Background

Gender-specific programming has recently emerged over the past decade as an increasingly important issue, in large part because the number of girls involved in the juvenile justice system has been growing over the past decade. Systems and practitioners are working to understand the phenomenon and to identify ways to address the issues this growing population of female delinquents face.

The increasing number and proportion of girls arrested has drawn much attention. While arrests of boys still account for the large majority of arrests, by 2004, 30 percent of all juveniles arrested were girls (Chesney–Lind, Morash, and Stevens 2008). Compared with rates 20 years earlier, this represents a 42.5 percent increase (Chesney–Lind, Morash, and Stevens 2008). Tracy, Kempf–Leonard, and Abramoske–James (2009) note that while the trend for males’ and females’ juvenile arrests similarly increases to 1997 then starts declining, important differences can be observed. For example, female arrest rates declined less than for males (1997–2001), then stabilized (2002–06), while rates for males continued to decline. As a result, where current rates for male arrests are lower than the baseline rates (1980–88), female rates are higher. For property index crimes, female arrests either decreased less or had slight increases. Such trends led Zahn and colleagues to note that the “juvenile ‘crime drop’ of the past decade reflects primarily changes in arrest rates for boys” (2008, 5).

It appears that girls are committing more serious crimes. While girls accounted in 2005 for only 18 percent of juvenile arrests for Violent Crime Index offenses, they accounted for 33 percent of arrests for simple assault (Zahn, Brumbaugh et al. 2008). In 2005 the arrest rate for girls for simple assault was triple that in 1980. What is especially striking in these trends in rates of arrest for girls across multiple categories is that they are significantly different from trends for boys: while rates of juvenile arrests from 1996 to 2005 generally showed a decrease, the rates of decrease for girls were significantly smaller than for boys, and for the category of simple assault, rates increased for girls though they decreased for boys (Zahn, Brumbaugh et al. 2008). Perhaps of even greater concern is the apparent drive in increases (or smaller declines) by females younger than 15. Tracy and colleagues (2009, 191) point out that “very young female delinquents contribute disproportionately to female arrests, as compared to young males,” especially among violent index offenses.

Overall, there has been an increase in delinquency cases being handled by juvenile courts, but again females represent a growing proportion of that caseload. In 2005, juvenile courts handled 1.7 million delinquency cases, which represent a 46 percent increase from 1985 (Livsey 2009). This increase over these 22decades, though, is not linear: the courts experienced a 61 percent increase in caseload between 1985 and 1997, followed by a 9 percent drop from 1997 to 2005 (Sickmund 2009). The drop, though, appears to be explained by the declining rates of male referrals to juvenile courts: referrals for girls increased from 1985 to 1997, then stabilized, while referrals for males starts to decline in 1998 (Tracy, Kempf–Leonard, and Abramoski–James 2009). Between 1985 and 2002, males saw a 29 percent increase in referrals to juvenile courts, while females saw an increase of 92 percent (Chesney–Lind, Morash, and Stevens 2008).

Also of concern are the outcomes of referrals to court: a higher percentage of females’ cases were “petitioned for formal processing and ultimately adjudicated. … Regardless of the year, females were handled more punitively than males” (Tracy, Kempf–Leonard, and Abramoski–James 2009, 195, 200–201). For instance, detention is not only a more frequently exercised option as of 2005 than it was in 1985, but also its growth has disproportionately affected females. The overall growth rate in detained cases for 1985–2005 is about the same as that of the growth in delinquency cases (48 percent increase in detained cases, compared with the 46 percent increase in overall delinquency cases), but the trend line follows a different pattern. Between 1997 and 2005, when the delinquency caseload decreased by 9 percent, there was actually a small increase of 2 percent in the number of juvenile detained (Sickmund, Sladky, and Kang 2008). Thus, the proportion of cases leading to detention grew. In 2006, more than 90,000 youths were in residential placements (Sickmund, Sladky, and Kang 2008). Over this period, though, the proportion of girls detained increased: during the 1989–98 period, a surge in the number of female delinquency cases produced a 56 percent increase in the detention of females compared with a 20 percent increase for males (Zahn, Brumbaugh et al. 2008). Tracy and colleagues remark that the data consistently show “girls are much more likely than boys to receive the harshest sanction available in a juvenile court—that is, commitment to a juvenile prison—for status offenses and even for technical violations of probation” (2009, 202).

