruited directly through juvenile just\ ice agencies, schools (e.g., through school social workers), and community-based age\ ncies working with girls with disruptive behavior disorders. Girls were not mandated t\ o participate in the intervention, and the only incentives provided were for completion o\ f the Time 2 (T2) survey (US$5) and the Time 3 (T3) interview (US$10). To be included in the study, girls must have reported some contact with the juvenile justice system\ or law enforcement for disruptive behaviors. Every girl in the study had involv\ ement with the juvenile justice system prior to enrollment in the program. Results of t\ his study were shared with all advocates and community partners through distribution of\ a commu- nity newsletter and presentations at two community-based meetings, which\ included youth participants. Fifty-two adolescent girls between the ages of 13 and 18 years (M = 15.2, SD = 1.4) participated in the GAP. Of these, all were invited to take part in the evaluation of the program, though participation in research was not required to receive se\ rvices. In total, 51 girls and their respective parents/guardians (98%) agreed to partic\ ipate in research, and 47 (90%) successfully completed the program and the longitudinal e\ valuation in its entirety (3 youth were not able to complete the full program becaus\ e of longer term incarceration in the department of corrections for crimes committed prio\ r to their enrollment in the GAP). The majority of girls identified as Black or African American (73%), most lived with at least one biological parent (75%), and abo\ ut 9% had at least one child currently in their custody. Most girls had not yet completed a high school education (90.2%) and were enrolled in 7th through 12th grade, with th\ e majority at a 9th- or 10th-grade education level (see Table 1). According to their own reports, girls had been arrested an average of about 3 times (SD = 1.96), with a range of zero to nine arrests and were about 13 years of age upon coming to attention of autho\ rities for com- mitting their first offense (SD = 1.81). Girls were also incarcerated in juvenile deten- tion an average of about 3 times (SD = 2.52) with a range of zero to nine incarcerations.

The most common offense reported by girls was assault (60.8%), followed by status offenses (e.g., running away from home (13.9%), theft or property offenses (8.9%), and technical violations of probation (e.g., not complying with probati\ on terms; 6.3%).

About 10% of offenses were included in an “other” category that captured other mi\ nor (e.g., public drinking) as well as serious crimes (e.g., robbery); s\ ee Table 1. Comparisons of GAP participants’ reports on the Youth Risk Behavior Surveillance Survey (YRBSS) suggest they are significantly more likely to report en\ gagement and injury in physical fights, dating violence, forced sexual intercourse, being threatened with weapons, suicide attempts, drug and alcohol use, early sexual activ\ ity, and engag- ing in unprotected sex. As such, they represent girls at the highest risk of continued incarceration and experience of health disparities around violence, sexu\ al health, and substance abuse and dependence.

The GAP Intervention Training of advocates. Advocates were advanced female undergraduate students receiv- ing course credit and were recruited from a large university. Every advocate attended 144 Feminist Criminology 11(2) Table 1. Demographic Information and Program Fidelity and Satisfaction.

Demographic information M (range)SD Age 15.23 (13-18)1.41 Categories% Ethnicity White/Caucasian21.2 Black/African American 3.1 Mixed ethnicity 5.8 Living with biological parent Yes75.0 No 20.4 Any children Yes8.9 No 91.1 Education Less than high school90/2 High school diploma/GED 9.8 Juvenile justice involvement M (range)SD Arrested 2.70 (0-9)1.96 Probation 0.60 (0-5)0.89 Detention 2.63 (0-9)2.52 Age at first arrest 13.16 (10-16)1.81 Offenses categories (%) Assault60.8 Status offenses 13.9 Theft/property 8.9 Probation violation 6.3 Other (e.g., robbery) 10.1 Victim relationship (%) Family11.7 Friend/acquaintance 43.2 Stranger 9.5 No victim 35.6 Program fidelity MSDRange Overall fidelity 1.450.451.00-2.43 Community resources 1.460.461.00-2.50 Emotional support 1.480.551.00-3.00 Youth guided 1.380.411.00-2.25 Strengths-based 1.510.591.00-3.33 Importance of program characteristics Community resources 4.270.892.00-5.00 Emotional support 4.221.041.00-5.00 Youth guided 4.490.792.00-5.00 Strengths-based 4.330.981.00-5.00 Program satisfaction Overall program 1.490.641.00-3.67 Meeting goals overall 1.820.871.00-4.50 Meeting specific goals 1.770.771.00-4.00 Note. Program fidelity range from 1 = strongly agree to 5 = strongly disagree; importance of program characteristics range from 1 = not at all important to 5 = very important; and program satisfaction range from 1 = extremely pleased to 7 = terrible. Javdani and Allen 145 a mandatory orientation session before enrollment in a two-semester serv\ ice-learning course. During the first semester, advocates participated in a 12-week i\ ntensive train- ing emphasizing girls’ pathways to delinquency, the history and struc\ ture of the juve- nile justice system, adolescent development, empathy and active listenin\ g skills, crisis response skills, safety and mandatory reporting, and implementation of the four phases of advocacy (Javdani & Allen, 2012; Sullivan, Sutherland, & Allen, 2003\ ). A particu- lar emphasis of this training was on the unique risks for girls involved\ in the juvenile justice system, including the meanings and motivations of girls’ viol\ ence (e.g., why girls might engage in fighting) and typical offenses (e.g., why girls \ might run away from home or be truant from school). This allowed advocates to be in tune with girls’ unique needs and to engage in assessment and resource mobilization to ta\ rget these needs. Upon completion of the training, and for the duration of the firs\ t and second semester, advocates were required to work 6 to 8 hr per week with and on\ behalf of a single female adolescent client. Over the course of engaging in their in\ terventions, advocates continued to receive intensive supervision through participati\ on in weekly supervision sessions, were provided weekly written supervision feedback,\ and received consistent access to supervisors (including a doctoral candidate course\ instructor [the first author], undergraduate student supervisors who were former advocat\ es, and a PhD-level faculty director [the second author]). Because the focus of the intervention was around increasing girls’ access to com- munity resources and making the community (including the systems in whi\ ch girls were involved) more responsive to girls’ needs and rights, every cohort of advocates also participated in intensive community reconnaissance to research existing local resources. Such resources included education, employment, housing, legal assistance, transportation, health care, material goods, financial assistance, extracurricular sup- port, social support, and services for children. Every local resource targeting these need areas was contacted and a full assessment of their services was conducted and compiled in a GAP resource binder used to guide individual interventions.

GAP intervention process. Advocacy involves four phases (Javdani & Allen, 2012; Sul- livan, 2011). During the assessment phase, advocates typically began the interven\ tion by getting to know their clients and gathering information about clients\ ’ lives, goals, and contexts (e.g., peer dynamics). This information helped advocates understand what clients wanted to accomplish during the intervention period (e.g.,\ enhance grades in school, successfully complete probation, obtain employment, join a sp\ orts team).

