Examining the link between traumatic events and delinquency among juvenile delinquent girls: A longitudinal study

Abstract

Researchers have postulated associations between childhood trauma and delinquency, but few have examined the direction of these relationships prospectively and, specifically, with samples of delinquent girls. The purpose of this study was to examine the relationship between traumatic events and delinquency for girls in the juvenile justice system using a cross-lagged model. Developmental differences in associations as a function of high school entry status were also examined. The sample included 166 girls in the juvenile justice system who were mandated to community-based out-of-home care due to chronic delinquency. Overall, study results provide evidence that trauma and delinquency risk pathways vary according to high school entry status. Implications for future research and practice are discussed.

Childhood victimization and exposure to violence are significant social problems, with an estimated 25% of children being exposed to family, school, and community violence around the world (Ammar, 2006; Cohen & Mannarino, 2008). Although most children who have experienced traumatic events exhibit resiliency and do not develop significant behavior problems, many become vulnerable to serious developmental interruptions and negative long-term consequences including substance abuse, risky sexual behavior, poor academic performance, mental health problems, and delinquent behavior (Brown, Henggeler, Brondino, & Pickrel, 1999; Lansford et al., 2002; Smith, Leve, & Chamberlain, 2006). Youth in detention are at substantially higher risk than youth in the community to have witnessed or been victimized by violence, with high rates of polyvictimization reported among this population (Abram et al., 2004; Wood, Foy, Goguen, Pynoos, & James, 2002). Despite the consistent finding that trauma history is associated with severe and pervasive maladaptation, the relationship between childhood experiences of violence and adolescent delinquency remains inconclusive. For delinquent girls, understanding the interplay between experiences of trauma and delinquency becomes critical given the increase of girls in the justice system, the disproportionate number of girls who experience extreme and repeated violence, and the lack of trauma informed services for this population (Hennessey, Ford, Mahoney, Ko, & Siegfried, 2004; Zahn et al., 2010).

Childhood Trauma Experiences and the Development of Antisocial Behavior

Individuals who have been maltreated or exposed to violence during childhood are more likely to have arrest records during adolescence or adulthood (Cernkovich, Lanctot, & Giordano, 2008; Maxfield & Widom, 1996; Thornberry, Ireland, & Smith, 2001). Further, the association between trauma experiences in childhood and the development of mental health concerns such as oppositional defiant disorder (Flisher et al., 1997; Merry & Andrews, 1994) and conduct disorder (Fergusson, Horwood & Lynskey, 1996) has been well established. Research suggests that timing, duration, and chronicity of traumatic events influence the risk and trajectory of delinquency, with children exposed to trauma at a younger age being at higher risk for poor developmental outcomes (Maas, Herrenkohl & Sousa, 2008; Moffit & Caspi, 2001; Silverthorne & Frick, 1999).

Researchers recently proposed new developmental theories of psychopathology for juvenile delinquents that conceptualize trauma as a pathway to psychological disturbance for adjudicated adolescents. Ford, Fraleigh, and Connor (2010) proposed an integrative theoretical model that suggests traumatic experiences overwhelm the executive functioning capacities of the brain, causing impairment in the mediation of thoughts, behaviors, and emotion. Accordingly, if this state of psychological and physiological alarm continues over time, an adolescent’s resources become depleted and s/he develops a rigid cognitive schema, has a decreased ability to regulate emotions, and learns fewer coping strategies. If the traumatic stress and symptoms continue to be ignored by those who comprise the child’s interpersonal environment, the adolescent then escalates toward “victim coping,” a perceived justification to take any means necessary to avoid revictimization. According to Ford and colleagues (2010), to carry out this defense mechanism, victim coping may result in the loss of empathy, inability to self-regulate, distorted cognitions, lack of impulse control, and other characteristics that can increase an adolescent’s tendency toward delinquent behavior.

Kerig and Becker (2010) review several additional theoretical models and the underlying processes thought to contribute to the development of delinquency within a trauma framework including emotional processing, cognitive processes, and interpersonal processes. The review highlights several promising themes across perspectives including the role of underlying existential experiences such as hurt, betrayal, and anxiety as well as emotional avoidance processes such as disassociation and numbing in the manifestation of adolescent delinquency. Allwood, Bell, and Horan (2011) found that numbing of fearful emotions was most strongly related to exposure to severe types of violence and subsequently with all types of delinquent behavior. Numbing of sadness was most significantly related to aggression only in the context of interpersonal relationships suggesting that interpersonal sadness, detachment, and diminished empathy are important mechanisms in the relationship between violence exposure and violence perpetration for adolescents. Further analysis revealed that numbing of fear and sadness was significantly related to delinquent behaviors only in the context of high arousal symptoms.