But while it appears that the severity of juvenile offenses has increased, there is also evidence that girls—despite media portrayals to the contrary—are not becoming more violent. To better understand the delinquency trends for girls, the Office of Juvenile Justice and Delinquency Prevention (OJJDP) convened the Girls Study Group to investigate available evidence on girls’ delinquency, why they might become delinquent, and whether there are interventions that work well with girls. The group concentrated on diverse data sources to determine to what degree increasing arrests resulted from increasing delinquency by girls and to what degree they resulted from official responses to delinquency. After examining data from the FBI’s Uniform Crime Reports, the Monitoring the Future study, and the National Crime Victimization Survey, the group determined that self-report data presented convincing evidence that girls’ delinquency behaviors had not increased, but that official responses such as arrest laws and changes in law enforcement policy largely accounted for the changes in official statistics. Evidence suggested that rates of arrest for both boys and girls were increased by mandatory and pro-arrest policies, but that these policies disproportionately affected girls. One explanation for this finding is that such policies lower the threshold for classifying and reporting assaults; domestic disputes that might once have been classified as status offenses might be classified as simple assault, resulting in arrests because of mandatory arrest laws (Zahn, Brumbaugh et al. 2008). Since girls more than boys tend to fight with family members, this law enforcement policy would affect girls more frequently than boys (Zahn, Brumbaugh et al. 2008).

Detention has often been used to handle the increasing number of girls in the juvenile justice system, but this appears to be an inappropriate response. Most detention centers are devoid of elaborate and comprehensive treatment services (Wordes and Jones 1998) that this population of girls needs.

Theoretical Context

Advocates of gender-specific programming note that girls differ developmentally from boys. Following elementary school, self-esteem drops significantly more for girls than for boys (Chesney–Lind and Sheldon 1998). As girls enter adolescence, they encounter a variety of stressful changes (physical, emotional, and psychological). They become more preoccupied with identity, appearance, family, and peer relationships (Greene et al., 1998). Girls often begin to ignore their sense of self, instead placing more importance on personal relationships (Debold, Wilson, and Malave 1993) and counting on others for validation (Taylor, Gilligan, and Sullivan 1995). At this stage in development, girls may begin to step back from competitive situations, fearing that distinguishing oneself presents a risk of being disliked (American Bar Association and National Bar Association 2001). “They begin to see themselves as others see them, and they orient their thinking and themselves toward others” (Debold, Wilson, and Malave 1993). Perceptions of self-worth, physical appearance, and social, academic, and athletic competence often sink to low levels (American Bar Association and National Bar Association 2001). While such behavior is normal, it serves as a breeding ground for certain risk factors in female populations: prior victimization, substance abuse, mental illness, spousal abuse.

Research has suggested that girls involved in the juvenile justice system do indeed have different profiles from boys’. Girls have higher rates of physical, emotional, and sexual abuse in their histories (Bloom et al. 2002; Zahn et al. 2009). To escape highly dysfunctional homes, girls will run away, which is one of the most prevalent risk factors for girls’ ultimate involvement with the juvenile justice system. Although girls and boys run away at about the same rate, girls are arrested more frequently than boys are for this status offense (Bloom et al. 2002). Girls appear to have greater odds of co-morbid mental health conditions and are particularly associated with major depression, posttraumatic stress disorder, separation anxiety, and disruptive disorders (Huefner and Mason 2009; Vincent et al. 2008; Zahn et al. 2009). Girls tend to be younger when detained than boys, have been detained for less-serious crimes or status offenses, and have higher rates of family dysfunction (Tracy, Kempf–Leonard, and Abramoski–James 2009; Zahn et al. 2009). Laurie Schaffner, in her ethnographic study of system-involved girls, insists that “the vast extent of emotional injury in the form of sexual and violent assault that young women in this population report experiencing cannot be understated” (2006, 2).