Based on this informal and formal assessment, advocates and clients begi\ n the second phase through collaboratively identifying areas of unmet need and mobili\ zing an array of resources tailored to meet these needs. Clients guided the process by\ choosing which resources were contacted and advocates helped to systematically ma\ ke phone calls, arrange in-person meetings, and identify critical individuals who\ could enhance girls’ chances of accessing the resources they needed. Next, in the third phase, each unmet need area was monitored to assess the extent to which resources se\ rved to meet girls’ needs. In cases where an original resource was not sufficiently addressing needs, advocates and clients worked together to identify and mobilize new forma\ l and 146 Feminist Criminology 11(2) informal resources that could more effectively meet girls’ needs. Finally, during the last phase, advocates intensify efforts to transfer “self-advocacy” skills. During this phase, advocates transferred skills and knowledge about advocacy and enc\ ouraged girls to take more active roles so they could implement advocacy on thei\ r own. This phase involved a psychoeducational component whereby advocates helped in\ struct girls on the phases and philosophy of advocacy, as well as an active component whereby girls became more involved in every aspect of the intervention (\ e.g., planning and leading meetings with probation officers, writing their own letters for juvenile court). It is important to note that advocacy phases were typically eng\ aged simultane- ously and in an individualized manner. For instance, assessment of unmet needs con- tinued for the duration of the intervention and the transfer of self-adv\ ocacy skills began as soon as the first contact with a community resource was made. Thus, though the advocacy model undergirding the intervention shares similarities with female-responsive programming and gender-responsive probation, it also includes notable differences from probation-based models. Specifically, though advo- cates in the program often work with probation officers, the goals of advocacy are much broader—indeed, our data suggest that girls report working on go\ als related to acquiring housing (30%), obtaining material goods (e.g., computers; 8\ 9%), health care (73%), education (e.g., tutoring; 70%), employment (41%), extracur\ ricular and cre- ative activities (37%), transportation needs (56%), social support (\ 54%), and safety (e.g., obtaining orders of protection against violent partners; 100%),\ in addition to legal needs, including meeting terms of probation (81%). Thus, advocacy represents a comprehensive approach to identifying and leveraging resources to help m\ eet a multi- tude of girls’ needs, as defined by girls themselves.

Procedures All procedures and measures were approved by the university’s institutional review board. Upon enrollment in the intervention, each youth and a parent or g\ uardian were invited to participate in research and relevant adolescent assent and re\ search consent documents were reviewed and signed by each participating family. Youth and parents were informed that their participation in research was completely volunt\ ary and would not affect whether the youth could receive intervention services. Youth were also informed they would be invited to complete surveys at three different points in time—T1 (Time 1), T2, and T3—each of which was voluntary. Clients were informed that their responses to all surveys would remain confidential and would \ not be shared with their advocates to promote disclosure of both negative and positive\ aspects of their intervention. Youth who agreed to participate with their caregiver’s permission completed several questionnaires before being assigned to an advocate (\ T1), com- pleted a subset of these questionnaires in the middle of their intervent\ ion (T2), and completed a final full set of questionnaires upon completion of the prog\ ram (T3).

Interview sessions, which including administration of surveys, were cond\ ucted by highly trained project staff, separate from the advocates who provided direct inter - vention to girls. Javdani and Allen 147 Measures Program fidelity and satisfaction. At T3, 17 items were designed in previous studies examining the four-phase advocacy model (e.g., Allen, Larsen, Trotter, & Sullivan, 2013) to assess fidelity to advocacy tenets defined by Sullivan and col\ leagues (1999) as being the most important to the intervention (e.g. “I feel suppor\ ted and encouraged by my advocate”; “the advocate in this program asked for my input \ when discussing what services I needed”). Specific subscales of fidelity assessed th\ e extent to which the intervention was focused on obtaining community resources, and was e\ xperienced as emotionally supportive, guided by youth, and strengths-based. Partici\ pants indi- cated how much they agreed with each statement, from 1 (strongly agree) to 5 (strongly disagree). Youth responded to six additional questions assessing how important they perceived each component as being, from 1 (not at all important) to 5 (very impor- tant). Youth also responded to questions assessing how satisfied they were \ with the program overall, meeting their goals in general, and meeting their speci\ fic needs, from 1 (extremely pleased) to 7 (terrible). Fidelity and Satisfaction scales and subscales were internally consistent, with Cronbach’s alphas between .73 and .9\ 5. Finally, the degree to which program dosage was delivered adequately was assessed by \ collecting data from advocates for every intervention week indicating (a) total h\ ours spent on intervention activities, (b) total hours spent with clients face-to-fa\ ce, and (c) total number of face-to-face contacts per week. These data were collected systematically through weekly progress reports completed by advocates and submitted to supervisors on a weekly basis.

Youth offenses and juvenile justice system risk. At T1 and T3, we used a self-reported measure based on official offense categories (Snyder, 2005; Snyder & Sickmund, 2006) to assess the extent to which youth engaged in behaviors for which they were, or could have been, brought to the attention of law enforcement (e.g., \ put them at risk of incarceration, got them in trouble at school; for example, “been s\ uspended from school”). Information was collected about whether youth engaged in a\ range of behav- iors that would put them at risk of juvenile justice involvement, includ\ ing status offenses (e.g., suspended from school, ran away from home), property offenses (e.g., vandalized property, stealing), violent acts (e.g., robbery, assault), other offenses (e.g., sold drugs, engaged in prostitution), as well as behaviors that would v\ iolate an existing court order (e.g., technical violation of probation). Each youth was a\ sked to indicate whether they had engaged in each risk behavior, as well as the number of times each behavior was engaged, both baseline (T1), and “over the course of the intervention” (T3). Three composite variables (at T1 and T3) were constructed to characterize the Proportion of Girls Engaging in any Delinquent Acts, the Proportion of Girls Engag- ing in Status Offenses, and the average Number of Delinquent Acts reported over time.

These data provide an assessment of the extent to which youth offenses may have changed over the course of the intervention.

Youth risk behaviors. At T1, T2, and T3, we administered a self-reported measure adapted from the CDC’s YRBSS system protocol (CDC, Department of Health and 148 Feminist Criminology 11(2) Table 2. Program Outcomes Pre-Intervention (Time 1; n = 51) and Post-Intervention (Time 3; n = 46).

