enf ap 3

International Journal of Offender Therapy and

Comparative Criminology 1 –33

© The Author(s) 2015 Reprints and permissions:

sagepub.com/journalsPermissions.nav DOI: 10.1177/0306624X15613992

ijo.sagepub.com

Article

The Relationship Between Adverse Childhood Experiences and Recidivism in a Sample of Juvenile Offenders in Community- Based Treatment

Kevin T. Wolff1, Michael T. Baglivio2, and Alex R. Piquero3

Abstract Adverse childhood experiences (ACEs) have been identified as a key risk factor for a range of negative life outcomes, including delinquency. Much less is known about how exposure to negative experiences relates to continued offending among juvenile offenders. In this study, we examine the effect of ACEs on recidivism in a large sample of previously referred youth from the State of Florida who were followed for 1 year after participation in community-based treatment. Results from a series of Cox hazard models suggest that ACEs increase the risk of subsequent arrest, with a higher prevalence of ACEs leading to a shorter time to recidivism. The relationship between ACEs and recidivism held quite well in demographic-specific analyses. Implications for empirical research on the long-term effects of traumatic childhood events and juvenile justice policy are discussed.

Keywords adverse childhood experiences, juvenile offenders, recidivism

1The City University of New York, New York City, USA 2G4S Youth Services, Tampa, FL, USA 3University of Texas at Dallas, Richardson, USA

Corresponding Author: Kevin T. Wolff, John Jay College of Criminal Justice, The City University of New York, 524 W. 59th Street, Room 2109N, New York, NY 10019, USA. Email: [email protected]

613992 IJOXXX10.1177/0306624X15613992International Journal of Offender Therapy and Comparative CriminologyWolff et al. research-article2015

2 International Journal of Offender Therapy and Comparative Criminology

The juvenile justice system was founded on the premise that youth are malleable and rehabilitative efforts paramount. Recently, cries for reform have stemmed predomi- nately from developmental neuroscience, maturity, and criminal career research (Cauffman, 2012; Farrington, Loeber, & Howell, 2012; Gibson & Krohn, 2012; Loeber & Farrington, 2012; Monahan, Steinberg, & Piquero, 2015; Woolard, 2012). During the last decade, the U.S. Supreme Court has handed down rulings preventing the death penalty for juveniles (Roper v. Simmons, 2005), life without parole for non- homicide (Graham v. Florida, 2011) and homicide offenses (Jackson v. Hobbs, 2012; Miller v. Alabama, 2012) committed by juveniles. Arguments have been made for a middle-tier young adult (18-24 years of age) justice system or “maturity discounts” in sentencing (Farrington et al., 2012). Furthermore, juvenile justice reforms have called for limiting juvenile residential program placements to the most severe “high-risk” offenders and realigning resources to community-based alternatives (Baglivio, Greenwald, & Russell, 2015). These reforms are premised on the hypothesis that com- munity-based alternatives yield better outcomes for youth served, or at minimum avoid the potential criminogenic effects of incarceration (see Gatti, Tremblay, & Vitaro, 2009; Loughran et al., 2009). As juvenile offenders have been shown to have higher rates of adverse childhood experiences (ACEs) than the general population (Dierkhising et al., 2013; Evans-Chase, 2014) and are 13 times less likely to have no ACEs (Baglivio, Epps, Swartz, Huq, & Hardt, 2014), examining the effect of trau- matic childhood events on subsequent offending of adjudicated juveniles is paramount if intentions to reduce recidivism are to be optimized.

The current study assesses the extent to which cumulative exposure to trauma, that is, the ACE score, relates to recidivism in a large sample of juvenile offenders who completed community-based services. Before we present the results of our investiga- tion, we provide a brief overview of the documented (non)success of juvenile services, distinguishing community-based services from deep-end, residential placements. Next, we discuss how ACEs may relate to juvenile recidivism, while highlighting the lack of official offending/re-offending studies using the composite ACE score. We then define measures and outline our analytic approach, followed by a presentation of results. We end by highlighting some future directions and policy implications regard- ing juvenile offenders.

Trends in Recidivism Patterns

Juvenile justice reform has been driven by fiscal considerations, as well as philo- sophical debate, but also by the growing evidence base regarding the ineffective- ness of deeper-end placements.. Recidivism rates of youth released from residential placement are large. The State of Florida, which provides the data used in the current study, is one state that has consistently reported juvenile offender recidi- vism rates separately by placement type, such as probation, day treatment, and residential services. The average recidivism rate (conviction of new offense within 1 year) for Florida juvenile offenders completing residential placement has ranged from 41% to 46% over the last 5 years (Florida Department of Juvenile Justice

Wolff et al. 3

[FDJJ], 2011, 2012, 2013, 2014, 2015a). The recidivism rate for probation super- vision has varied slightly from 18% to 19% during the same time period.

In juvenile justice systems, youth are often afforded the chance at probation/ community supervision prior to placement in residential programs, as per the Comprehensive Strategy for Serious Violent Offenders and progressively gradu- ated sanctions (Howell, 2009; Wilson & Howell, 1993).1 If progressive systems should use structured decision-making tools, such as a disposition matrix and graduated response options favoring least restrictive alternatives and providing community-based services prior to considerations of residential placement, the question turns to considering the types of characteristics that are related to a juve- nile offender successfully completing probation. As indicated above, the height- ened exposure to traumatic childhood events among juvenile offenders necessitates the examination of the effects of that trauma on subsequent offending. This is especially the case as juvenile probation recidivism rates within Florida have not declined, even as juvenile arrests have decreased 36% from 2009 to 2013 (FDJJ, 2015b).

One would be remiss to not highlight arguments against solely (or at all) con- sidering recidivism as a marker of probation success (Dilulio, 1992; Petersilia, 1993). One perspective in this debate concerns whether the focus of community corrections (such as probation) should be surveillance/control as punishment, or rehabilitation/treatment and service. Potential intermediate outcome measures for probation have been proposed, such as collection of fines, fees, and restitution; monitoring community service hour completion; or securing legal employment. Specifically for juveniles, arguably staying in school or earning a high school diploma, gaining employment, or refraining from drug use are admirable goals and may result in a more positive life trajectory. Arguments have been made that eval- uating any outcome, such as recidivism, after the supervision of the offender is complete is inappropriate, at least with respect to adult community corrections (Petersilia, 1993). For example, police, prosecutorial, and court success are not measured on offender behavior post-arrest or post-judgment, yet perhaps unfairly correctional services are. Nonetheless, returning to the notion that the juvenile justice system was predicated on a philosophy of rehabilitation and that it is never too late to rehabilitate a juvenile offender (Cullen, Vose, Jonson, & Unnever, 2007; Nagin, Piquero, Scott, & Steinberg, 2006; Piquero, Cullen, Unnever, Piquero, & Gordon, 2010), lowering recidivism rates and delaying re-offending more gener- ally is germane. While arguments may persist, the predominance of the Risk– Need–Responsivity (RNR) paradigm in adult and juvenile corrections brings risk reduction, service provision, and, as a result, recidivism to the forefront (Andrews, Bonta, & Wormith, 2011; Howell, Lipsey, & Wilson, 2014; McGrath & Thompson, 2012; Peterson-Badali, Skilling, & Haqanee, 2015; Vose, Lowenkamp, Smith, & Cullen, 2009). As recently articulated by the Pew Center on the States (2011), “successful efforts to improve public safety and control corrections costs should start with defining, measuring, tracking and rewarding correctional agencies’ per- formance in terms of recidivism reduction” (p. 27).

4 International Journal of Offender Therapy and Comparative Criminology

ACEs as an Offending Risk

Research in several disciplines has noted the importance of ACEs as being a risk factor for a wide range of negative outcomes (Agnew, 1992; Anda, Butchart, Felitti, & Brown, 2010; Widom, 1989). Youth are exposed to traumatic experiences which can have repercussions on neurological development as well as mental health. The preva- lence of child maltreatment is higher in dysfunctional and maladaptive families (Hertzig, 1983; Moffitt, 1993), and exposure to traumatic life experiences is higher in disadvantaged communities (Baglivio, Wolff, Epps, & Nelson, 2015) increasing the likelihood that most of these victims never access needed services and, as a conse- quence, exhibit a worsening of their health (Anda & Brown, 2010; Bronfenbrenner, Moen, & Garbarino, 1984; Gilbert et al., 2009; Mercy & Saul, 2009).

A study of 17,421 privately insured, well-educated adult patients first identified specific negative childhood events, termed ACEs, which were found to be positively associated with several leading causes of death (Felitti et al., 1998). The 10 adverse experiences which stem from this seminal study are emotional abuse, physical abuse, sexual abuse, emotional neglect, physical neglect, violent treatment toward mother, household substance abuse, household mental illness, parental separation or divorce, and having an incarcerated household member (Centers for Disease Control and Prevention, 2015). The ACE concept acknowledges the complex and cumulative nature of risk factors through the process of summing risk factors and associating the composite score with relevant outcomes developed by Rutter (1983). The ACE score is expressed as the sum of the 10 exposure types, each measured dichotomously, such that an exposure is counted as one point regardless of the number of incidents, longev- ity, or the severity of the exposure to that type.