Although these theories provide promising ideas about the reciprocal impact of trauma and delinquency, few empirical studies have examined the relationship between these constructs over time. In addition, efforts to reduce and prevent traumatization, delinquency, and recidivism rates are driven by empirical evidence gathered on predominately male samples that exhibit markedly different trauma and delinquency constellations (Fontaine, 2009; Leve & Chamberlain, 2004). Understanding the similarities and differences in the sequelea of delinquent behavior within a developmental perspective becomes essential to identifying risk and protective factors related to the development of delinquency and repeat trauma experiences.

Trauma Experiences and Delinquent Behavior Among Girls

Although more than 90% of delinquent youth have experienced a traumatically stressful life event, delinquent girls are perhaps the most at-risk group of youth. Girls represent the fastest growing segment of the juvenile-justice system, with female delinquent youth representing 30% of all juvenile arrests (Zahn et al., 2010). From 1997 to 2006, arrests for aggravated assaults decreased for boys by 24% and, for girls, only by 10%. According to the Federal Bureau of Investigation’s arrest statistics, during this same time period, arrests for simple assault decreased for boys by 4% but increased for girls by 19% (Snyder, 2008). In addition to the growth in delinquency rates, victimization experiences reported by delinquent girls, in particular extreme and repeated victimization, are exponentially greater than those reported by male juvenile offenders and females in the community (Belknap & Holsinger, 2006; Bender, 2010).

Researchers have noted sex differences in responses to abuse experiences, which may result from the fact that girls are at greater risk than boys for repeated interpersonal victimization such as sexual abuse by family members or being direct targets of community violence (Chamberlain & Moore 2002; Freyd, 2001; Herrera & McCloskey, 2003). Kerig and Bennett (2013) found distinct gender differences in peritraumatic reactions within the context of experiencing interpersonal traumas for a large sample of juvenile justice-involved youth. The authors reported a general trend of increased endorsement of peritraumatic reactions within interpersonal experiences of trauma for all youth; however, girls more often reported symptoms of disassociation in the context of interpersonal traumas while boys more often reported disassociation and confusion in the context of non-interpersonal traumas. The growth in delinquent behavior for girls combined with high incidences and differential experiences of trauma has created momentum in the field to examine the antecedents and trajectories of delinquent behavior specific to girls.

In their review of 46 empirical studies that examined the developmental trajectories of females’ antisocial behavior, Fontaine, Carbonneau, Vitaro, Barker, and Tremblay (2009) found the majority of results indicated that: (a) an early onset/life course-persistent trajectory does exist for females, (b) adolescent-onset of antisocial behavior seems to be more prevalent for females than males, and (c) differences in risk and protective factors that follow certain trajectories of delinquency (i.e., adolescence-limited trajectory and adolescence-delayed-onset trajectory) are inconclusive for females. The results of this study also highlight that females’ antisocial behavior can emerge at different ages although girls within the early onset/life course-persistent trajectory are more likely to have antisocial behavior and other maladjustment problems in adulthood.

The authors suggested several hypotheses that may contribute to the lack of cohesion and consistency in findings related to trajectories for females; however, none of these suggestions incorporate looking at the parallel developmental trajectories of trauma experiences and delinquency. Given the overwhelming correlational evidence that points to high experiences of trauma and related negative outcomes for delinquent youth, it becomes crucial to include a more complex developmental conceptualization of childhood trauma in our investigation of antisocial behavior and delinquency, especially for girls.