Changes in the way girls are handled in the juvenile justice system were encouraged by the language of the 1992 reauthorization of the Juvenile Justice and Delinquency Prevention Act, which prohibited gender bias and ensured that girls would have adequate access to services (Zahn et al. 2009). Foley notes that two general theoretical positions characterize much of the literature on gender-specific programming. One is the “ feminist pathways theory [which] argues that childhood events, particularly traumas, are precursors to risk factors for (usually) girls’ and women’s offending behaviors” (2008, 263). The other category is underpinned by “relational-cultural theory [, which] argues that girls’ and women’s development is based on connection with others and the relationships of meaning to delinquent girls should be addressed through prevention and treatment approaches” (2008, 263). Foley’s review of gender-specific programs notes that the few programs with a theoretical grounding were based on the latter theory, though most had no theoretical grounding at all. This lack is problematic, because traditional “correctional programming for female offenders has been based on profiles of male criminality or pathways to crime” (Covington and Bloom 2003, p. 10). Most research has concentrated on male populations—perhaps unsurprisingly, since males form the large majority of those in detention (85 percent in 2006).

Another barrier to effective gender-specific programming (especially for practitioners) is the apparent disconnect between the gender-specific literature, which largely concentrates on identifying the unique causes of female delinquency, and the “what works” literature, which concentrates on principles of effective intervention (Hubbard and Matthews 2008). Hubbard and Matthews (2008) note the differences that characterize the two literatures in terms of guiding principles and substantive areas (e.g., theoretical foundations, program goals, consideration of risk; assessment techniques, therapeutic approach). Of course, the effort to identify particular risk (or protective) factors largely fits into the popular risk/protective factors model (see e.g., Catalano and Hawkins 1995; Hawkins, Catalano, and Miller 1992; Hawkins et al. 2000) used to identify effective programs. OJJDP’s Girls Study Group examined the available literature on girls and identified factors that may predict or prevent delinquency. Many of these factors apply equally to both boys and girls; others seem particularly influential for girls. Those identified as equally influential for both sexes included family dynamics, involvement in school, the level of neighborhood disadvantage, and the availability of community-based programs. Factors that seem more to affect girls’ delinquent behaviors include early puberty (which can lead to increased conflict with parents and associations with older boys or men), sexual abuse or maltreatment, depression and anxiety, and romantic partners. Other factors that may work to support resilience—at least for some behaviors—include the presence of a caring adult, school success, and religiosity. School connectedness appeared not to function as either a risk or protective factor (Zahn, Hawkins et al. 2008). One study found that, for girls, anxiety disorder was identified as a risk factor for recidivism while dysthymia was found to have a protective influence (Plattner et al. 2009). One study disputes the notion of a female-specific pathway to serious, violent, and chronic offending: Johansson and Kempf–Leonard (2009) looked at a sample of 10,405 youths, one third of whom were girls, and concluded that Howell’s (2003) hypothesis that five risk factors are more important for girls is not empirically supported.

Evidence of Impact

Several recent reviews on gender-specific programming suggest that the evidence is thus far weak for the effectiveness of gender-specific programming. For instance, Chesney–Lind, Morash, and Stevens (2008) conclude after an assessment of eight girl-specific programs that “knowledge of what works for girls has been little advanced from when Lipsey (1992) examined the literature and showed minimal relevant evaluation research” (2008, 178). Zahn and colleagues (2009) found in their review of program evaluation evidence that there are few findings based on rigorous methodologies. Of the 62 programs they identified as serving only girls and specifically targeting delinquency or system-involved girls, they were able to identify only 18 that had at least one evaluation. Nine of these were for system-involved girls. Two evaluations used a randomized control design; two used a quasi-experimental design with control groups; the remainder used before-and-after measures on selected variables. They conclude that these evaluations offer mixed evidence about the effectiveness of such programming, with the two randomized control evaluations showing no evidence for long-term impact on recidivism. They also looked at programs used with both boys and girls and found that there is evidence that comprehensive programs that address multiple risk factors (e.g., Multisystemic Therapy, Multidimensional Treatment Foster Care) can be effective in reducing recidivism. They point out that this does not mean that gender-specific programming does not work, but that the development and evaluation of such programs are in their infancy.