The concept of a composite score is central to understanding the effect of ACEs. This is based on the high prevalence of ACEs and level of interrelatedness among ACEs. Empirical research has supported these theoretical assertions documenting interrelatedness (Anda et al., 2010; Baglivio & Epps, 2015; Dong et al., 2004), further supporting the concept that examining individual trauma type exposures separately, or to ascertain unique effects of a few, misses the broader context in which they occur. This research on interrelatedness shows that exposures to additional ACEs are more likely given exposure to any particular ACE, suggesting that ACE exposure is non- random. The cumulative assessment of multiple trauma types is necessitated by this non-randomness, which leads to the tenet that an individual’s composite ACE expo- sure is central, rather than which particular ACE leads to what outcomes. The current study uses the 10-item ACE score, as it is the score espoused by the Centers for Disease Control and Prevention (CDC; 2015) and is backed by multi-disciplinary research.

The ACE score is simple, conceptually. Specific exposures are assessed in a simple binary fashion and summed to arrive at a composite score, regardless of frequency or severity of exposure. The empirical research demonstrating the relevance of the ACE score on negative outcomes is extensive (as reviewed below). A dose–response rela- tionship has been found with a higher cumulative stressor composite ACE score indi- cating increased odds of heart disease, cancer, chronic lung disease, skeletal fractures,

Wolff et al. 5

and liver disease (Anda et al., 2010; Anda et al., 2006; Chartier, Walker, & Naimark, 2010; Dube, Felitti, Dong, Giles, & Anda, 2003). More relevant to the current study is the effect of ACEs on short-term outcomes closer to the abuse exposure. A retrospec- tive cross-sectional survey of 1,500 randomly sampled individuals (stratified by eco- nomic disadvantage) aged 18 to 70 in the United Kingdom found that higher cumulative ACE scores increased the odds of smoking, heavy drinking, and morbid obesity, as well as increased risk for poor educational and employment outcomes, recent inpatient hospital care, incarceration, and recent involvement in violence (Bellis, Lowey, Leckenby, Hughes, & Harrison, 2014). Specifically with respect to violence and crimi- nal behavior, higher ACE scores were significantly associated with (a) a higher preva- lence of both hitting someone and having been hit within the last 12 months (especially for those with four or more ACEs) as well as (b) having spent at least one night in jail in the last 12 months (more than 8 times more likely for those with four or more ACEs; Bellis et al., 2014).

Studies examining the impact of ACEs (as measured with the composite ACE score) on behaviors in adolescent samples as opposed to relying on long-term retro- spective recall of childhood abuse in adult samples have yielded similar results. Hamburger and colleagues found that among seventh- through 12th-grade students, those witnessing domestic violence, or having a history of physical or sexual abuse, were up to 3 times more likely to have early-onset alcohol use (Hamburger, Leeb, & Swahn, 2008). Examining six types of ACEs on more than 130,000 students, Duke et al. found that each additional type of ACE exposure increased the risk of violence perpetration by 35% to 144% (Duke, Pettingell, McMorris, & Borowsky, 2010), including both interpersonal violence (including delinquency, weapon-carrying, fight- ing, bullying, and dating violence) and self-directed violence (attempted suicide, self- mutilation). The future public health repercussions for at-risk juveniles are exacerbated in comparison with negative life events evidenced in prior ACE work, as higher preva- lence of ACEs has been found in special populations, such as children of alcoholics (Dube et al., 2001) and juveniles with justice system involvement (Baglivio, Epps, et al., 2014), when compared with the mostly middle-class original ACE study population.

Although the studies reviewed above point to a link between traumatic childhood events and antisocial behavior, much less research has examined those exposures as a predictor of time to re-arrest within a recidivism framework. While 50% of the 0-10 ACE score is composed of examples of childhood maltreatment (physical abuse, sex- ual abuse, emotional abuse, physical neglect, and emotional neglect), the remaining 50% is composed of traumatic childhood events, not specifically maltreatment. As no prior work has examined ACE scores and time to failure, we generate hypotheses from studies on individual childhood maltreatment exposures.

Benda (2005) examined time to failure of 600 adult men and women boot camp graduates and found that exposure to childhood physical and sexual abuse predicted a shorter time to recidivism. Both abuse exposures were, however, stronger predictors for women than for men. Examining hazard models for violent re-arrest in a sample of almost 2,000 California Youth Authority parolees, Lattimore, Visher, and Linster

6 International Journal of Offender Therapy and Comparative Criminology

(1995) found that serious youthful offenders with a history of intra-family violence or abuse, or parental neglect, were more likely to be re-arrested for violence, and these exposures predicted a shorter time to committing a subsequent violent offense. In a study of serious juvenile sex offenders, Miner (2002) found that youth who were vic- tims of sexual abuse (of which 86% of the sample reported) were at decreased risk for non-sexual recidivism in early adulthood, while childhood neglect was insignificant in predicting time to recidivism. In a recent analysis examining recidivism for first-time violent juvenile offenders, Ryan and colleagues found that youth with open child wel- fare cases exhibited an increased risk in recidivating early compared with those youth without open cases, even after controlling for age, gender, race, and disposition (Ryan, Abrams, & Huang, 2014).

Additional studies of official maltreatment have found youth, and males in particu- lar, who experience maltreatment are more likely to exhibit violent behavior and delin- quency (Barrett, Katsiyannis, Zhang, & Zhang, 2014; Chen, Propp, deLara, & Corvo, 2011; Cicchetti & Manly, 2001; Yu-Ling Chiu, Ryan, & Herz, 2011). Childhood mal- treatment (as measured by ever being placed in Child Protective Services) was predic- tive of recidivism in a sample of 96,565 juvenile offenders (Barrett et al., 2014). Child abuse/neglect exposure has been shown to double the risk of arrest for violent offenses for girls (Maxfield & Widom, 1996). Experiencing childhood physical abuse and other maltreatment leads to higher self-reported total, violent, and property offending, even after controlling for prior delinquency (Teague, Mazerolle, Legosz, & Sanderson, 2008). Longitudinal, prospective studies report similar findings. Smith and colleagues found that a history of childhood maltreatment significantly increases the chances of involvement in delinquency as measured by both official and self-reported delin- quency (increasing the risk of being arrested, the frequency of arrests, and more seri- ous and violent forms of self-reported delinquency; Smith & Thornberry, 1995). Prospectively examining the effects of childhood physical abuse on later intimate part- ner violence using the Seattle Social Development data evidenced a strong direct effect of the abuse for males, yet the quality of a female’s relationship with an intimate partner mediated the effect of childhood abuse on later violence to a partner (Herrenkohl et al., 2004). Using data from the prospective Lehigh Longitudinal Study, Moylan and colleagues found youth exposed to child abuse and/or domestic violence at increased risk for internalizing and externalizing outcomes in adolescence (Moylan et al., 2010). Further analyses focusing on a cumulative stressor approach showed that only those youth with dual exposure were at elevated risk compared with non-exposed youth. Gender differences were insignificant.

What has not been previously examined is the relationship between the cumulative stressor ACE score and official offending among a high-risk group such as juvenile offenders. In addition, prior work has not examined the possibility of a relationship between ACE exposure and time to re-offending among delinquent youth. Based on findings from the childhood maltreatment studies reviewed above, we hypothesize that higher ACE scores will be predictive of increased re-offending, specifically a faster time to recidivism than juveniles with lower ACE scores. Few studies have examined whether abuse and exposure to domestic violence affect males and females

Wolff et al. 7

differently, despite expressed interest in gender differences (Edleson, 1999; Herrenkohl, Sousa, Tajima, Herrenkohl, & Moylan, 2008; Sternberg, et al., 1993). With respect to gender differences in maltreatment exposure among justice-involved youth, females have reported higher levels of exposure to sexual assault and interpersonal victimiza- tion, though males do report higher rates of witnessing violence (Cauffman, Feldman, Waterman, & Steiner, 1998; Ford, Chapman, Hawker, & Albert, 2007; Wood, Foy, Layne, Pynoos, & James, 2002). Examining trauma histories of 658 justice-involved youth, Dierkhising and colleagues found relatively similar rates of exposure to each of 19 different types of traumas (though females had significantly higher rates of sexual abuse and sexual assault; Dierkhising et al., 2013). With respect to justice system involvement, males who experience maltreatment are prone to violent behavior and delinquency (Chen et al., 2011; Mass, Herrenkohl, & Sousa, 2008; Yu-Ling Chiu et al., 2011). Other studies have found that significantly more maltreated females (including all forms of abuse) committed violent offenses as juveniles or adults than non-maltreated females, while no significant differences in prevalence rates of violent offending were found for maltreated versus non-maltreated males (Herrera & McCloskey, 2001; Widom & Maxfield, 2001). Additional prior work examining an offending population and physical abuse in particular has not found sex differences for heightened risk of violent offending (Teague et al., 2008).

Almost no ACE studies have examined ACE prevalence differences across gender or race/ethnicity in justice-involved samples. Examining gender differences in ACE exposure using the same measure employed in the current study and a different sample of Florida juvenile offenders, Baglivio, Epps, et al. (2014) found similar ACE preva- lence rates across gender, with the exception of sexual abuse where the female rate was more than 4 times that of males. Recent work has highlighted the inattention to race/ethnic differences in offending, charging the field with increasing its focus on race/ethnic relevancy (Piquero, 2015), and such a focus is apt in light of findings that community disadvantage and concentrated affluence affect ACE exposure (Baglivio, Wolff, Epps, & Nelson, 2015), conditions which disproportionately affect racial and ethnic minorities. As well, knowledge of how ACE varies (or does not) across sub- groups is important for the development of more effective intervention strategies, as there may need to be gender and/or race-specific interventions if in fact ACE is not a general risk factor. Therefore, even descriptive knowledge on this front is important. As well, researchers (see Chesney-Lind, 2001; Chesney-Lind & Sheldon, 1992) have argued for gender-specific programming, while other researchers (see Unnever & Gabbidon, 2011) have articulated race-specific theories (for offending) in general, with a focus on African American youth. The current study aims to address those gaps, by examining the effect of ACE on time to failure and racial/ethnic and gender differ- ences in the effect of ACEs on time to recidivism.