Outcomes assessed at two time points Multivariate F(6, 40) = 16.01, p = .00 Outcome Time 1, M (SD)Time 2, M (SD) Time 3, M (SD) Univariate test Proportion of delinquent girls 4.15 (2.78)—1.04 (0.87) t(1, 45) = 8.51, p < .01 Proportion of status offending girls 2.11 (0.95)—0.72 (0.66) t(1, 45) = 9.44, p < .01 Average number of delinquent acts 17.84 (40.12)—5.22 (11.4) t(1, 45) = 2.07 p < .05 YRBSS physical violence composite 2.36 (0.41)—0.36 (0.11) t(1, 45) = 6.42, p < .01 YRBSS sex risk behaviors composite 0.36 (0.14) —0.27 (0.09) t(1, 27) = 0.85, p = .40 Resilience 53.18 (12.22)—56.5 (10.22) t(1, 42) = −2.05, p < .05 Outcomes assessed at three time points Multivariate F(7, 39) = 6.31, p = .00 Outcome Time 1, M (SD)Time 2, M (SD) Time 3, M (SD) Univariate test Self-efficacy 27.50 (7.38)27.98 (6.31)31.00 (5.570)F(2, 36) = 4.40, p < .05 STAXI state anger 21.42 (10.82)20.61 (8.13)16.14 (1.42)F(2, 38) = 7.28, p < .05 STAXI trait anger 21.70 (7.47)19.57 (5.21)18.64 (5.14)F(2, 38) = 5.26, p < .05 STAXI anger expression out 18.84 (5.84)16.82 (4.44)16.68 (4.62)F(2, 38) = 4.74, p < .05 STAXI anger expression in 16.84 (5.34)16.66 (3.76)15.98 (4.37)F(2, 38) = 0.91, ns Brief Symptom Inventory 12.51 (13.01)11.21 (9.70)6.09 (7.350)F(2, 34) = 7.45, p < .05 YRBSS substance use composite 4.32 (1.76)6.03 (2.23)1.52 (0.56)F(2, 39) = 3.64, p < .01 Note. YRBSS = Youth Risk Behavior Survey; STAXI = State-Trait Anger Expressio\ n Inventory–2. Human Services, 2009), which is a protocol used to assess youth risk beh\ aviors nation- ally (e.g., “Have you ever used marijuana?”). The YRBSS is one of the most compre- hensive measures and informs several areas of risk, including safety beh\ aviors (e.g., driving under the influence of alcohol); other-directed violence (e.g., being in a physi- cal fight, carrying a weapon); cigarette, alcohol, and other drug use (e.g., current fre- quency and severity of use); and risky sexual activity (e.g., unprotec\ ted sex). National and regional norms and confidence intervals provided by the CDC (www.cdc.gov/ HealthyYouth/yrbss/index.htm) allow for directly comparing the risk profiles of\ current participants with a large sample of girls from the same region. All risk categories were assessed at T1 and are used to characterize the current sample and compare girls’\ behav- iors with national averages. Subsets of questions assessing the most rel\ evant categories of risk for a juvenile justice sample were administered at T2 and T3, using the same question and response option format recommended by the YRBSS Specifically, youth risk behavior composites for Violence (frequency of times in a physical fight; carrying weapons), Substance Use (current cigarette use, binge drinking, and marijuana use), and Sexual Risk Behaviors (using alcohol before intercourse, not engaging in protected sex, including using birth control, condoms, etc.) were created and exa\ mined for changes over time. Questions were formulated such that youth reported ab\ out the same time frame (e.g., “in the last month”) so comparisons between T1, T2, and T3 were equivalent. Means and standard deviations are reported in Table 2. Javdani and Allen 149 Resilience. At T1 and T3, we assessed youth’s general perceptions of the extent to which they perceived themselves as being resilient individuals who could\ handle and overcome life challenges using a shortened version of the Resilience Sca\ le (Wagnild & Young, 1993) adapted by Neill and Dias (2001). Items assessed youth ch\ aracteris- tics consistent with the construct of resilience (e.g., “I am determ\ ined”), on a scale from 1 to 7 (1 = agree, 7 = disagree). This scale had high internal consistency at both T1 and T3 (Cronbach’s αs, T1 = .86, T3 = .91). These questions assessed girls’ per - ceived resilience at the time of data collection and did not ask girls t\ o think about a particular period of time.

General self-efficacy. At T1, T2, and T3, we assessed youth’s perceived skills and capac- ity to cope in the face of stress and challenges using the General Self-\ Efficacy Scale, a measure that has demonstrated construct and cross-cultural validity (\ Schwarzer, 1993). Youth responded to questions regarding how efficacious they perceived that their thoughts and actions were if they were going through a difficult time in their lives (e.g., “I can solve most problems if I invest the necessary effort”) using a scale from 1 to 4 (1 = not at all true, 4 = exactly true). This scale had high internal consistency across time points (Cronbach’s αs, T1 = .87, T2 = .90, T3 = .90). These questions assessed girls’ perceived self-efficacy at the time of data collection and did not ask girls to think about a particular period of time.

Internalizing symptoms (Brief Symptom Inventory–18 [BSI-18]). At T1, T2, and T3, we assessed key internalizing tendencies related to psychological distress around depres- sion and anxiety using the BSI-18 (Derogatis, 1992). The BSI was developed as a clinical assessment tool that provides an index of global distress (Glo\ bal Severity Index), as well as subscales informing Somatization (e.g., “nausea \ or upset stomach”), Depression (e.g., “feeling worthless”), and Anxiety (e.g., “feeling afraid”). Youth responded about each symptom for “the past 30 days” on a scale fro\ m 0 to 4 (0 = not at all, 4 = extremely). Questions were formulated such that youth reported about the same time frame (“in the past 30 days”) so that comparisons betw\ een T1, T2, and T3 were equivalent. Although the BSI-18 is typically used with adults aged 18 years and older, recent work has demonstrated the utility of administering this to yout\ h samples (e.g., Piersma, Boes, & Reaume, 1994), given the developmentally appro\ priate lan- guage used in the measure (i.e., items are all at a sixth-grade reading\ level). The BSI had high internal consistency across time points (Cronbach’s αs, T1 = .94, T2 = .87, T3 = .87).

State and trait anger. At T1, T2, and T3, we assessed youths’ experiences, expressions, and capacity to control their state and trait anger, including whether anger is expressed inward or outward, using the State-Trait Anger Expression Inventory–2 (STAXI-2; Spielberger, 1988). Summary scores were calculated using standard guidelines and t\ he STAXI-2 interpretive report (Vagg & Spielberger, 2000). In addition, 11 separate sub- scales were generated around 4 specific areas: State Anger (State Anger, State Anger Feelings, State Anger Verbal, and State Anger Physical), Trait Anger (Trait Anger, 150 Feminist Criminology 11(2) Trait Anger Temperament, Trait Anger Reaction), Anger Expression (Anger In and Anger Out), and Anger Control (Anger Control In and Anger Control Out; see Spiel- berger, 1988, for a full description). Youth rated each item on a scale from 1 to 4 (1 = not at all or almost never, 4 = very much so or almost always). All subscales demon- strated moderate to high internal consistency (Cronbach’s αs, T1 range = .78-.99, T2 range = .66-.95, T3 range = .65-.88). Questions assessed girls’ perceived state anger at the time of data collection (i.e., “right now . . .”) as well as\ trait anger (i.e., “thinking about yourself in general”) so comparisons between T1, T2, and T3 were equivalent.

Data Analytic Plan Descriptive analyses are used to characterize the sample (Table 1) and report program satisfaction and fidelity. Substantive analyses focus on examining the extent to which engagement in the GAP intervention was associated with reductions in key risk experi- ences and behaviors, and promotion of resilience and self-efficacy. To reduce the prob- ability of Type I error, two sets of repeated-measures MANOVA tests were conducted:

(a) one on outcomes assessed at 2 points in time and (b) one on outc\ omes assessed at 3 points in time (see Table 2). Thus, two sets of analyses were used to examine the extent to which the GAP intervention had a holistic impact on key outcomes. Follow-up univariate tests of specific outcomes are reported to delineate specific\ domains of impact, and are organized by substantive area (instead of whether they were assessed at 2 or 3 time points).

Results Program Satisfaction and Fidelity Fidelity was monitored over the course of the intervention through inten\ sive weekly supervision of advocates’ interventions. This included a process by which advocates received weekly written supervision feedback, and provided supervisors w\ ith written logs of their intervention efforts and goals. Formal assessments of fidelity were also gathered at T3. Participants described the program as having a high fidelity impleme\ n- tation along a number of dimensions emphasized in Sullivan’s training manual for advocacy intervention (Sullivan et al., 2003). Specifically, girls perceived the program as being, on average, highly community-based and centralizing access to \ resources, effective in providing emotional support, being youth driven, and focusing\ on youths’ strengths and assets. In addition, participants perceived these principl\ es as being “important” or “extremely important” to the program. Partici\ pants also reported high satisfaction with the program and were on average “pleased” or “\ extremely pleased” with their involvement in the advocacy program overall (87% of particip\ ants), the extent to which the program helped them reach their overall goals (80% \ of partici- pants), and the degree to which they were able to reach the specific go\ als they targeted during their intervention (77% of participants; see Table 1). In addition, according to advocate reports from weekly progress charts, a total of 3,432 intervent\ ion hours Javdani and Allen 151 (about 8 hr per week) were completed, with 1,152 hr of direct face-to-face contact with clients (about 2.5 hr per week), and 712 distinct face-to-face contact\ s (about 1.5 con- tacts per week). Overall, these results suggest that the intervention w\ as delivered with the degree of dosage required.