Zahn and colleagues (2009) also note that the evaluations they found failed to distinguish between different groups of girls for whom programs do/do not work. Given recent work on trajectories of crime, such information could be invaluable. Colman and colleagues (2009), for instance, prospectively tracked 499 girls discharged from juvenile justice facilities from 16 to 28 years of age. They identified four distinct early adult offending paths: rare/nonoffending, low chronic, low rising, and high chronic. The four groups can follow very different paths, varying from low recidivism rates to very high, chronic recidivism. Being able to identify girls who might follow one trajectory or another has implications for program design and evaluation. If a program can shift a system-involved girl from a chronically recidivating track to a low recidivating trajectory, that might not be captured by certain evaluation designs.

While the evidence for the effectiveness of gender-specific programming is still relatively scant, this could be attributed to two interrelated problems. First, programs for girls are scarce. Many have noted the lack of gender-specific, culturally competent services available to those in the juvenile justice system (Bloom et al. 2002; Gaarder, Rodriguez, and Zatz 2004). Chesney–Lind, Morash, and Stevens (2008) attribute the growth in use of detention to the lack of available alternatives. Also, research indicates inadequate training on how to work with female juveniles for those in the juvenile justice system. Bloom and colleagues’ statewide assessment in California found a clear need on the part of judges, prosecutors, and public defenders for more information and training on working with minor female offenders and on meeting their needs. They conclude: “This education and training should include gender difference in delinquency, substance abuse education, the developmental stages of female adolescence, and available programs and appropriate placements and limitations” (2002, 547). Gaarder, Rodriguez, and Zatz (2004) found a similar lack of information and training. Their research suggested that many in the juvenile justice system function on the basis of racial and gender stereotypes, which prevents girls in the system from getting access to needed services. They found that girls are seen as “criers, liars, and manipulators,” and they documented the ways that such stereotypes deny the realities these girls are facing. Second, the overall number of girls in the system is small compared with the number of boys, which makes evaluation activities more challenging (Tracy, Kempf–Leonard, and Abramoski–James 2009).

Matthews and Hubbard (2009) identify five elements that could be used to develop effective programs for girls—programs that could help shed more light on the promise of gender-specific programming. These elements are

1.      The use of assessments (Austin, Johnson, and Weitzer 2005), for the importance of objective assessment tools)

2.      The incorporation of a therapeutic or helping alliance, which speaks to the desire for girls to have “someone to speak to” (Chesney–Lind, Morash, and Stevens 2008; Schaffner 2003) within a collaborative relationship

3.      The use of a gender-responsive cognitive–behavioral approach

4.      The promotion of healthy connections, which would speak to the relational–cultural theory of gender development and programming (Foley 2008)

5.      The recognition of within-girl differences, which could help distinguish subgroups of girls for whom particular programs are effective (Zahn et al. 2009)

These elements can help bridge the divide between the “what works” literature and the gender-specific literature.

The support for evaluations of gender-specific programming will help ensure that programs are evaluated with rigorous designs. There are programs that show promising results. For instance, Movimiento Ascendencia (Upward Movement) was established in Pueblo, Colo., to provide girls with positive alternatives to substance use and gang involvement. Girls in the treatment group showed a greater reduction in delinquency than girls in the control group during the preprogram and postprogram periods. Similarly, Girls’ Circle is a strengths-based support group that addresses the specialized needs of girls ages 9–18 by integrating relational–cultural theory, resiliency practices, and skills training into a specific format designed to increase positive connection, personal and collective strengths, and competence in girls. Evaluations found significant increases (compared with pretest scores) in posttest body image scores, perceived social support, resiliency through bonding to school, and level of self-efficacy and decreases in self-harming behavior and alcohol use.