Common Risk Factors

Several risk factors that have been shown in a multitude of prior work to effect recidi- vism must be controlled so as to assess the true effects of ACE on re-offending.

8 International Journal of Offender Therapy and Comparative Criminology

Specifically, age at first offense, antisocial peer associations, substance use, school behavior, history of running away, and the worst prior offense will be discussed (which we control for in our analyses, as indicated below). More than 20 longitudinal studies have reported a significant relationship between early onset and later crime, with simi- lar findings for both males and females (Howell, 2009; see also Krohn, Thornberry, Rivera, & LeBlanc, 2001). Youth classified as early-onset offenders have a 2 to 3 times higher risk of later violence, serious offenses, and chronic offending, and are more likely to carry weapons, become gang involved, and use drugs or alcohol (Howell, 2009, 2012; Krohn et al., 2001; Loeber & Farrington, 2001; Loeber et al., 2003). The link between early age of onset of offending behavior and higher frequency and greater seriousness of crime has been consistently demonstrated (DeLisi & Piquero, 2011; Loeber & Farrington, 1998; Tolan, 1987; Tolan & Thomas, 1995). Delinquent peer association has consistently been shown as one of the strongest predictors of delin- quency risk (Akers, 1998; Osgood, Wilson, O’Malley, Bachman, & Johnston, 1996). Juvenile offenders who associated exclusively with antisocial peers or gang members (as per dichotomous measure) were found more likely to belong to Florida juvenile offending prevalence trajectory groups which began offending earlier in adolescence (Baglivio, Wolff, Piquero, & Epps, 2015). Additional prior work using samples of Florida juvenile offenders has indicated that school problems and a history of running away both significantly predict re-offending (Baglivio, Greenwald, & Russell, 2015). Research findings show offenders report higher substance use rates, and substance users have higher rates of offending (Rosenfeld, White, & Esbensen, 2012). Substance use is included in the “Big Eight” risk factors of the RNR model promoted by Andrews and Bonta (2003) upon which many risk/needs assessment instruments have been con- structed (such as the Community Positive Achievement Change Tool [C-PACT] used in the current study). Similarly, many reviews of the predictors of juvenile recidivism suggest that substance abuse represents a salient predictor of continued criminal involvement (Cottle, Lee, & Heilbrun, 2001; Dowden & Brown, 2002). Based on this prior work, and as indicated in reviews of the literature on risk factors for juvenile recidivism (e.g., Loeber & Farrington, 1998, 2012), it is expected that those juveniles with earlier age of onset, antisocial peer associations, more serious prior offending, difficulties in school, and substance use will be more likely to recidivate and to do so more quickly.

Although the current study is not an empirical assessment of a specific theoretical framework, the focus on ACE (and its components) is consistent with several crimi- nological (and psychological) theories including in particular Agnew’s (1992) General Strain Theory. Specifically, Agnew expanded the role of strain to be three- fold: (a) the presentation of noxious stimuli, (b) the removal of positive stimuli, and (c) the failure to achieve positive goals. The ACE measure focuses mainly on the first two of these, with its introduction of negative and/or adverse events and experiences within the youth’s lives. In fact, Agnew specifically notes types of abuses and neglect as consistent with several types of strains, and existing literature has empirically tested some of these types of abuses and neglect within his framework (see Agnew, 2001; Reid & Piquero, 2013). In addition, Agnew (2001) postulated that strains

Wolff et al. 9

interpreted as unjust or voluntary/intentional are more likely to lead to antisocial behavior. We argue that this is most likely the case with many adverse childhood experiences.

Current Study

The purpose of the current study is to examine the effects of ACE exposure on the time between completion of a community-based juvenile justice service and re-arrest among a multi-year statewide sample of juvenile offenders in the State of Florida. The use of time to failure provides more details than a simple dichotomous indicator of recidivism because it specifically focuses on the time to re-offending. There is ample literature indicating that the timing of failure is an important outcome because those youth who recidivate in less time may be different (i.e., have different types and/or levels of risk) than those who recidivate later (i.e., a youth who re-offends 5 days after release is likely to be different from one who does not recidivate until 300 days after release). The classic works by Maltz (1984) and Schmidt and Witte (1988), as well as the juvenile offender recidivism literature examining release cohorts from the California Youth Authority by Visher, Lattimore, and Linster (1991) and North Carolina juvenile training school releases (Dean, Brame, & Piquero, 1996), highlight the importance of such an investigation. A traditional dichotomous analysis that focuses only on the likelihood of recidivism may fail to give proper weight to indi- vidual characteristics that affect the timing of that event (Schmidt & Witte, 1989). We also explore the effects of ACE exposure on time to failure across gender and race/ ethnicity. Importantly, we control for common risk factors as mentioned in the review above as well as the type of service the juvenile completed preceding a 365-day fol- low-up period, to determine whether any potential effect of the ACE score remains a salient predictor of recidivism. We group these additional risk factors into demo- graphic, individual risk factors, and personal history risk domains. We examine the significance of ACE on time to recidivism independently, then in conjunction with demographics, before adding the risk measures of each domain (individual followed by personal history) in a third comprehensive model of juvenile recidivism. In short, the current study extends prior research by examining (a) the relationship between ACE scores and official recidivism, (b) time to failure using the ACE score, and (c) variation in the ACE–recidivism relationship across gender and race/ethnicity.

Data

Data were drawn from the FDJJ archival data records. The FDJJ maintains a centralized database, the Juvenile Justice Information System (JJIS), that contains complete social, offense, placement, and risk assessment history data for all youth referred for delin- quency (equivalent to an adult arrest). The current study uses 3 years of FDJJ commu- nity-based service completions. The individual-level measures of interest were taken from the C-PACT risk/needs assessment panel used by the FDJJ. Data for this study are inclusive of all youth within Florida who completed a FDJJ community-based service

10 International Journal of Offender Therapy and Comparative Criminology

between July 1, 2009, and June 31, 2012, that were administered by the Full C-PACT. This resulted in a final sample of 27,867 unduplicated youth.2

The C-PACT is a fourth-generation actuarial risk/needs assessment designed to classify youth according to risk to re-offend levels of low, moderate, moderate–high, and high, as well as rank-order criminogenic needs. The pre-screen version has 46 items, while the Full assessment has 126. Both produce an identical overall risk rating (if an individual were administered a Pre-screen and a Full assessment, the overall risk level would be the same). Results of the assessment are integrated into individualized case plans for each youth. The C-PACT has undergone several validation studies, all using different samples of Florida juvenile offenders with a total sample in excess of 130,000 (Baglivio, 2009; Baglivio & Jackowski, 2013; Baird et al., 2013; Winokur- Early, Hand, & Blankenship, 2012). The predictive ability of the C-PACT, as mea- sured using the commonly reported area under curve (AUC) score, ranged between a low of 0.59 and a high of 0.632 in those evaluations, with similar AUC scores across gender, race, and dispositional placement. Few other instruments have been examined for the same geographic area with multiple samples (Schwalbe, 2007), especially for juvenile offenders. The most recent reliability study (Baird et al., 2013) used video- taped interviews and an offense history file to assess reliability across raters, finding an intra-class correlation coefficient (ICC) of .83 for the overall C-PACT risk level, and only 4% (five items) with less than 75% agreement with an “expert” rater. Reliability of the C-PACT is enhanced in that all criminal history components (such as age at first arrest, number of adjudicated misdemeanors, felonies, and prior place- ments) are automated from the JJIS database and not dependent on the recall of the youth. Non-criminal history items are based on self-report and corroborated when possible (such as attendance and school grades with teachers and parents). All staff administering the C-PACT must complete a standardized 2-day motivational inter- viewing training plus a standardized 3-day C-PACT and case planning training.

The exit Full C-PACT assessment (the risk/needs assessment administered just prior to the youth completing the service) was used, rather than the first or C-PACT reassessments during service, to ensure that we captured each youth’s risk immedi- ately proximate to the 1-year recidivism follow-up. Using only the Full C-PACT was necessitated as the C-PACT Pre-Screen does not contain all items needed to create ACE scores. The final sample represented 21% of all community-based completions over the study period.3 While using youth assessed with the Full C-PACT leads to our analytic sample being of somewhat higher risk, 55% of the sample was assessed as low or moderate risk to re-offend on the C-PACT.

Measures

Days to Failure

The FDJJ JJIS centralized database maintains records of all official delinquency refer- rals for law violations. For the current study, official offending (recidivism) is mea- sured as subsequent delinquency referral (equivalent to an adult arrest) for a new law

Wolff et al. 11

violation within 12 months from the day of service completion. Therefore, each youth was tracked exactly 365 days from the date that youth completed the community- based FDJJ service. The first arrest for each youth during the 365-day period was counted as recidivism, with the number of days from service completion to first sub- sequent arrest as days to failure. Youth who were not re-arrested within the 365 days post-completion were “censored” in the survival analyses (discussed below) regard- less of whether that youth had an arrest outside of the 365-day follow-up. There are no technical or non-law violations to consider, as all youth completed supervision preced- ing the follow-up period.