Repeated-Measures MANOVAs Table 2 reports results of repeated-measures MANOVAs and follow-up univariate analyses. As depicted, girls participating in the GAP intervention demonstrated sig- nificant changes in outcomes over time in the desired directions. Specif\ ically, indica- tors assessed at baseline (T1) and then again post-intervention (T3)\ demonstrated statistically significant improvements over time, F(6, 40) = 16.01, p = .00. Observed power for these analyses ranged from .67 to .93, with an overall average\ effect size ( ηp 2) of .29. Indicators assessed at baseline, mid-, and post-intervention \ also demon- strated significant improvements over time, F(7, 39) = 6.31, p = .00. Observed power for these analyses ranged from .53 to .99, with an overall average effect size ( ηp 2) of .86. These results indicate a holistic intervention effect on reducing risk and distress and promoting self-efficacy and resilience, as elaborated below.

Delinquency and youth risk behaviors. Overall, fewer girls endorsed offending behaviors across every offense category over the course of their participation in the GAP (i.e., at T3), with an average of 22.6% of girls reporting offending behavior at T1 compared with 5.6% of girls reporting these behaviors at T3. Table 2 demonstrates that signifi- cantly fewer proportion of girls reported engaging in all delinquent beh\ aviors, t(1, 45) = 8.51, p < .01, including status offenses, t(1, 45) = 9.44, p < .01, over the course of their participation in the GAP. A closer examination of specific status offend- ing behaviors suggests that running away was the most common behavior se\ lf-reported by girls. Specifically, 37% of girls enrolled in the program reported ever having run away from home in their lifetimes. A comparison of running away in the 6 months pre- ceding the intervention with runaway behaviors during the intervention (\ which lasted approximately 6 months) suggests a significant decrease in running away\ behaviors over time, t (1, 44) = 3.61, p < .01. Thus, results suggest that advocacy is a promising approach to reducing a particularly risky behavior for girls, running aw\ ay from home. Furthermore, girls reported fewer total acts of delinquency post-interve\ ntion com- pared with baseline, t(1, 45) = 2.07, p < .05. These findings are both corroborated and elaborated by the YRBSS results, which suggest that girls engaged in fewer acts of physical violence, including assaults and weapon-carrying over time, t(1, 45) = 6.42, p < .01. In addition, girls engaged in significantly fewer substance use \ behaviors, including cigarette use, binge drinking, and marijuana use over the cour\ se of the inter - vention (Table 2), F(2, 39) = 3.64, p < .01. The only risk-related behavior that did not demonstrate significant change over time was that of sexual risk–rela\ ted behaviors, such that girls enrolled in the intervention were no more likely to enga\ ge in protected sex at baseline compared with post-intervention. This lack of significant change could be due to low variability in reported sexual risk-taking behaviors. Specifically, though 152 Feminist Criminology 11(2) 72% of girls reported being sexually active, 11% reported unprotected sex at T1 and 11% reported unprotected sex at T3.

Resilience and General Self-Efficacy GAP participants reported significantly greater resilience at T3 compared with T1, t(1, 42) = −2.05, p < .05, suggesting greater perceived capacity to cope with challenges and reach long-term goals. This finding is complemented by parallel gains in per - ceived self-efficacy at T3 compared with T1 and T2, F(2, 36) = 4.40, p < .05, suggest- ing that GAP participants perceived a greater capacity to use problem-solving skills in their daily lives (see Table 2).

Internalizing Symptoms GAP participants reported significantly fewer global distress symptoms suggested by a decrease in the BSI Global Severity Index over time, F(2, 34) = 7.45, p < .05. BSI subscales suggest that this linear decrease was driven by reductions in \ anxiety, F(2, 34) = 5.54, p < .05, and depression, F(2, 34) = 8.19, p < .05, but not somatization, for example, headaches, F(2, 34) = 1.76, ns.

State and Trait Anger Overall, GAP participants reported significantly reduced state anger, F(2, 38) = 7.28, p < .05, including through feeling state anger, F(2, 38) = 5.86, p < .05, and feeling like expressing state anger verbally, F(2, 37) = 8.51, p < .05, and physically, F(2, 37) = 4.91, p < .05. A similar pattern was found for trait anger, suggesting a decrease in ten- dencies to experience anger across situations, F(2, 38) = 5.26, p < .05, including reduc- tions in angry temperament, F(2, 38) = 6.84, p < .05, but not necessarily through reductions in angry reactions, F(2, 38) = 0.99, ns. GAP participants were less likely to report expressing anger outward (e.g., through punching walls or fighti\ ng), F(2, 38) = 4.74, p < .05, and more likely to control their anger inward (i.e., self-regulate upon feeling angry), F(2, 37) = 3.68, p < .05. In sum, these findings demonstrate reduced feelings of state and trait anger and increased skill in controlling the\ expression of angry feelings.

Discussion Girls’ increased involvement at all levels of the juvenile justice system unde\ rscores the need to better understand and target girls’ disruptive behaviors—a topic of growing concern for psychologists, juvenile justice service providers, and polic\ y makers. The current study describes an innovative intervention model for girls at ri\ sk of engage- ment in disruptive behaviors and, in turn, justice system involvement, a\ nd provides promising preliminary support for the effectiveness of the program in a real-world setting. Over the course of the intervention, girls reported greater lev\ els of resilience and self-efficacy, and decreased risk and distress, including violence, substance use, Javdani and Allen 153 delinquency, depression, anxiety, and anger. Indeed, results suggest that girls are at decreased risk of engaging in disruptive behaviors due to demonstrated d\ ecreases in perceived distress, including anger and internalizing symptoms, reductions in engag- ing in externalizing and risky behaviors, such as physical fights and aggression, and gains in problem-solving skills, including perceived self-efficacy and resilience and greater ability to control the expression of anger. Although more evidence is needed before concluding that the GAP caused these observed gains, these results are never - theless promising, given that the intervention is designed based on prev\ ious models that have demonstrated effectiveness using RCT designs (Sullivan & Bybee, 1999).

Moreover, the development of the GAP is timely given the urgent need to provide gender-responsive interventions for the growing justice-involved populations o\ f girls and women locally and nationally (Snyder & Sickmund, 2006), and can be used to generate further research using multi-site evaluations and delineating c\ hange mechanisms. Although more research is needed to examine the degree to which particular mech- anisms of change contribute to the gains observed for GAP clients, we draw from multiple promising frameworks to propose potential mechanisms of change \ that are in keeping with the theoretical grounding of the GAP model. Specifically, there are five key mechanisms of change that may have additively or interactively contr\ ibuted to gains. For illustrative purposes, we present a short description of a de\ -identified inter - vention based on a composite of cases, and delineate potential change mechanisms with reference to this example. Trisha is a 16-year-old girl who has been arrested several times for fighting. She lives with her sister and her sister’s boyfriend because her mom uses drugs and is in a violent relationship. Trisha’s sister recently kicked her out of the house for using marijuana and Trisha is facing expulsion from school for truancy and fighting, but says\ that she “just doesn’t care anymore.” Before the end of their first meeting, Trisha’s advocate identified several of Trisha’s strengths. Specifically, Trisha’s fighting often occurred in response to threats made to her by a particular classmate regarding tension over tha\ t classmate’s boyfriend, and her substance use was a coping strategy for dealing with \ her life at home.