Present Study

Despite the important advances made in the field, trauma and delinquency rates for girls continue to rise and a theoretical and empirical framework for understanding the risk and protective factors specific to girls’ antisocial and delinquent behavior does not exist (Zahn et al., 2010). The purpose of this study was to use longitudinal date collected prospectively to examine how traumatic events and delinquency were related over a 12-month period for girls who were mandated to community-based out-of-home treatment programs due to chronic delinquency. We also investigated whether there were differences in the associations between traumatic events and delinquency between girls who entered treatment prior to high school entry versus after high school entry. The transition to high school is a particularly important developmental turning point in an adolescent’s life given the biological, cognitive, relational, and environmental changes that co-occur. It is well documented that an increase in the likelihood of deviant behaviors such as antisocial behavior and substance use occurs during transition to adolescence (Van Ryzin, Stormshak, & Dishion, 2011). In addition, children experiencing transitions that co-occur with cumulative or simultaneous events (i.e., experience of trauma and transition to high school) can significantly alter developmental pathways to risk (Cicchetti & Rogosch, 2002). Consequently, investigating the timing of trauma experiences in relation to high school entry status may provide further insight into a more comprehensive developmental trajectory of trauma and delinquency for girls.

We hypothesized that: (a) traumatic events and delinquency would each show moderate stability over the 12-month study period, (b) traumatic events at T1 would predict delinquency at T2, (c) delinquency at T1 would predict traumatic events at T2, and (d) these associations would vary across age groups, with associations being stronger for girls who had not entered high school at the time of treatment.

Method

Participants

Eligible participants included girls in the juvenile justice system (N = 166) who were mandated to community-based out-of-home care due to chronic delinquency. Participants were enrolled in one of two cohorts of girls who participated in a randomized-control trial (ns = 81 and 85 for cohorts 1 and 2, respectively) conducted in the Northwestern United States between 1997 and 2006 to contrast Multidimensional Treatment Foster Care (MTFC; Chamberlain, 2003) and out-of-home services-as-usual (i.e., typical Group Care; GC). All participants were referred by the criminal justice system and screened for eligibility criteria. Inclusion criteria included the following: (a) female, (b) 13–17 years old, (c) at least one criminal referral in the prior 12 months, and (d) placed in out-of-home care within 12 months following referral. Girls pregnant at the time of referral were excluded from enrollment. The project coordinator randomly assigned (coin flip) girls to MTFC (n = 81) or GC (n = 85) and independent evaluators were uninformed of group assignment at all time points.

Participants’ mean age at baseline (T1) was 15.30 years old (SD = 1.17). Of the total sample, 68.1% of girls identified as Caucasian, 16.9% as Multiracial, 11.4% as Hispanic, 1.8% as African-American, 0.6% as Native American, 0.6% as Asian, and 0.6% as Other/Unknown. At T1, 63% of the girls lived in single-parent families and 54% of the girls lived in families earning less than $10,000 per year. The girls had been removed from their home at T1 and 43% had been returned to their biological parent(s) home at T2 (34% were in a treatment setting, 13% were in detention or jail, 5% were on the run, and 4% were living independently). There were no group differences on other demographic characteristics, delinquency rates, or childhood maltreatment at baseline.

Procedure

Girls and their legal guardians were assessed at baseline (T1) and 12 months post-baseline (T2). At both time points, each girl and her current caregiver participated separately in a 2-hr assessment, which was conducted by staff members who were not informed of group assignment and who were not involved in intervention delivery. Although the present study does not focus specifically on intervention outcomes, we briefly describe the intervention here and include it as a covariate in our analyses.

Multidimensional Treatment Foster Care (MTFC) intervention condition

MTFC (Chamberlain, 2003) is an empirically evaluated, randomized intervention for youth who are involved in the juvenile justice system. MTFC is designed to prevent recurring problems with delinquency and to assist youth with transitioning successfully into the community after treatment. In MTFC, experienced program supervisors with small caseloads (i.e., 10 MTFC families) supervise the clinical staff, coordinate aspects of each youth’s placement, and maintain daily contact with the MTFC parents. The intervention is individualized to meet specific behavior problems and aftercare needs; however, all participants receive the basic MTFC intervention components: daily telephone contact with foster parents to monitor case progress, weekly group supervision and support meetings for foster parents, an in-home, daily point-and-level program, individual therapy, weekly meetings with behavioral support specialists in community settings, family therapy for the aftercare placement family focused on parent management strategies, close monitoring of school behavior, case management to coordinate the interventions, 24-hr, on-call staff support for foster and aftercare parents, and psychiatric consultation as needed.

Control intervention condition

Girls randomly assigned to the typical group care control condition (GC) were placed in community-based programs around the state that represented typical out-of-home care by the juvenile justice system. Each community-based program had between 1 and 41 youth in residence (M [SD] = 10.67 [11.06]). The typical group care program types included residential treatment (45.5%), foster care (39.4%), and other treatment (15.2%).