Independent Variables

ACE score. Although created to classify youth according to levels of risk to re-offend, the C-PACT risk/needs assessment captures items related to ACEs. The Full C-PACT contains items to create all 10 ACEs of the composite ACE score. These C-PACT items were used to create ACE scores for each youth. The exact items, responses, and coding used to create ACE scores from C-PACT data have been reported elsewhere (Baglivio, Epps, et al., 2014). Each exposure was binary (yes/no), and exposures were summed for a cumulative ACE score ranging from 0 (unexposed to any) to 10 (exposed to all 10 categories).

In contrast to ACE studies with adults, the current study suffered less from the chal- lenges of retrospective recall of childhood events as the exposures are more contem- porary for the young sample. In keeping with prior ACE studies in the social and medical sciences (Dong et al., 2004; Dube, Williamson, Thompson, Felitti, & Anda, 2004), we ascertained the following 10 ACEs: emotional abuse, physical abuse, sexual abuse, emotional neglect, physical neglect, family violence, household substance abuse, household mental illness, parental separation or divorce, and household mem- ber incarceration.

Different from previous studies of ACE using FDJJ data, the current study uses the Full C-PACT assessment immediately prior to completion of the FDJJ community- based service for each youth to create the ACE score. This ensures appropriate time order of having the ACE score prior to the re-offending measure. A brief description of each ACE and responses indicating being exposed are as follows: (a) emotional abuse—Parents/caretakers were hostile, berating, and/or belittling to youth; (b) physi- cal abuse—The youth reported being a victim of physical abuse was victimized or physically abused by a family member; (c) sexual abuse—The youth reported being the victim of sexual abuse/rape; (d) emotional neglect—The youth reported no support network, little or no willingness to support the youth by the family, or that youth does not feel close to any family member; (e) physical neglect—The youth has a history of being a victim of neglect (includes a negligent or dangerous act or omission that con- stitutes a clear and present danger to the child’s health, welfare, or safety, such as failure to provide food, shelter, clothing, nurturing, or health care); (f) family vio- lence—The level of conflict between parents included verbal intimidation, yelling, heated arguments, threats of physical abuse, and domestic violence, or the youth has

12 International Journal of Offender Therapy and Comparative Criminology

witnessed violence at home or in a foster/group home; (g) household substance abuse—Problem history of parents and/or siblings in the household includes alcohol or drug problems; (h) household mental illness—Problem history of parents and/or siblings in the household includes mental health problems; (i) parental separation/ divorce—Youth does not live with both mother and father; and (j) incarceration of household member—There is a jail/prison history of family members.

In addition to the ACE measure, the following measures included in the current study are grouped into domains to assess the importance of each domain in examining time to failure. We group measures into demographic, individual risk factors, and a personal history risk factor domain. All items are collected by the C-PACT risk/needs assessment administered by the FDJJ or maintained in the JJIS centralized database.

Demographics. We include age, gender, race, and ethnicity as demographic controls. Age was measured continuously as the youth’s age at date of completion of the FDJJ service. Gender was dichotomous (female = 0, male = 1), while race/ethnicity is mea- sured using a set of dichotomous variables with 1 = Black, 1 = Hispanic, 1 = “Other” (with White being the reference group in each instance).

Individual risk factors Antisocial peer association. A self-reported antisocial peer association measure of the

youth’s friendship network was used (=1 if youth reported having exclusively antiso- cial peers or associating with gang members, else = 0). Prior research has used a single self-reported item for gang membership (Melde & Esbensen, 2011; Sweeten, Pyrooz, & Piquero, 2013) and has indicated the validity of self-report measures for gang mem- bership (Krohn, Ward, Thornberry, Lizotte, & Chu, 2011; Thornberry, Krohn, Lizotte, & Chard-Wierschem, 1993; Thornberry, Krohn, Lizotte, Smith, & Tobin, 2003). The peer dichotomy variable of the current study has been used in prior work examining Florida juvenile offenders and found predictive of recidivism (Baglivio, Jackowski, Greenwald, & Howell, 2014). This dichotomous measure has also been found pre- dictive of membership in juvenile offending prevalence trajectories with an earlier age of onset and more juvenile offending (Baglivio, Wolff, Piquero, & Epps, 2015). In addition, the utility of antisocial peer measures as a risk of offending has been demonstrated across subgroups, including juvenile and adult offenders, as well as across gender and race/ethnicity (see reviews in Akers, 1998; Loeber & Farrington, 1998), a finding which has held across both male and female Florida juvenile offend- ers (Baglivio, 2009).4

Substance use. Substance use is measured using two items from the C-PACT, one measuring past alcohol use, and the other measuring past drug use. The items were combined to create the current substance use measure, coded 0 for no past use (of either alcohol or drugs), 1 for past use, and 2 for past use where such use caused problems in family conflict, health, pro-social peer associations, withdrawal, increased tolerance to drugs/alcohol, or contributed to criminal behavior. An identical measure of substance use, categorized in the same manner, was shown to predict early-onset juvenile preva-

Wolff et al. 13

lence offending trajectories (Baglivio, Wolff, Piquero, & Epps, 2015). Substance use as assessed by the C-PACT has also been demonstrated to predict recidivism for Flor- ida juvenile offenders generally, Florida serious, violent, and chronic (SVC) offenders (Baglivio, Jackowski, et al., 2014), and male Florida offenders (Baglivio, 2009).

Personal history risk factors School behavior. The current study includes two school problems: (a) age at first

suspension or expulsion, which ranged from 0 to 4 with higher values indicating a younger age at first suspension/expulsion (=0 if never expelled, 1 for expulsion between ages 16 and 18, 2 for ages 14 and 15, 3 for 10 through 13, and 4 for ages five-nine) and (b) number of suspensions/expulsions, which ranged from 1 to 6 with higher values indicating more suspensions/expulsions (=1 if never suspended, 2 for one suspension, 3 for two-three suspensions, 4 for four or five suspensions, 5 for six- seven suspensions, and 6 for more than seven suspensions/expulsions). These two items were standardized to create the school behavior scale (α = .834).

History of running away. The number of times a youth ran away or was kicked out of the house where the youth did not voluntarily return within 24 hours was included. The measure includes instances reported as well as not reported to law enforcement. The item ranges from no instances, 1, 2 to 3, 4 to 5, and more than five instances (coded 1-5, respectively). The grouping of instances is based on that captured by the C-PACT.

Age at first offense. The age of each youth at the time of their first referral for delinquency (equivalent to an adult arrest) was measured by the C-PACT. Categories include 12 and below, 13 to 14, 15, 16, and above 16 (coded 1-5, respectively, with higher values indicating an older age at first arrest). The grouping of categories is based on that captured by the C-PACT.

Worst prior offense. The most serious offense a youth has ever been adjudicated for was included and covers misdemeanor, felony other, property felony, and violent fel- ony categories. Each category was made dichotomous (such as violent felony = 1) and entered in the models presented below, with misdemeanor being the reference group.

History of residential placement. This measure is dichotomous for whether the youth in the current study had a history of residential commitment placement with the FDJJ. Forty percent of youth included in the current sample met this criterion. In Florida, a youth is committed to a residential program only by a judge, for an indeterminate length of time, and must complete an individualized treatment plan.

Service type. The type of community-based service the juvenile completed pre- ceding the 365-day follow-up was included as a control variable.5 Service type was grouped into five types: diversion services, probation supervision, day treatment, redi- rections, and aftercare services. Each type was made dichotomous (e.g., probation = 1) and diversion services served as the reference group. Diversion services are non-

14 International Journal of Offender Therapy and Comparative Criminology

judicial alternatives used to keep less serious youth offenders from being processed. Diversion services range from teen courts, mediation services, and mentoring pro- grams, to services for family problems and substance use issues.

Probation supervision is ordered by the court when a youth has been determined to have committed a delinquent act. All youth placed on probation have a juvenile proba- tion officer assigned who monitors compliance with court restrictions and sanctions, conducts C-PACT risk/needs assessments that drive an individualized case plan, and makes appropriate referrals for services.

Day treatment programs are facility-based treatment programs that provide delin- quency interventions, and vocational and educational training in the afternoons and evenings and often on weekends. Youth report daily to the facility but do not stay overnight. Day treatment programs are all contracted providers who serve as the case manager and conduct C-PACT assessments and reassessments driving individualized case plans.

Redirection services were “established by the Florida Legislature to use commu- nity-based alternatives in lieu of residential commitment for those youth that meet certain criteria” (FDJJ, 2015a, p.3). Redirection services involved community-based intensive family therapy programs and during the time period used in the current study consisted predominately of Multi-Systemic Therapy (MST) and Functional Family Therapy (FFT), with some youth receiving Brief Strategic Family Therapy (BSFT) or Parenting With Love and Limits (PLL). All providers of these family therapies were trained according to protocols of the respective model developers and monitored for fidelity under one “managing entity” contract. Juvenile probation officers still oversee the case and conduct C-PACT assessments. The redirection service is essentially an “overlay” addition to probation.