Indeed, Trisha’s fights are either directly with current or former male partners, or ot\ her girls competing for these relationships because Trish feels that her worth is “tied up with” whether she can “keep a man.” After getting to know Trisha better, her advocate noticed that she is incredibly bright and reads several books a week and loves t\ o go to the library.

Together, Trisha and her advocate were able to advocate for Trisha at school; she was given the choice to alter her schedule to avoid her threatening classmat\ e and also able to leave class to see the school social worker whenever she felt threatened. She also joined a support group at a local counseling center, and her advocate helped her get a job at the local library. Trisha said that her advocate “really got me” and helped her learn \ how to advocate for herself. Theorized Mechanisms of Change Non-specific or common factors. Because of the strong relationship-based approach to change employed in the GAP, it is likely that part of the gains can be a\ ttributed to the 154 Feminist Criminology 11(2) non-specific or common factors described in the therapy literature, whic\ h hinge on building a strong alliance and rapport (e.g., Asay & Lambert, 1999; Mes\ ser & Wampold, 2002). The relationship-building approach used in the GAP is i\ ndeed grounded in humanist theoretical models of non-directive, client-centere\ d communi- cation, which foreground empathic and active listening, genuineness, and uncondi- tional positive regard (Rogers, 1951). These common factors have long \ been described as necessary and predictive of therapeutic gains (W. R. Miller, Taylor, & West, 1980; Patterson, 1984), and may well account for some of the gains observed i\ n the GAP.

For instance, in Trisha’s case, her advocate invested a large proport\ ion of their inter- vention, especially initially, engaged in active listening and communica\ ting empathic understanding. These steps contribute to the alliance between Trisha and\ her advocate, can promote further trust and information sharing (Gaston, 1990), and \ are highly endorsed by clients in the present study.

Gender-based empowerment. Several theorists and practitioners have suggested that gender-based oppression in girls’ proximal contexts can promote lower functioning and increased distress (e.g., Chesney-Lind & Pasko, 2004; Girshick, 199\ 9). Gender- based oppression is based on power differentials that systematically accord fewer social and economic resources to girls and women, and can become “int\ ernalized” through processes such as learned helplessness and adoption of hegemonic\ self-reject- ing views (Prilleltensky & Gonick, 1996). There is a growing evidence base linking low social status to psychological problems in living (Okazaki, 2009).\ The model employed by the GAP quite directly targets oppression in girls’ lives by attending to gender-salient “crucial contexts” (Javdani et al., 2011a) and having an action orienta- tion geared toward increasing access to desired resources (i.e., changing girls’ con- texts; Allen, Lehrner, Davis, & Javdani, 2013). Furthermore, this action orientation is grounded in a feminist framework based on the Community Advocacy Project model employed with survivors of domestic violence (Sullivan & Bybee, 1999; a\ lso see Allen, Larsen, et al., 2013; Allen, Lehrner, et al., 2013), which deliberately locates barriers in girls’ lives using a gender oppression lens and advocates for policy-level change (e.g., Belknap & Holsinger, 2006). For instance, the multiple fights that Trisha engages in are not viewed solely as a result of intrapsychic factors, su\ ch as impulsivity.

Instead, Trisha’s fighting behavior is located in a social context that places value on \ Trisha as a girl who can “keep a man” (J. Miller, 2008). Indeed, Trisha’s fights are either directly with current or former male partners, or with other girl\ s competing for these relationships. Over the course of advocacy supervision, these beha\ viors are con- sistently interpreted using a gendered lens, and such processes are open\ ly discussed and targeted.

Contextual competence. Given the strong action orientation of the GAP, one of the effects of participation in the program may be to help girls become more e\ ngaged in a greater breadth of contexts. This process has been termed “contextual competence,” and refers to the extent to which youth are actively involved in home, p\ eer, school, community, and cultural contexts (Seidman & Pedersen, 2003), a process linked t\ o Javdani and Allen 155 decreased violence at the community level (Zeldin, 2004). Higher engag\ ement in a greater breadth of contexts, in turn, is associated with greater self-es\ teem and lower depression among at-risk youth (Pedersen et al., 2005). Because GAP clients reported accessing resources across several need areas, participation in the prog\ ram may increase both their level and breadth of engagement in relevant contexts\ . In the case of Trisha, her advocate worked to increase her engagement in school by creat\ ing an indi- vidualized response for in-school arguments and helped create a bridge between Tri- sha and her social worker. Trisha also became an active visitor to the local library, which allowed her to have greater computer and printer access to complet\ e her school- work and complete employment applications online. Finally, Trisha became involved in a local support group and expanded her peer context to include youth \ who were seeking support services and trying to abstain from using substances. In\ tandem, Tri- sha may have become more contextually competent through participation in\ advocacy by becoming more engaged in existing (e.g., school) and new (e.g., li\ brary, counsel- ing) contexts.

Comprehensive, youth-driven systems of care framework. System of care approaches are among the most promising intervention frameworks promoted for youth at r\ isk of social and behavioral difficulties, particularly by the SAMHSA (e.g., systemsofcare.

samhsa.org; also see Cross, Bazron, Dennis, & Isaacs, 1989; Stroul & Friedman, 1986). This approach assumes that youth have multiple needs and life complexiti\ es, and thus supports an individualized and coordinated community-based system, driven by youth, that involves many key stakeholders (e.g., service providers,\ law enforce- ment, teachers) in the response to youth mental health difficulties, and has demon- strated empirical support (though effects do vary across sites; see Bickman, Noser, & Summerfelt, 1999; Vinson, Brannan, Baughman, Wilce, & Gawron, 2001). The advo- cacy model of the GAP seems highly aligned with the system of care framework, given the deliberate efforts of GAP advocates to attend to youths’ multitude of needs, involve critical individuals who can help meet these needs, and coordina\ te services in a way that is both individualized and driven by youth themselves. In Trisha’s interven- tion, a team of important stakeholders was invited to help address Trisha’s school dif- ficulties, and included Trisha, Trisha’s mother, sister, advocate, school resource officer (i.e., school police officer), and her school social worker. Trisha and her advocate prepared for the meeting by writing out a list of Trisha’s goals and used the meeting time to brainstorm how the people “at the table” could help suppor\ t Trisha to reach those goals. The result was an individualized response to fights at school (by the school resource officer and school social worker), and a better understanding of Tri- sha’s daily challenges by her probation officer, including how she is tardy to her first period because of child care responsibilities. This latter realization resulted in a posi- tive court/probation report for Trisha and active efforts by the probation officer and school social worker to generate alternative ways for Trisha to “make up” tardy arriv- als (e.g., arranging with her first period teacher for Trisha to stay after school).

Trauma-informed principles. Given the severe victimization histories of juvenile justice- involved girls, including GAP clients, alignment with developing trauma-informed 156 Feminist Criminology 11(2) practices is essential (Harris & Fallot, 2001; Morrissey et al., 2005; \ SAMHSA National Center for Trauma Informed Care: www.samhsa.gov/nctic/; National Child Traumatic Stress Network: www.nctsn.org/). Although future work with the GAP can deliber - ately incorporate emerging trauma-informed elements, the GAP is aligned with sev- eral current trauma-informed principles, which may have contributed to participants’ gains. Among them are the promotion of safety by attending to victimization risk and engaging in evidence-based safety planning (Sullivan & Bybee, 1999); trustworthi- ness by creating an infrastructure of high accountability for advocates; choice by sup- porting multiple alternative solutions to meeting goals and needs; collaboration by actively involving girls in directing and planning their interventions; \ and empower - ment by following a rights-based model that supposes each girl has a right to positive youth development. In Trisha’s case, she reported multiple traumas related to directly witnessing violence against her mother, being involved in violent fights, and being exposed to community violence through her peer group—risk factors rel\ ated to girls’ disruptive behaviors in particular (Javdani, Abdul-Adil, Suarez, Nichols, & Farmer, 2014). Her advocate worked to enhance her voice and choice over the cou\ rse of the intervention; a process that may have promoted gains in self-efficacy and problem solving.