Measures

Childhood trauma

Traumatic events at T1 and T2 were measured using caregiver report items from the Traumatic Stress Index (TSI; Norris, 1990). The TSI was designed to measure the occurrence of traumatic events using six indicators: (a) being the victim of a robbery, (b) being the victim of an assault, (c) being involved in motor vehicle accident, (d) losing a friend/family member, (e) being the victim of natural or manmade disasters, and (f) other unique traumatic events. At T1, caregivers indicated whether the adolescent participant had experienced each of the six types of trauma over their lifetime (cohort 1) or during the prior 12-months (cohort 2); at T2, caregivers were asked about the prior 12-months (both cohorts). A total trauma event score for each participant was calculated by summing the total scores for each of the six indicators. The range of possible scores was 0–6, with higher scores indicating more types of traumatic events experienced.

At T1, caregiver report consisted of the original parent figure (i.e., biological parent, step-parent, adoptive parent or grandparent) approximately 96% of the time, At T2, approximately 40% of the time it was the original parent figure, and the remaining 60% of the time it was a foster parent, staff person, or other relative. Although the T2 reporter was not the same as T1 reporter approximately 60% of the time, there were no significant differences in mean levels of trauma at T2 as a function of biological parent report (M = .73, SD = .90) or other caregiver report (M = .71, SD = 87), t (137) = .12, p = .91.

Delinquency

A delinquency construct was created to determine the level of criminal/antisocial behavior at each assessment (Chamberlain, Leve, & DeGarmo, 2007). The construct was comprised of three indicators assessing girls’ behavior during the prior 12-month period: (a) number of criminal referrals, (b) number of days in locked settings, and (c) self-reported delinquency, as has been previous used with this sample. In addition, we conducted post hoc analyses for each measure separately, to ascertain whether the pattern of effects seemed to be driven by a specific indicator of delinquency.

Number of criminal referrals was measured using frequency counts of official criminal records from state police and circuit courts. The number of days spent in locked settings was measured using frequency counts of girls’ self report of total days spent in detention, correctional facilities, jail, or prison. Self reported delinquency was measured using the general delinquency subscale of the Elliott Self-report of Delinquency Scale (Elliott, Huizinga, & Ageton, 1985), which is a 21-item self-report measure of delinquency. Each girl was asked how many times she violated certain laws during the preceding 12 months. This measure has shown to have acceptable internal consistency with this sample (α = .91; Chamberlain et al., 2007).

The weighted average of the three indicators (number of criminal referrals, number of days in locked settings and self-reported delinquency) at T1 and T2 was computed to create the total delinquency score. To form this weighted average, the Elliot Self-report of Delinquency Scale items were divided by the max frequency over both waves, creating a 0 to 1 score at each time point for each participant. The other two indicators, number of days spent in a locked setting and number of criminal referrals, were logarithmically transformed and then shifted and divided to fit a 0 to 1 scale based again on the max frequency over all waves. The average of the three scores was then calculated to create the delinquency construct (see Chamberlain et al., 2007). Higher scores indicate a higher rate of delinquency for that participant.

Data Analyses

Two models were examined to test study hypotheses. Model 1, a two-wave, two-variable cross-lagged structural equation model was tested to examine the relationship between trauma experiences and delinquency over time (Figure 1). Model 2 was analyzed to investigate whether there were differences in the associations between traumatic events experienced and delinquency over time between girls who entered treatment prior to high school entry (n = 73; Figure 2a) and those who entered treatment after high school (n = 93; Figure 2b). Due to possible cohort differences and/or a treatment type effect, both cohort and treatment condition were included as covariates in initial analyses.

Figure 1.

Full Model Cross-Lagged Path Analysis

Note. * p < .05. ** p < .01

Figure 2.

Group Based Cross-Lagged Path Analysis

Figure 2a. Pre-High School Entry

Note. * p < .05. ** p < .01

Figure 2b. Post-High School Entry

Note. * p < .05. ** p < .01

The full cross-lagged model created a saturated model allowing for observation of the pathways for the full model as well as pre- and post-high school entry groups. Descriptive statistics for all study variables (i.e., traumatic events and delinquency) were examined including non-normal distributions, skewness and kurtosis. Missing data were accounted for using maximum likelihood estimation (MLR) to provide an unbiased estimate of model coefficients.