Finally, for the purposes of this study, aftercare services were composed of post- commitment supervision. Aftercare youth are those youth released from residential placement and subsequently supervised in the community under probation supervi- sion. Therefore, all aftercare services youth have a history of residential placement, whereas not all youth in any other service type have such history (though some may, just not immediately preceding the service type placement examined in the current study). Like probation youth, aftercare youth have an assigned juvenile probation officer and an individualized case plan based on C-PACT assessment results.6

Table 1 presents the descriptive statistics for the variables included in our analysis. The mean number of ACEs in the sample is 2.64. Seventy-seven percent of the sample is male, with an average age of 16 at the time of their completion of FDJJ services. Blacks make up the largest proportion of the sample (46%), followed by Whites (38%), Hispanics (15%), and other races (a total of 0.5%). Nearly 80% of the youth had previ- ously been adjudicated for a felony offense, with 47% having committed a violent felony. The largest proportion of the sample was on probation (37%), while most oth- ers were receiving aftercare or participating in a diversion program (26.1% and 22.5%, respectively). Just more than 40% of youth were re-arrested within the 1-year follow- up period, with the average number of days to the first subsequent arrest at 277.

Wolff et al. 15

Table 1. Descriptive Statistics for Analysis of ACEs and Juvenile Recidivism (N = 27,867).

Variable Variable definition M SD

Dependent variables Time to failure Number of days between completion and re-arrest 277.02 124.43 Re-arrest Failure dummy, re-arrest within 365 days of completion

of services 0.406 0.491

Independent variables Sum of ACE score Total of ACEs for each youth, range = 0-10 2.642 1.776 Demographics Age Age at time of release from DJJ service 16.308 1.635 Gender Coded 0 for females, 1 for males 0.77 0.421 Black Coded 1 for Black youth, 0 for all others 0.463 0.499 Hispanic Coded 1 for Hispanic youth, 0 for all others 0.152 0.359 Other race Coded 1 for Asian, Native American, or Pacific Islander,

0 for all others 0.005 0.07

Individual risk factors Substance abuse Coded 0 for no past use, 1 for past use, and 2 for past

use that caused problems 0.892 0.748

Antisocial peers Coded 1 if youth reported having exclusively antisocial peers or gang members, 0 for all others

0.05 0.218

Personal history risk factors School behavior Two-item standardized index, including age at first

suspension or expulsion and number of suspensions and expulsions (α = .834)

0 1.852

History of running away

Coded 0 for no instances, 1 for one instance, 2 for two to three instances, 3 for four to five instances, and 4 for more than five instances

1.707 1.217

Age at first offense Coded 5 for above 16 years old, 4 for 16, 3 for 15, 2 for 13 to 14, and 1 for 12 and below

2.191 1.119

Residential placement history

Coded 1 for no prior residential treatment, 2 for one residential stay, and 3 for two or more

1.368 0.591

Prior misdemeanor Worst prior offense was a misdemeanor = 1, else = 0 0.207 0.405 Prior other felony Worst prior offense was a felony = 1, else = 0 0.041 0.198 Prior property

felony Worst prior offense was a property felony = 1, else = 0 0.281 0.45

Prior violent felony Worst prior offense was a violent felony = 1, else = 0 0.47 0.499 Diversion Service type = diversion, coded 1, else = 0 0.225 0.418 Probation Service type = probation, coded 1, else = 0 0.373 0.484 Redirection Service type = redirection, coded 1, else = 0 0.047 0.212 Day treatment Service type = day treatment, coded 1, else = 0 0.093 0.291 Aftercare Service type = aftercare, coded 1, else = 0 0.261 0.439

Note. ACEs = Adverse childhood experiences; DJJ = Department of Juvenile Justice.

Analytic Strategy

To assess the impact of ACEs and other covariates on juvenile recidivism, we use a series of Cox proportional hazard models. This form of survival analysis is used to examine the extent to which traumatic experiences are related to recidivism, in

16 International Journal of Offender Therapy and Comparative Criminology

our case time to failure. Proportional hazard models are ideal for analyzing time- dependent censored outcomes, such as days to re-arrest within a follow-up period and the outcome of our research. Furthermore, survival analysis accommodates censored data, in which we do not observe the outcome of interest (re-arrest) due to the ending of the follow-up period. The Cox method is also a special form of the generalized proportional hazard model that is less restrictive and allows the baseline hazard ratio (HR) to take any form, rather than being specified a priori (Cox, 1972; Hosmer et al., 2008). The model is written as follows:

ln ACE DEMO RISK,1 2 3hi t t( )( ) = ( ) + + +α β β β

where hi(t) is the risk of recidivism on day t for youth i, α(t) is the baseline hazard or intercept, ACE is the measure of childhood trauma, DEMO is a matrix of demo- graphic covariates (age, gender, race), and RISK is a matrix of the individual-level risk factors described above. The outcome h(t) is defined as the probability of recidivism on day t knowing the actual failure day T, dropping those who have already failed, and adjusting for the passage of time during the 1-year follow-up window Δt.

Results

Modeling ACE Effects on Time to Recidivism

The main objective of our analysis is to assess the effect of ACE on juvenile recidivism, controlling for demographic characteristics and other relevant risk factors for offend- ing. Cox proportional hazard regressions are used to model the days to failure using a broad measure of recidivism (re-arrest). A HR greater than 1 indicates a shorter time to failure, that youth with that particular characteristic or risk factor recidivated more quickly than offenders in the comparison group, controlling for other variables in the model (Chung, Schmidt, & Witte, 1991). Several iterative regression models were con- structed to evaluate the effect of a summative measure of ACE on time to failure. First, the main effect of ACEs on time to failure was assessed. The second model includes demographic characteristics to assess any group differences which may exist in the relationship between ACE and the timing of re-arrest. Finally, individual-level risk fac- tors were introduced into the regression model to uncover any confounding relation- ships between ACE and the risk of failure. The results of this analysis for the full sample, as well as male and female youth separately, are presented in Table 2.

Most central to the research question addressed in the current study, the results shown in Table 2 reveal that youth who reported a greater number of ACEs were sig- nificantly more likely to be re-arrested earlier in the follow-up period. As indicated by the HRs presented in both the full and gender-specific models, an additional ACE was associated with a shorter time to failure, net of the commonly considered demographic and risk factors. Although the observed effect is reduced once other risk factors are included, the effect remains significant across all models presented in Table 2, includ- ing the models estimated on the full sample as well as gender-specific subsamples.

Wolff et al. 17

Table 2. Multivariate Analysis of ACEs and Juvenile Recidivism.

Full sample (N = 27,867) Males (n = 21,462) Females (n = 6,405)

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Independent variables ACE 1.037** 1.078** 1.022** 1.059** 1.073** 1.019** 1.096** 1.101** 1.035**

(0.005) (0.006) (0.006) (0.006) (0.006) (0.007) (0.013) (0.013) (0.014) Demographics Age — 1.056** 1.011 — 1.069** 1.020* — 0.975 0.953*

— (0.006) (0.008) — (0.007) (0.008) — (0.015) (0.019) Gender — 1.946** 1.732** — — — — — —

— (0.051) (0.048) — — — — — — Black — 1.461** 1.390** — 1.513** 1.445** — 1.215** 1.088

— (0.031) (0.031) — (0.035) (0.035) — (0.062) (0.062) Hispanic — 1.154** 1.177** — 1.197** 1.224** — 0.910 0.903

— (0.034) (0.035) — (0.038) (0.039) — (0.077) (0.077) Other race — 0.927 0.994 — 0.996 1.098 — 0.696 0.694

— (0.144) (0.154) — (0.172) (0.189) — (0.247) (0.247) Individual risk factors Substance abuse — — 1.157** — — 1.168** — — 1.054

— — (0.017) — — (0.018) — — (0.040) Antisocial peers — — 1.045 — — 1.006 — — 1.292*

— — (0.042) — — (0.045) — — (0.130) Personal history risk factors School behavior — — 1.064** — — 1.066** — — 1.056**

— — (0.006) — — (0.007) — — (0.015) History of running away — — 1.039** — — 1.035** — — 1.065**

— — (0.009) — — (0.010) — — (0.018) Age at first offense — — 0.894** — — 0.897** — — 0.876**

— — (0.010) — — (0.011) — — (0.026) Residential placement history — — 1.107** — — 1.096** — — 1.182*

— — (0.028) — — (0.029) — — (0.085) Prior other felony — — 1.082 — — 1.073 — — 1.077

— — (0.060) — — (0.066) — — (0.136) Prior property felony — — 1.148** — — 1.125** — — 1.199**

— — (0.036) — — (0.040) — — (0.084) Prior violent felony — — 1.098** — — 1.059 — — 1.261**

— — (0.034) — — (0.038) — — (0.077) Probation — — 1.263** — — 1.289** — — 1.191*

— — (0.042) — — (0.048) — — (0.087) Redirection — — 1.704** — — 1.797** — — 1.402**

— — (0.084) — — (0.100) — — (0.155) Day treatment — — 1.642** — — 1.687** — — 1.500**

— — (0.066) — — (0.076) — — (0.139) Aftercare — — 1.442** — — 1.501** — — 1.154

— — (0.064) — — (0.073) — — (0.133)

Note. Hazard ratios reported with standard errors in parentheses. ACEs = Adverse childhood experiences. *p < .05. **p < .01.

The timing and likelihood of re-arrest was also influenced by the individual charac- teristics of the youth. Specifically, demographic characteristics as well as indicators of criminal history and personal risk factors were significantly related to a shorter time to

18 International Journal of Offender Therapy and Comparative Criminology

failure. Black and Hispanic youth were more likely to fail earlier in comparison with Whites. Youth who reported past substance abuse failed more quickly and were re- arrested within the follow-up period, though antisocial peers were not a significant predictor in the full model.