Strengths and Limitations The current article presents an innovative intervention, the GAP, based on empirically supported models and consistent with several nationally recognized FRS c\ riteria for one of the most underserved and at-risk youth populations—juvenile ju\ stice-involved adolescent girls. The evaluation of the GAP was conducted using a prospective design, limiting threats of recall bias and allowing for risk and protective fac\ tors to be assessed systematically before, during, and after the intervention. Use of multip\ le measures provided convergent validity, given observed gains in both self-efficacy and resilience (protective factors) as well as decreases in depression and anxiety (internalizing), state and trait anger, and self-and other-directed violence and substance use (externalizing risk). There are several limitations that should be considered in light of the \ findings pre- sented. Perhaps most importantly, the current study did not employ a control group design or random assignment to groups due to limited resources, decreasing our ability to make causal inferences about the effects of the GAP on girls’ risk and protective factors. Future research can undoubtedly help strengthen the evidence ba\ se by employ- ing a more rigorous methodological design including random assignment to the inter - vention condition. However, the use of several prospective indicators and multiple measures provides compelling evidence and allows for a fair investigatio\ n of changes in girls’ risk and protective factors over time. Another methodological limitation is the relatively small sample size and implementation of intervention in a single site.

Although future research with larger, multi-site data sets is necessary, we note that participants included almost every girl referred to the GAP over the course of 3 years, so there is almost no self-selection bias characterizing our participant\ s. Also, despite Javdani and Allen 157 having low statistical power, several gains were detected in our outcomes, with effect sizes in the moderate to large category. However, we underscore that data do not sup- port the hypothesis that the GAP intervention is associated with reductions in risky sexual behavior. As noted, this could be due to low statistical power, given low base rates of this risk factor. We encourage future research with a larger participant pools that assess risky sexual activity using methods sensitive to social desi\ rability. In addition, though several relevant risk and protective factors were in\ vestigated, official juvenile justice recidivism data were not available for the curre\ nt pool of par - ticipants, another important avenue for future research. The outcomes assessed in this study, however, include multiple convergent measures as well as behavioral indicators (e.g., decrease in physical fights) that are known correlates of recid\ ivism (Van Dorn & Williams, 2003), suggesting that changes in recidivism could be expected\ on average.

Similarly, our results and fidelity assessments are informed by perceptual, self-\ reported data and may be influenced by single method bias and social desirability\ . The percep- tual measures we employed are among the most oft-used and include state \ and national norms with which to compare our girls’ risk profiles (e.g., the YRBSS), and use of self-report data is consistent with the program’s philosophical orientation of relying on girls’ own accounts (e.g., trustworthiness). It is also notable that girls r\ eported rela- tively high levels of risk behavior suggesting social desirability was n\ ot muting their self-report of these challenges. Finally, intervention hours (total and face-to-face) and number of face-to-face contacts were assessed every week for every advoc\ ate, provid- ing advocate reports that corroborate overall programmatic fidelity. It is also important to consider that while this intervention focuses on being gender- responsive, girls involved in the juvenile justice system are characterized by multiple marginalities, including by virtue of their race, ethnicity, class, and sexual orientation (Maniglia, 2003). Attention to these rays of diversity is a critical next step for the GAP\ and future evaluations of the program. Finally, because our girls represent a forensic sample, we not only caution readers with regard to the generalizability \ of findings but also note that participants reported a range of experiences and risk and\ legal profiles, and spanned early to late adolescence in age. Finally, we note that this particular intervention model relies on student advo\ cates and is perhaps best suited to settings in which university–community partnerships can be fostered. While this raises issues of sustainability of the intervention for other\ types of settings, we note that the Advocacy model has also been implemented with non- student paraprofessional (e.g., “cultural brokers”), leveraging key individuals within communities who already have strong relational ties with potential progr\ am partici- pants (Hess, Barr, & Hunt, 2009). Thus, Advocacy could be implemented and evalu- ated with non-student paraprofessionals and is currently being incorpora\ ted into some traditional human service settings by Sullivan, Allen, and colleagues. In addition, reli- ance on students has the potential to build university–community part\ nerships, which can be mutually beneficial by addressing the training needs of students \ and simultane- ously providing low cost or (in this case) free of charge interventions (Holsinger, 2008, 2012). Indeed, advocacy has been used across a wide variety of universi\ ty-based set- tings with, for example, survivors of domestic violence (Allen, Larsen, et al., 2013; 158 Feminist Criminology 11(2) Sullivan & Bybee, 1999). Moreover, the reliance on students and university-based settings does not inherently limit sustainability per se. Indeed, resear\ ch suggests that the strong collaborative ties between universities and communities can result in highly sustainable models, particularly if advocacy is incorporated into the cu\ rriculum of university settings (Holsinger, 2008). Taken together, this study presents a model for an ecological intervention for girls with disruptive behaviors that place them at high risk of juvenile justi\ ce system involvement. Promising preliminary results support its implementation at\ several stages of risk, including before formal system (e.g., justice, school)\ sanctioning, as an intermediate or immediate response to disruptive behavior, to promote positive com- munity reentry, or as an alternative to incarceration.

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, authorsh\ ip, and/or publication of this article.

References Acoca, L., & Dedel, K. (1998). Outside/inside: The violation of Americ\ an girls at home, on the streets, and in the juvenile justice system. Crime & Delinquency, 44, 561-589.

Allen, N. E., Larsen, S. E., Trotter, J. L., & Sullivan, C. (2013). Exploring the core service deliv- ery processes of an evidence-based community advocacy program for women \ with abusive partners. Journal of Community Psychology, 41, 1-18.

Allen, N. E., Lehrner, A., Davis, S., & Javdani, S. (2013). Gender as ecology: One understand- ing of men’s use of violence against women. In C. Raghavan (Ed.), Domestic violence:

Methodologies in dialogue (pp. 102-116). Boston: MA. Northeastern University Press.

Asay, T., & Lambert, M. (1999). The empirical case for the common fact\ ors in therapy. In M. Hubble, B. Duncan, & S. Miller (Eds.), The heart and soul of change (pp. 37-51).

Washington, DC: American Psychological Association.

Bardone, A. M., Moffitt, T. E., Caspi, M., Dickson, N., & Silva, P. A. (1996). Adult men- tal health and social outcomes of adolescent girls with depression and c\ onduct disorder.

Development and Psychopathology, 8, 811-829.

Belknap, J. (2001). The image of the female victim. The Invisible Woman. Gender, Crime, and Justice. University of Colorado, 207-221.

Belknap, J. (2010). “Offending women”: A double entendre. The Journal of Criminal Law and Criminology, 100, 1061-1098.

Belknap, J., Gaarder, E., Holsinger, K., McDaniels Wilson, C., & Cady, B\ . (2011). Using girls’ voices and words to study their problems. In M. Kerr, H. Stattin, R. Eng\ els, G. Overbeek, & A. K. Andershed (Eds.), Understanding girls’ problem behavior: How girls’ delinquency develops in the context of maturity and health, co-occurring problems, a\ nd relationships (pp. 95-115). Oxford: United Kingdom. Wiley-Blackwell.