Results

Descriptive statistics for trauma and delinquency (overall and by high school entry status) are reported in Table 1 and correlations are reported in Table 2. Inclusion of cohort and treatment variables did not significantly alter model results, and thus, following the principles of parsimony, were excluded from the final models. Examination of variable distributions suggested that variables did not depart substantially from normality. The cross-lagged full model provided an exact fit for the data, χ² (5, N = 166) = 27.01, p < 0.001 and, therefore, fit indices are not reported. When testing missing data assumptions, Little’s Missing Completely at Random (MCAR) test was significant, χ² (11) = 21.24, p = .03, indicating that the data may not be missing completely at random. After comparing the data for girls with (n = 24) and without (n = 142) missing data, however, we found no significant differences on delinquency at T1 or T2 or on trauma at T1 or T2: t(27.29) = 1.65, p = 0.11; t(157) = 0.88, p = 0.45; t(163) = 0.41, p = 0.50; T2, t(142) = -0.45, p = 0.65, respectively. Additionally, non-significant results for the chi-square test of independence were found between missingness and age groupings, χ² (1) = 0.41, p = .52, and missingness and ethnicity, χ² (5) = 3.15, p = .68. These results suggested that it would be reasonable to proceed with the MLR approach for handling missing data.

Table 1.

Means and Standard Deviations for Trauma and Delinquency Construct Scores

Variable	Pre-High school	Post-High school	Overall
Trauma construct
T1	1.79 (0.13)	2.18 (0.13)	2.01 (1.24)
T2	0.82 (0.13)	0.65 (0.08)	0.72 (0.87)
Delinquency construct
T1	0.48 (0.02)	0.47 (0.02)	0.47 (0.17)
T2	0.28 (0.02)	0.21 (0.02)	0.24 (0.20)

Table 2.

Correlation Matrix by Pre-High School and Post-High School Entry Status

Variable	T1 Del	T2 Del	T1 Trauma	T2 Trauma
T1 Del		.367**	.015	−.107
T2 Del	.310**		.018	.123
T1 Trauma	.129	.263*		.209
T2 Trauma	.116	.170	.072

Note. Del = Delinquency. Values for the Pre-high school entry sample are located below the diagonal and values for the post-high school entry sample are located above the diagonal.

^*p < 0.05.

^**p < 0.01.

Cross-Lag Prediction: Full Model

In Model 1, delinquency at T1 significantly predicted delinquency at T2 (β = 0.35, p < 0.001). Traumatic events at T1 did not predict traumatic events at T2. These results suggest that, for the full sample, rank-order levels of delinquency are moderately stable over time, whereas the number of earlier traumatic events does not necessarily predict the number of future traumatic events. Further examination of reciprocal pathways revealed that the association between delinquency at T2 and trauma at T2 was also significant (β = 0.17, p = 0.045).

Cross-Lag Prediction: Developmental Differences Based on High School Entry Status

In Model 2a, for pre-high school entry girls, significant pathways were found between delinquency at T1 and T2 (β = 0.30, p = 0.01). Examination of the cross-lagged paths showed that trauma events at T1 significantly predicted delinquency at T2 (β = 0.24, p = 0.025). In Model 2b, for post-high school girls, delinquency at T1 again predicted delinquency at T2 (β = 0.36, p < 0.001). In addition, trauma events at T1 significantly predicted trauma events at T2 (β = 0.21, p = 0.049). There were no significant associations between trauma events and delinquency found for girls who entered treatment post-high school.

Analysis of each indicator within the delinquency construct (number of criminal referrals, number of days in locked settings and self-reported delinquency) was conducted to determine whether the pattern of results was variable specific. The models for criminal referrals and self-reported delinquency produced similar patterns of developmental differences based on high school entry status. The model using number of days in locked settings deviated from this pattern, primarily due to the lack of stability of this variable over time. Therefore, the composite variable of delinquency utilizing all three indicators was retained.

Discussion

During the past decade, there has been a significant increase in criminal referrals of girls for minor and violent offenses, with a disproportionate number of girls in detention reporting histories of maltreatment. The purpose of this study, therefore, was to examine the relationship between trauma events and delinquency for juvenile delinquent females. Study results point to a prospective predictive relationship between previous trauma event exposure and future delinquency for pre-high school entry girls only. For older girls, results suggest that previous trauma experiences were predictive of future trauma experiences, but not future delinquency. Closer examination of early exposure to trauma during specific developmentally sensitive periods, therefore, may add more clarity to our understanding of girls’ risk for early arrests rates and delinquent behavior.