Turning to the gender-specific models, substance abuse was not significant for females, although it was for male youth. The impact of antisocial peers was significant for females but not for males. Similarly, those who first offended earlier in life were more likely to be re-arrested earlier. School misconduct, a history of running away, and previous residential placement were also associated with earlier failure within both the full and gender-specific samples. Although youth who committed a violent or property felony offense were more likely to be re-arrested earlier within the follow-up period when compared with misdemeanants, the effect was the largest for those youth who committed a property offense (with violent felony failing to achieve significance in the male model). This finding is consistent with Wright and Rodriguez (2014), who found that youth who had been previously convicted on a property offense were at highest risk for future involvement in criminal activity and is in line with prior research indicating that the severity of a presenting offense has no, or even an inverse, significant associa- tion with re-offending (Baglivio, Greenwald and Russell,2015; Grattet, Lin, & Petersilia, 2011; Langan & Levin, 2002; Piquero, Jennings, & Barnes, 2012). In comparison with the diversion youth, those youth who were placed in any of the more intensive forms of supervision were also more likely to be re-arrested more quickly.7 These results also held across the three samples considered in Table 2 (full, males only, and females only).

Race/Ethnicity and Gender-Specific Models

Next, to examine the race/ethnic-specific effects of ACE on time to failure, the sample was segmented into White, Black, and Hispanic youth. An identical set of models was run for each racial/ethnic group (see Table 3). Results indicate that ACE is a significant predictor of early failure for both White and Black youth, although the effect for Hispanic youth becomes insignificant once the individual and personal risk factors are introduced beyond demographics (see Model 9; Table 3). These results suggest that ACE represents a salient predictor of recidivism and earlier failure, one which is rel- evant for both Black and White youth, net of other commonly considered risk factors. Across race/ethnicity, gender, substance abuse, school behavior, running away, age at first offense, and all service types deeper than diversion were significant in the hypoth- esized directions. Of note, antisocial peer association, as operationalized currently, was not significant for any race/ethnic group.

Finally, to assess the effect of ACEs on re-offending for each of these six sub- groups, we also estimated the same set of survival analysis models for Gender × Race/ Ethnic Groups (see Table 4).8 Although the HRs remain in the predicted direction, our findings indicate that ACEs are not a significant predictor of time to failure among White girls (HR = 1.019, p > .05), Black boys (HR = 1.016, p > .05), or Hispanic boys (HR = 1.004, p > .05) and girls (HR = 1.085, p > .05). Higher ACE scores remain a significant predictor of earlier failure for White males (HR = 1.025, p < .05) and Black

Wolff et al. 19

Table 3. Multivariate Analysis of ACEs and Juvenile Recidivism by Race/Ethnicity.

White youth (n = 10,624) Black youth (n = 12,899) Hispanic youth (n = 4,224)

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Independent variables ACE 1.061** 1.085** 1.024* 1.032** 1.069** 1.020* 1.048** 1.088** 1.015

(0.009) (0.009) (0.010) (0.008) (0.009) (0.009) (0.015) (0.016) (0.017) Demographics Age — 1.050** 1.014 — 1.050** 0.998 — 1.097** 1.066**

— (0.011) (0.014) — (0.008) (0.010) — (0.018) (0.023) Gender — 1.675** 1.466** — 2.080** 1.874** — 2.234** 1.998**

— (0.073) (0.068) — (0.076) (0.072) — (0.180) (0.168) Individual risk factors Substance

abuse — — 1.183** — — 1.131** — — 1.172** — — (0.030) — — (0.022) — — (0.044)

Antisocial peers

— — 1.145 — — 0.998 — — 1.040 — — (0.083) — — (0.055) — — (0.113)

Personal history risk factors School

behavior — — 1.080** — — 1.058** — — 1.050** — — (0.011) — — (0.008) — — (0.016)

History of running away

— — 1.037** — — 1.033** — — 1.053* — — (0.014) — — (0.013) — — (0.024)

Age at first offense

— — 0.880** — — 0.912** — — 0.862** — — (0.016) — — (0.014) — — (0.024)

Residential placement history

— — 1.102* — — 1.140** — — 0.983 — — (0.047) — — (0.038) — — (0.083)

Prior other felony

— — 1.138 — — 1.049 — — 0.980 — — (0.100) — — (0.089) — — (0.130)

Prior property felony

— — 1.221** — — 1.106* — — 1.084 — — (0.061) — — (0.052) — — (0.085)

Prior violent felony

— — 1.185** — — 1.064 — — 0.972 — — (0.061) — — (0.047) — — (0.076)

Probation — — 1.212** — — 1.290** — — 1.305** — — (0.065) — — (0.063) — — (0.107)

Redirection — — 1.500** — — 1.846** — — 1.812** — — (0.124) — — (0.134) — — (0.216)

Day treatment — — 1.585** — — 1.685** — — 1.689** — — (0.111) — — (0.096) — — (0.179)

Aftercare — — 1.353** — — 1.438** — — 1.755** — — (0.099) — — (0.091) — — (0.222)

Note. Hazard ratios reported with standard errors in parentheses. ACEs = Adverse childhood experiences. *p < .05. **p < .01.

females (H.R. = 1.044, p < .05). Collectively, these findings reveal that the extent to which ACE has an impact on juvenile recidivism may vary for particular subgroups. These findings are discussed in greater detail in the following section.

20

T a b

le 4

. M

ul ti va

ri at

e A

na ly

si s

o f A

C Es

a nd

J uv

en ile

R ec

id iv

is m

b y

G en

de r

an d

R ac

e/ Et

hn ic

it y.

W hi

te f em

al es

(n

= 2

,5 36

) W

hi te

m al

es (

n =

8 ,0

88 )

B la

ck f em

al es

(n

= 3

,0 57

) B

la ck

m al

es (

n =

9 ,8

42 )

H is

pa ni

c fe

m al

es

(n =

7 76

) H

is pa

ni c

m al

es

(n =

3 ,4

48 )

(1

) (2

) (3

) (4

) (5

) (6

) (7

) (8

) (9

) (1

0) (1

1) (1

2) (1

3) (1

4) (1

5) (1

6) (1

7) (1

8)

In de

pe nd

en t

va ri

ab le

s

A C

E 1.

07 7*

* 1.

07 7*

* 1.

01 9

1. 08

5* *

1. 08

8* *

1. 02

5* 1.

10 1*

* 1.

10 3*

* 1.

04 4*

1. 05

9* *

1. 06

1* *

1. 01

6 1.

19 7*

* 1.

19 7*

* 1.

08 5

1. 06

4* *

1. 07

1* *

1. 00

4 (0

.0 20

) (0

.0 20

) (0

.0 21

) (0

.0 10

) (0

.0 10

) (0

.0 11

) (0

.0 19

) (0

.0 19

) (0

.0 20

) (0

.0 10

) (0

.0 10

) (0

.0 10

) (0

.0 45

) (0

.0 45

) (0

.0 50

) (0

.0 17

) (0

.0 17

) (0

.0 18

) D

em o gr

ap hi

cs

A ge

— 0.

97 4

0. 94

9 —

1. 06

5* *

1. 02

5 —

0. 96

7 0.

93 3*

* —

1. 06

2* *

1. 00

8 —

1. 02

5 1.

07 2

— 1.

10 4*

* 1.

06 2*

* —

(0 .0

25 )

(0 .0

33 )

— (0

.0 12

) (0

.0 15

) —

(0 .0

20 )

(0 .0

24 )

— (0

.0 09

) (0

.0 11

) —

(0 .0

53 )

(0 .0

72 )

— (0

.0 20

) (0

.0 24

) In

di vi

du al

r is

k fa

ct o rs

Su

bs ta

nc e

ab us

e —

— 1.

14 7*

— —

1. 18

7* *

— —

0. 96

3 —

— 1.

15 6*

* —

— 1.

08 6

— —

1. 18

5* *

— —

(0 .0

70 )

— —

(0 .0

33 )

— —

(0 .0

52 )

— —

(0 .0

25 )

— —

(0 .1

33 )

— —

(0 .0

47 )

A

nt is

o ci

al

pe er

s —

— 1.

35 0

— —

1. 11

0 —

— 1.

30 9

— —

0. 96

7 —

— 1.

14 6

— —

1. 02

8 —

— (0

.2 07

) —

— (0

.0 91

) —

— (0

.1 95

) —

— (0

.0 57

) —

— (0

.3 54

) —

— (0

.1 20

) Pe

rs o na

l h is

to ry

r is

k fa

ct o rs

Sc

ho o l

be ha

vi o r

— —

1. 05

4* —

— 1.

08 5*

* —

— 1.

04 4*

— —

1. 06

1* *

— —

1. 13

7* *

— —

1. 04

3* *

— —

(0 .0

24 )

— —

(0 .0

12 )

— —

(0 .0

21 )

— —

(0 .0

09 )

— —

(0 .0

56 )

— —

(0 .0

17 )

H

is to

ry o

f ru

nn in

g aw

ay

— —

1. 04

4 —

— 1.

04 3*

* —

— 1.

08 2*

* —

— 1.

01 9

— —

1. 09

2 —

— 1.

04 4

— —

(0 .0

29 )

— —

(0 .0

16 )

— —

(0 .0

26 )

— —

(0 .0

15 )

— —

(0 .0

61 )

— —

(0 .0

27 )

A

ge a

t fir

st

o ffe

ns e

— —

0. 88

2* *

— —

0. 88

0* *

— —

0. 88

2* *

— —

0. 91

7* *

— —

0. 81

9* —

— 0.