Belknap, J., & Holsinger, K. (2006). The gendered nature of risk facto\ rs for delinquency. Feminist Criminology, 1, 48-71. Javdani and Allen 159 Belknap, J., Holsinger, K., & Dunn, M. (1997). Understanding incarcera\ ted girls: The results of a focus group study. The Prison Journal, 77, 381-404.

Bickman, L., Noser, K., & Summerfelt, W. (1999). Long-term effects of \ a system of care on children and adolescents. The Journal of Behavioral Health, 26, 185-202.

Bloom, B. (1998). Women with mental health and substance abuse problem\ s on probation and parole. Offender Program Report: Social and behavioral rehabilitation in prisons\ , jails and the community, 2, 1-13.

Bloom, B., & Covington, S. (2001, November). Effective gender-responsi\ ve interventions in juvenile justice: Addressing the lives of delinquent girls. In Annual Meeting of the American Society of Criminology. Atlanta, GA.

Bloom, B., Owen, B., Deschenes, E. P., & Rosenbaum, J. (2002). Moving \ Toward Justice for Female Juvenile Offenders in the New Millennium Modeling Gender-Specific\ Policies and Programs. Journal of Contemporary Criminal Justice, 18(1), 37-56.

Bloom, B., Owen, B., Rosenbaum, J., & Piper, E. (2003). Focusing on girls and young women: A gendered perspective on female delinquency. Women & Criminal Justice, 14(2/3), 117- 136.

Bronfenbrenner, U. (1992). Ecological systems theory. In R. Basta (Ed.), Annals of child devel- opment – Six theories of child development: Rebised formulations and \ current issues (pp.

187-250). Greenwich, CT: JAI Press.

Bush-Baskette, Stephanie. (2004). “The war on drugs and the incarce\ ration of mothers.” Gendered (in) justice: Theory and practice in feminist criminology, 236-244.

Caspi, A., Lynam, D., Moffitt, T. E., & Silva, P. A. (1993). Unravelin\ g girls’ delinquency: Biological, dispositional, and contextual contributions to adolescent mi\ sbehavior.

Developmental Psychology, 29, 19-30.

Cauffman, E. (2008). Understanding the female offender. Future of Children, 18, 119-142.

Chesney-Lind, M. (1989). Girls’ crime and woman’s place: Toward a feminist model of female delinquency. Crime & Delinquency, 35, 5-29.

Chesney-Lind, M., & Okamoto, S. K. (2001). Gender matters: Patterns in girls’ delinquency and gender responsive programming. Journal of Forensic Psychology Practice, 1(3), 1-30.

Chesney-Lind, M., & Pasko, L. (2004). Girls, women, and crime. Thousand Oaks, CA: SAGE.

Chesney-Lind, M., & Shelden, R. G. (2004). Girls, delinquency, and juvenile justice. Belmont, CA.

Cross, T. L., Bazron, B. J., Dennis, K. W., & Isaacs, M. R. (1989). Towards a culturally compe- tent system of care. Washington, DC: Georgetown University.

Davidson, W. S., & Rapp, C. (1976). Child advocacy in the justice syst\ em. Social Work, 21, 225-232.

Derogatis, L. R. (1992). The Brief Symptom Inventory (BSI): Administration, scoring, and pro- cedures manual II. Baltimore, MD: Clinical Psychometric Research.

Fisher, R. J. (1993). Social desirability bias and the validity of ind\ irect questioning. Journal of Consumer Research, 20, 303-315.

Gaarder, E., Rodriguez, N., & Zatz, M. (2004). Criers, liars, and mani\ pulators: Probation offi- cers’ views of girls. Justice Quarterly, 21, 547-578.

Gaston, L. (1990). The concept of the alliance and its role in psychot\ herapy: Theoretical and empirical considerations. Psychotherapy, 27, 143-153.

Girshick, L. B. (1999). No safe haven. Boston, MA: Northeastern University Press.

Harris, M., & Fallot, R. D. (2001). Envisioning a trauma-informed service system: A vital para- digm shift. New Directions for Mental Health Services, 89, 3-22. 160 Feminist Criminology 11(2) Hess, J. Z., Barr, S. C., & Hunt, G. D. (2009). The practice of family\ mentoring and advo- cacy: A theoretical investigation of critical issues. Families in Society: The Journal of Contemporary Social Services, 90, 189-195.

Hinshaw, S., & Kranz, R. (2010). The tripe bind. New York, NY: Ballantine Books.

Holsinger, K. (2008). Teaching to make a difference. Feminist Criminology, 3, 319-335.

Holsinger, K. (2012). Teaching justice: Solving social justice problems through university edu\ - cation. London, United Kingdom: Ashgate Publishing Ltd.

Holsinger, K., & Ayers, P. (2004). Mentoring girls in juvenile facilit\ ies: Connecting college students with incarcerated girls. Journal of Criminal Justice Education, 15, 351-372.

Javdani, S. (2013). Gender matters: Using an ecological lens to unders\ tand female crime and dis- ruptive behavior. In B. L. Russell (Ed.), Perceptions of female offenders: How stereotypes and social norms affect criminal justice responses (pp. 9-24). New York, NY: Springer.

Javdani, S., Abdul-Adil, J., Suarez, L., Nichols, S., & Farmer, A. D. (2014). Gender differ- ences in the effects of community violence on mental health outcomes in \ a sample of low- income youth receiving psychiatric care. American Journal of Community Psychology, 53, 235-248.

Javdani, S., & Allen, N. E. (2012). The girls’ advocacy project instructor manual. Champaign: University of Illinois at Urbana–Champaign.

Javdani, S., Rodriguez, E. M., Nichols, S., Emerson, E., & Donenberg, G. (2014). Risking it for love: Romantic relationships and early pubertal development confer r\ isk for disrup- tive behavior disorders in African-American girls receiving psychiatric \ care. Journal of Abnormal Child Psychology, 42, 1325-1340.

Javdani, S., Sadeh, N., & Verona, E. (2011a). Expanding our lens: Fema\ le pathways to antiso- cial behavior in adolescence and adulthood. Clinical Psychology Review, 31, 1324-1348.

Javdani, S., Sadeh, N., & Verona, E. (2011b). Gendered social forces: \ An examination of the impact of the justice systems’ response on women and girls’ cr\ iminal trajectories.

Psychology, Public Policy, and Law, 17, 161-211.

Kruttschnitt, C. (1996). Contributions of quantitative methods to the \ study of gender and crime. Journal of Quantitative Criminology, 12, 135-161.

Larson, R. (2000). Toward a psychology of positive youth development. \ American Psychologist, 55, 170-183.

Lerner, J. V., Phelps, E., Forman, Y., & Bowers, E. P. (2009). Positiv\ e youth development. In R. M. Lerner & L. Steinberg (Eds.), Handbook of adolescent psychology (3rd ed., pp. 524- 558). Hoboken, NJ: John Wiley.

Maniglia, R. (2003, May). Meeting the needs juvenile female offenders. Prepared for Juvenile female offenders: How do you meet their needs in your juvenile justice s\ etting? Office of Juvenile Justice and Delinquency Prevention, U.S. Department of Justice, Washington, DC.

Messer, S. B., & Wampold, B. E. (2002). Let’s face facts: Common fa\ ctors are more potent than specific therapy ingredients. Clinical Psychology: Science and Practice, 9, 21-25.

Miller, J. (2008). Getting played: African American girls, Urban inequality, and gendered v\ io- lence. New York: New York University Press.

Miller, W. R., Taylor, C. A., & West, J. C. (1980). Focused versus bro\ ad-spectrum behavior therapy. Journal of Consulting and Clinical Psychology, 48, 590-601.