Study results also extend extant research showing that an increase in trauma experiences may not be associated directly with an increase in delinquency, but rather, may play out through a myriad of individual and environmental influences for which age is an important component. Results also support literature showing that juvenile delinquent girls experience high rates of maltreatment during childhood and continue to experience traumatic events throughout adolescence. Although the sample was restricted in range and variance, with all participants showing a high level of delinquency and trauma experiences, it appears that the timing of trauma experiences may impact the trajectory of future trauma and delinquency differently for older and younger girls.

Limitations

Several limitations of the current study are important to consider. First, although a multi-method and multi-agent report approach was used to measure delinquency, trauma was measured solely by current caregiver report. Youth may have been in different settings at the time of each data collection point, and so it is possible that the caregiver who reporting the traumatic event at each assessment time point may have not been the same. In addition, youth self-report could not be included because it was not available for both cohorts. Reliance on current caregiver report alone does not allow for a comprehensive measurement of girls’ experiences of trauma, especially interpersonal traumas that may include family members. Furthermore, girls may not have trusted or felt comfortable reporting their experiences to their caregivers, and/or it may not have been part of the case file information shared with the caregiver. Although caregivers play an important part in youth’s experience of and recovery from trauma, caregiver-youth disagreement about exposure to traumatic events and subsequent effects on youth outcomes is a growing area of research (Oransky, Hahn & Stover, 2013), especially for youth in detention. Future studies that include caregiver-youth report discrepancies of traumatic events would significantly contribute to a more accurate picture of the impact of trauma on delinquency outcomes for girls.

Second, the trauma construct utilized in this study was a count of traumatic experiences (i.e., robbery, assault) but did not measure interpersonal or relational traumas, which are disproportionately experienced by girls (Chamberlain & Moore 2002; Freyd, 2001; Herrera & McCloskey, 2003). This measurement method also does not include attention to specifics of the traumatic incident such as the severity, frequency, and duration of the event. We included a developmental cut off point for age rather than determining age of initial exposure to a traumatic event. In addition, the timeframe used to assess trauma differed between cohorts, with cohort 1 reporting lifetime trauma and cohort 2 reporting trauma in the past year. Although no significant mean differences on trauma events were reported by cohort, future studies would benefit from a more cohesive measurement of trauma experiences. Future studies could benefit from incorporating a more thorough assessment of age and additional casual mechanisms that may influence the association between trauma and delinquency over time, such as family and peer support, coping strategies (e.g., substance use), cognitive processes (e.g., attributions of stigmatization or shame), and emotional processes (e.g., dissociation, experiential avoidance; Ford, 2010; Kerig & Becker, 2012; Lansford et al., 2006). A more complex representation of trauma experiences will be important to further our understanding of the heterogeneity of abuse experiences that account for individual differences and responses to abuse.

A third limitation refers to the absence of direct correlations between study variables, which reduced the likelihood of accounting for casual relationships across time. It is also important to note that, although findings were in the expected direction, we did not correct for multiple statistical tests given the exploratory nature of this study. Lastly, girls in this sample are locally representative, with the majority of girls self-identifying as European-American. Thus, study results may not generalize to more ethnically and geographically diverse samples. Future studies would benefit from exploring these associations among a more diverse sample as well as using a contextual and ecological approach that allows for the intersection of gender with race, ethnicity, social class, ability, and sexual orientation (Gaarder, Rodriguez, & Zatz, 2004).

Implications for Practice

Results of this study support the use of a developmental conceptualization of trauma as well as the inclusion of trauma experiences in the investigation of delinquent trajectories for girls. Early intervention and assessment practices of delinquent behavior that include a specific trauma focus, and target contextual and individual factors accordingly, will help promote resiliency and reduce future negative outcomes for delinquent girls. As researchers and clinicians are able to more closely identify how children’s responses to trauma are affected by their pre-trauma life experiences and mental health, services can be tailored to preventing children from chronic victimization and participation in antisocial behavior. This is particularly important for girls who have not yet entered high school, a population with significant association between traumatic experiences and later delinquency. Strengthening our efforts to apply a developmental, strength-based, trauma informed framework to research and practice would increase the opportunities for delinquent girls to heal and live healthy and productive lives.