87 1*

* —

— (0

.0 41

) —

— (0

.0 18

) —

— (0

.0 37

) —

— (0

.0 15

) —

— (0

.0 73

) —

— (0

.0 26

)

( co

nt in

ue d)

21

W hi

te f em

al es

(n

= 2

,5 36

) W

hi te

m al

es (

n =

8 ,0

88 )

B la

ck f em

al es

(n

= 3

,0 57

) B

la ck

m al

es (

n =

9 ,8

42 )

H is

pa ni

c fe

m al

es

(n =

7 76

) H

is pa

ni c

m al

es

(n =

3 ,4

48 )

(1

) (2

) (3

) (4

) (5

) (6

) (7

) (8

) (9

) (1

0) (1

1) (1

2) (1

3) (1

4) (1

5) (1

6) (1

7) (1

8)

R

es id

en tia

l pl

ac em

en t

hi st

o ry

— —

1. 28

3 —

— 1.

12 4

— —

0. 85

1 —

— 1.

03 5

— —

0. 65

3 —

— 1.

04 1

— —

(0 .2

01 )

— —

(0 .1

19 )

— —

(0 .2

22 )

— —

(0 .0

94 )

— —

(0 .3

08 )

— —

(0 .1

46 )

Pr

io r

o th

er

fe lo

ny —

— 1.

23 9*

— —

1. 20

9* *

— —

1. 19

1 —

— 1.

05 5

— —

1. 21

3 —

— 1.

09 8

— —

(0 .1

33 )

— —

(0 .0

70 )

— —

(0 .1

28 )

— —

(0 .0

57 )

— —

(0 .2

38 )

— —

(0 .0

95 )

Pr

io r

pr o pe

rt y

fe lo

ny

— —

1. 41

7* *

— —

1. 14

4* —

— 1.

27 3*

* —

— 0.

99 8

— —

0. 72

5 —

— 1.

01 5

— —

(0 .1

45 )

— —

(0 .0

68 )

— —

(0 .1

08 )

— —

(0 .0

52 )

— —

(0 .1

44 )

— —

(0 .0

88 )

Pr

io r

vi o le

nt

fe lo

ny

— —

1. 07

6 —

— 1.

10 6*

— —

1. 32

2* *

— —

1. 12

4* *

— —

1. 16

7 —

— 0.

97 3

— —

(0 .1

27 )

— —

(0 .0

51 )

— —

(0 .1

28 )

— —

(0 .0

40 )

— —

(0 .3

20 )

— —

(0 .0

87 )

Pr

o ba

ti o n

— —

1. 04

1 —

— 1.

28 4*

* —

— 1.

37 1*

* —

— 1.

27 0*

* —

— 1.

09 8

— —

1. 33

3* *

— —

(0 .1

16 )

— —

(0 .0

79 )

— —

(0 .1

50 )

— —

(0 .0

70 )

— —

(0 .2

51 )

— —

(0 .1

18 )

R

ed ir

ec tio

n —

— 1.

00 4

— —

1. 68

4* *

— —

1. 78

1* *

— —

1. 87

8* *

— —

1. 53

4 —

— 1.

85 4*

* —

— (0

.1 91

) —

— (0

.1 56

) —

— (0

.2 77

) —

— (0

.1 55

) —

— (0

.4 92

) —

— (0

.2 39

)

D ay

tr

ea tm

en t

— —

1. 13

8 —

— 1.

73 4*

* —

— 1.

83 1*

* —

— 1.

65 2*

* —

— 1.

70 7

— —

1. 69

5* *

— —

(0 .1

93 )

— —

(0 .1

36 )

— —

(0 .2

31 )

— —

(0 .1

06 )

— —

(0 .4

87 )

— —

(0 .1

94 )

A

ft er

ca re

— —

1. 07

5 —

— 1.

45 4*

* —

— 1.

24 4

— —

1. 44

4* *

— —

1. 00

3 —

— 1.

83 0*

* —

— (0

.1 93

) —

— (0

.1 19

) —

— (0

.2 06

) —

— (0

.1 00

) —

— (0

.4 09

) —

— (0

.2 47

)

N ot

e. H

az ar

d ra

ti o s

re po

rt ed

w it h

st an

da rd

e rr

o rs

in p

ar en

th es

es . A

C Es

= A

dv er

se c

hi ld

ho o d

ex pe

ri en

ce s.

* p <

.0 5.

* *p

< .0

1.

T a b

le 4

. (c

o n

ti n

u e d

)

22 International Journal of Offender Therapy and Comparative Criminology

Discussion

Juvenile offenders represent an especially important group of theoretical and policy interest (Mulvey et al., 2004). Also, assessing the extent to which particular risk fac- tors are associated with continued or curtailed re-offending represents an important line of inquiry that can be used not only to better develop theoretical models but also to importantly assist policy makers in designing evidence-based programs aimed at helping youth move away from a continued criminal career. One of these risk factors, ACEs, was the focus of the current investigation. Specifically, our analysis of the 1-year recidivism patterns among a large sample of previously adjudicated delinquent youth who participated in community-based services in the State of Florida was car- ried out in an effort to contribute in several ways to this growing body of research, especially with respect to the lack of research assessing demographic differences. Several key findings emerged from our analysis.

First, our results show that when included independently or with demographic con- trols, a higher ACE shortened time to failure for all gender and race/ethnic subgroups. Even when individual and personal history risk factors were included in our gender- specific models, higher ACE scores predicted shorter time to failure for both males and females. Considering race/ethnicity, full models show that higher ACE scores increase hazard rate for Whites and Blacks but were not predictive of time to failure of Hispanics. These results provide evidence of a relatively robust relationship between ACEs and recidivism.9

Our findings point to several important areas of future scholarship. First, as our study only examined the cumulative ACE score effects on re-arrest, future efforts should examine whether some type of ACE threshold exists, although we recognize that a dose–response relationship may not exist across all gender and race/ethnic sub- groups. The CDC (2015) considered ACE scores of 4 or higher as the extreme tail of ACE distribution. Perhaps juveniles with at least 4 ACE exposures differ from those with less than 4, or perhaps some other threshold exists. Additional research should consider thresholds and further explore whether thresholds differ across gender and race/ethnic groups.

Second, while the effect of ACE scores on recidivism was robust for White males and Black females once all additional risk factors were included, subsequent analysis should examine resiliency differences across gender and race/ethnic subgroups. Although risk factors across domains were examined, investigations into protective factors may provide additional insight into gender and race/ethnic differences. On this score, researchers have argued for examining the existence of buffering protective fac- tors, which predict a low probability of a negative outcome (such as re-arrest) in the presence of risk factors, such that the buffering factor attenuates the impact of a risk (Lösel & Farrington, 2012). Additional work should seek to uncover which protective factors, such as support network for the family, low impulsivity, or optimism for the future, may mitigate the effects of ACEs on re-arrest.

Third, recall that the results of our analysis suggest that ACEs played much less of a role in time to re-arrest among Hispanic youth in our sample. For both the full sample

Wolff et al. 23

of Hispanic youth and the gender-specific subsamples, ACEs were not a significant predictor of continued criminal involvement. Supplementary analyses showed that, on average, Hispanic youth have experienced fewer traumas when compared with their peers of different racial and ethnic backgrounds (2.21 for Hispanics compared with 2.53 for Blacks and 2.94 for Whites; all differences statistically significant). The lower rate of ACEs among Hispanics may explain the insignificant effects if, as suggested above, a threshold exists. It is also possible that Hispanic youth in the State of Florida are better able to cope with these stressors and traumas in a way that does not translate into crime and violence, that is, they are more resilient (Holleran & Jung, 2005). Accordingly, future research should explore the potential for both individual and cul- tural factors to moderate the relationship between ACEs and the negative life events often being connected to them.

Finally, while we were able to link ACEs to recidivism, there remains a need to unpack the theoretical mechanisms by which ACEs increase re-offending and a more accelerated time to re-offending. Fleshing out these mechanisms requires extensive data, much of which is not collected by—or contained in—juvenile justice data infor- mation systems. Nevertheless, focusing on both internalization and externalization of the experienced stress and then how a youth deals with that stress is an important ques- tion, one that future empirical work should explore.

Aside from this healthy research agenda, it is important to be cautious of our conclu- sions because of some data limitations. For example, as is the case with any study using official records, there exists the possibility of underreporting with respect to offending. The use of self-reported offending data should be adopted in subsequent work.10

Second, recall that we examined the characteristics of youth while including a measure of the service type which the youth completed (probation, day treatment, etc.). Unfortunately, the data did not contain information for the services or interven- tions received while participating in each service type (e.g., Mulvey, Schubert, & Chung, 2007). Furthermore, no measure of dosage (intensity or duration received) of any of those interventions received during placement was available. Future research should strive to collect the appropriate data to examine these issues.11 Finally, the current study included only youth assessed with the Full C-PACT, which was neces- sary to create ACE scores. Youth assessed with the Pre-Screen C-PACT were not included. Arguably, using only the Full C-PACT results in a higher risk sample, yet limits generalizability to all juvenile offenders in Florida. While our sample repre- sented 21% of all releases during the 3-year study period, it should be noted that 39% of the final sample was assessed as low risk and 16% as moderate risk on the vali- dated C-PACT. In addition, future research should examine the effects of ACEs on re-offending outside of Florida to assess generalizability of the current findings. Florida data are unique in that not only are both juvenile and adult recidivism avail- able (necessary as some youth were or turned 18 during the follow-up) but also risk assessment data enabling creation of ACE scores as well as prominent risk factors are captured for all higher risk youth. Although Florida is a large, diverse state, the diver- sity of which is evident in the current sample, replications would enhance the merit of ACEs as predictors of juvenile recidivism.