Moffitt, T. E., Caspi, A., Rutter, M., & Silva, P. A. (2001). Sex differences in antisocial behav- ior. Cambridge, UK: Cambridge University Press.

Morash, M. (2010). Introduction. In C. Renzettti (Ed.), Women on probation & parole: A femi- nist critique of community programs & services (pp. 1-24). Boston, MA: Northeastern. Javdani and Allen 161 Morrissey, J. P., Jackson, E. W., Ellis, A. R., Amaro, H., Brown, V. B.,\ & Najavits, L. M.

(2005). Twelve-month outcomes of trauma-informed interventions for wom\ en with co- occurring disorders. Psychiatric Services, 56, 1213-1222.

Neill, J. T., & Dias, K. L. (2001). Adventure education and resilience\ : The double edge sword. Journal of Adventure Education and Outdoor Learning, 1, 35-42.

Okazaki, S. (2009). Impact of racism on ethnic minority mental health.\ Perspectives on Psychological Science, 4, 103-107.

Patterson, C. H. (1984). Empathy, warmth, and genuineness in psychothe\ rapy: A review of reviews. Psychotherapy: Theory, Research, Practice, Training, 21, 431-438.

Pedersen, S., Seidman, E., Yoshikawa, H., Rivera, A. C., Allen, L., & Ab\ er, J. L. (2005). Contextual competence. American Journal of Community Psychology, 35, 65-81.

Prilleltensky, I., & Gonick, L. (1996). Politics change, oppression re\ mains: On the psychology and politics of oppression. Political Psychology, 17, 127-148.

Piersma, H. L., Boes, J. L., & Reaume, W. M. (1994). Unidimensionality\ of the Brief Symptom Inventory (BSI) in adult and adolescent inpatients. Journal of personality Assessment, 63(2), 338-344.

Rappaport, J. (1981). In praise of paradox: A social policy of empowerment over prevention. American Journal of Community Psychology, 9, 1-25.

Rogers, C. (1951). Client-centered therapy. Boston, MA: Houghton Mifflin.

Rogers, C. (1981). The foundations of a person-centered approach. In C. Rogers (Ed.), A way of being (pp. 113-136). Boston, MA: Houghton Mifflin.

Schwarzer, R. (1993). Measurement of perceived self-efficacy: Psychometric scales for cross- cultural research. Berlin, Germany: Freie Universitat Berlin.

Seidman, E., & Pedersen, S. (2003). Holistic contextual perspectives o\ n risk, protection, and competence among low income urban adolescents. In S. S. Luthar (Ed.), Resilience and vulnerability (pp. 318-343). New York, NY: Cambridge University Press.

Sharp, S. F., & Marcus-Mendoza, S. T. (2001). It’s a family affair:\ Incarcerated women and their families. Women & Criminal Justice, 12(4), 21-49.

Smith, E. P., Wolf, A. M., Cantillon, D. M., Thomas, O., & Davidson, W. \ S. (2004). The adoles- cent diversion project: 25 years of research on an ecological model of i\ ntervention. Journal of Prevention & Intervention in the Community, 27, 29-47.

Snyder, H. N. (2005). Juvenile arrests 2003 (Juvenile Justice Bulletin). Washington, DC: U.S. Department of Justice, Office of Juvenile Justice and Delinquency Preven\ tion.

Snyder, H. N., & Sickmund, M. (2006). Juvenile offenders and victims: 2006 national report. Washington, DC: Office of Juvenile Justice and Delinquency Prevention.

Spielberger, C. (1988). State-Trait Anger Expression Inventory–Research Edition (Professional manual). Odessa, FL: Psychological Assessment Resources.

Stroul, B. A., & Friedman, R. M. (1986). A system of care for children and youth with severe emotional disturbances. Washington, DC: CASSP Technical Assistance Center.

Sullivan, C. M. (1997). Preventing family violence: An experimental intervention—Final per- formance summary report. Atlanta, GA: Centers for Disease Control and Prevention.

Sullivan, C. M. (2011). Community advocacy project manual. East Lansing: Michigan State University.

Sullivan, C. M., & Bybee, D. (1999). Reducing violence using community\ based advocacy for women with abusive partners. Journal of Consulting and Clinical Psychology, 67, 43-53.

Sullivan, C. M., Nguyen, H., Allen, N., Bybee, D., & Juras, J. (2001).\ Beyond searching for deficits: Evidence that physically and emotionally abused women are nurturing parents.

Journal of Emotional Abuse, 2(1), 51-71. 162 Feminist Criminology 11(2) Sullivan, C. M., Sutherland, C., & Allen, N. (2003). Training paraprofessionals to successfully advocate for women with abusive partners. East Lansing: Michigan State University.

Vagg, P. R., & Spielberger, C. D. (2000). State-Trait Anger Expression Inventory–2: Interpretive Report (STAXI-2: IR) Comprehensive handbook of psychological assessmen\ t, 2, 70-86.

Van Dorn, R., & Williams, J. (2003). Correlates associated with escala\ tion of delinquent behav- ior in incarcerated youths. Social Work, 48, 523-531.

Vinson, N. B., Brannan, A. M., Baughman, L. N., Wilce, M., & Gawron, T. (2001). The system- of-care model implementation in 27 communities. Journal of Emotional and Behavioral Disorders, 9, 30-42.

Wagnild, G. M., & Young, H. M. (1993). Development and psychometric ev\ aluation of the Resilience Scale. Journal of Nursing Measurement, 1, 165-178.

Werner, N. E., & Crick, N. R. (2004). Maladaptive peer relationships a\ nd the development of relational and physical aggression. Social Development, 13, 495-514.

Zahn, M. A. (2009). Determining what works for girls in the juvenile j\ ustice system: A sum- mary of evaluation evidence. Crime & Delinquency, 55, 266-293.

Zahn, M. A., Hawkins, S. R., Chiancone, J., & Whitworth, A. (2008). The girls study group— Charting the way to delinquency prevention for girls. U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention.\ Zahn, M. A., Agnew, R., Fishbein, D., Miller, S., Winn, D-M., Dakoff, G.\ , . . . & Chesney-Lind, M. (2010). Causes and correlates of girls’ delinquency. Washington, DC: U.S. Department of Justice, US Department of Justice, Office of Justice Programs, Office\ of Juvenile Justice and Delinquency Prevention.

Zeldin, S. (2004). Preventing youth violence through the promotion of \ community engagement and membership. Journal of Community Psychology, 32, 623-641.

Author Biographies Shabnam Javdani, PhD, is an assistant professor of applied psychology at New York University. Her research involves examining the distinct pathways throug\ h which women and girls engage in violence, and understanding how the systems that respond to violence can detri- mentally influence outcomes for families. Her research is also character\ ized by a social justice focus, with an emphasis on the development, implementation, and evaluati\ on of advocacy, men- tal health, and sociopolitical development programming for young women i\ nvolved in the juve- nile justice system. She has contributed to the scholarly literature via\ numerous articles and is committed to bridging research and action. Her intervention models are c\ urrently being imple- mented within the New York City Department of Juvenile Justice.

Nicole E. Allen, PhD, is an associate professor of community psychology at the Universi\ ty of Illinois at Urbana–Champaign. Her research examines community collabo\ ration and systems change processes with a focus on the community response to intimate part\ ner violence and systems of care development and implementation for youth and families. S\ he has contributed to the scholarly literature via numerous articles and presentations and is committed to bridging scholarship and action by working closely with community partners in bot\ h research and action.

She also directs an advocacy program for women with abusive partners and\ girls at risk of entry into or residing in juvenile detention.