24 International Journal of Offender Therapy and Comparative Criminology

In closing, we observe that a great deal of research is documenting that higher ACE scores are related to a host of negative short- and long-term outcomes across domains of health (Anda et al., 2010), neurological development (Danese & McEwen, 2012; Shalev et al., 2013), achievement (Bellis et al., 2014), and recently offending (Bellis et al., 2014; Fox, Perez, Cass, Baglivio, & Epps, 2015). There is also evidence that exposure to traumatic childhood events among special populations, such as juvenile offenders, is more prevalent than that found in nationally representative samples (Baglivio et al., 2014). Policies for universal screening for ACEs may serve as one important approach to guiding treatment provision. The simplicity of the ACE score strengthens the policy implications as there are limited costs to implementation of a 10-item additive yes/no binary screening tool. Additional work must be undertaken regarding the effectiveness of various treatment models and interventions at attenuat- ing the impact of ACEs on short- and long-term negative outcomes. Juvenile justice systems should keep in mind that, although they deal with adjudicated youthful offend- ers, a good portion of these youth arrive with a wide range of circumstances—several of which include experiencing traumatic and adverse events. Prevention and interven- tion efforts should ensure that this knowledge is not only obtained but also used in matching youth with appropriate treatment if criminal careers are to be rerouted away from persistence and transition into early adulthood, especially adult antisocial behav- ior (see Loeber & Farrington, 2012).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

1. We acknowledge that systems are on a wide continuum with respect to their use of grad- uated sanctions, and whether those graduations are afforded to all youth, regardless of offense. Our main view is that the vast majority of juveniles are afforded community- based sanctions upon adjudication of an offense rather than are placed in deeper-end, more restrictive settings.

2. It is important to note that the current study is not an evaluation of community-based ser- vices per se, but instead is focused on the relationship between adverse childhood experi- ences (ACEs) and recidivism.

3. Of those not receiving a Full Community Positive Achievement Change Tool (C-PACT), 65% were diversion youth, 30% probation supervision youth, the remaining 5% split between day treatment, redirections, and aftercare youth. Youth who do not receive a Full C-PACT assessment are all assessed as low or moderate risk on the C-PACT Pre-Screen. Accordingly, these youth do not require a Full C-PACT as per Florida Department of Juvenile Justice (FDJJ) policy (though obviously because our sample was 55% low and

Wolff et al. 25

moderate risk, these youth may receive a Full C-PACT). All youth being placed into day treatment and redirection services are supposed to receive a Full C-PACT (hence, the low percentage of those youth who did not).

4. In response to an anonymous reviewer’s comments, an alternative operationalization of the antisocial peer variable was explored. For these ancillary analyses (results available upon request), we created an ordinal measure of antisocial peers coded: 1 if youth reported hav- ing exclusively pro-social friends, 2 if youth had pro-social and antisocial friends, and 3 if youth reported having exclusively antisocial or gang member peers. Results pertaining to our focal variable (ACE) are substantively identical to those reported in the main “Results” section; however, the effect of antisocial peers was a significant predictor of continued justice system involvement in this alternative specification.

5. We reiterate that assessing the effectiveness of different service types is not the focus of the current study. We include service type only as a control variable for the purposes of removing confounding of the effects of what placement the juvenile completed from those of the ACE score on re-offending.

6. With a large number of risk factors being taken into consideration, there may be a concern of multicollinearity. In analyses not shown, we assessed the degree of collinearity present between the measures included in the current study. Results suggest that multicollinearity is not an issue in the current analysis as all correlations were less than r = .50.

7. Although not of central focus to the current work, the results presented in Table 2 sug- gest that youth who received any services other than diversion were more likely to be re-arrested earlier on during the follow-up period. Interestingly, those youth assigned to redirection or community-based intensive family therapy programs had, in most cases, the greatest odds of earlier re-arrest. These results were striking, given the vast literature on the efficacy of these approaches and led us to examine composition of those youth who received these services. In general, youth who received these community-based therapy interventions exhibited higher risk on many of the individual-level risk factors considered. For example, when compared with the other treatment groups, a higher proportion of redi- rection youth had committed violent felonies, offended earlier in life, had more problems in school, and reported having more delinquent peers. At the same time, they were less likely to have run away or been kicked out of their home and were not more likely to have abused drugs or alcohol. Overall, the large effect of redirection is likely a function of the type of youth who received that service, rather than the effect of the therapy therein.

8. The small number of cases among youth classified as “other” precluded their inclusion in the race/ethnicity and gender-specific models. Full results of the models referenced here are available from the authors upon request.

9. At the same time, once we further disaggregated the sample into gender and race/ethnic groups, higher ACE scores with demographic controls were significant across all gender and race/eth- nic groups (such as White females). However, upon inclusion of the individual and personal history risks, ACEs were significant predictors of White male and Black female re-arrest only.

10. Prior work has indicated that official measures show shorter offending careers, later age of onset, and a later age of desistance (Farrington, Ttofi, Crago, & Coid, 2014). However, in an analysis of longitudinal data from the Cambridge Study in Delinquent Development, it has been found that the probability of a self-reported offense leading to conviction was highest at ages 15 to 18, while the probability of a convicted offense being self-reported decreases with age from a 10- to 14-year-old group to 42- to 47-year-olds (Farrington, Piquero, & Jennings, 2013). Based on these findings, the current study and its use of offi- cial measure are arguably comparable with other studies based on self-reports.

26 International Journal of Offender Therapy and Comparative Criminology

11. Nevertheless, we did use the C-PACT risk/needs assessment at exit from placement to predict recidivism/time to failure, thereby assessing the risk/needs profile of the youth after any changes based on programming. Had we examined offenses committed during service or started the follow-up period from the date of service admission, then interven- tion information would have been paramount. In fact, we argue that any study examining service type effectiveness from the date of admission is essentially measuring the effective- ness of the actual services received and necessitates the need for dosage, as well as fidelity measures. Recently developed tools for this purpose include the Risk–Need–Responsivity (RNR) Simulation Tool (Taxman, Caudy, & Pattavina, 2013) and Lipsey’s Standardized Program Evaluation Protocol (SPEP; Howell & Lipsey, 2012; Lipsey & Howell, 2012; Lipsey, Howell, Kelly, Chapman, & Carver, 2010).

References

Agnew, R. (1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-88.

Agnew, R. (2001). Building on the foundation of general strain theory: Specifying the types of strain most likely to lead to crime and delinquency. Journal of Research in Crime & Delinquency, 38, 319-361.

Akers, R. L. (1998). Social structure and social learning. Los Angeles, CA: Roxbury. Anda, R. F., & Brown, D. W. (2010). Adverse childhood experiences & population health in

Washington: The face of a chronic public health disaster. Olympia, WA: Family Policy Council. Anda, R. F., Butchart, A., Felitti, V. J., & Brown, D. W. (2010). Building a framework for

global surveillance of the public health implications of adverse childhood experiences. American Journal of Preventive Medicine, 39, 93-98.

Anda, R. F., Felitti, V. J., Bremner, J. D., Walker, J. D., Whitfield, C., Perry, B. D., . . . Giles, W. H. (2006). The enduring effects of abuse and related adverse experiences in childhood: A convergence of evidence from neurobiology and epidemiology. European Archives of Psychiatry and Clinical Neuroscience, 256, 174-186.

Andrews, D. A., & Bonta, J. (2003). The psychology of criminal conduct (3rd ed.). Cincinnati, OH: Anderson.

Andrews, D. A., Bonta, J., & Wormith, J. S. (2011). The risk-need-responsivity (RNR) model: Does adding the good lives model contribute to effective crime prevention? Criminal Justice and Behavior, 38, 735-755.

Baglivio, M. T. (2009). The assessment of risk to recidivate among a juvenile offending popula- tion. Journal of Criminal Justice, 37, 596-607.

Baglivio, M. T., & Epps, N. (2015). The interrelatedness of adverse childhood experiences among high-risk juvenile offenders. Youth Violence and Juvenile Justice. doi:10.1177/ 1541204014566286

Baglivio, M. T., Epps, N., Swartz, K., Huq, M. S., & Hardt, N. S. (2014). The prevalence of adverse childhood experiences (ACE) in the lives of juvenile offenders. Journal of Juvenile Justice, 3, 1-23.

Baglivio, M. T., Greenwald, M. A., & Russell, M. (2015). Assessing the implications of a struc- tured decision-making tool for recidivism in a statewide analysis: Disposition matrix for court recommendations made by juvenile probation officers. Criminology & Public Policy, 14, 5-49.

Baglivio, M. T., & Jackowski, K. (2013). Examining the validity of a juvenile offending risk assessment instrument across gender and race/ethnicity. Youth Violence and Juvenile Justice, 11, 26-43.

Wolff et al. 27

Baglivio, M. T., Jackowski, K., Greenwald, M. A., & Howell, J. C. (2014). Serious, violent, and chronic juvenile offenders. Criminology & Public Policy, 13, 83-116.

Baglivio, M. T., Wolff, K. T., Epps, N., & Nelson, R. (2015). Predicting adverse childhood experiences: The importance of neighborhood context in youth trauma among delinquent youth. Crime & Delinquency. doi:10.1177/0011128715570628

Baglivio, M. T., Wolff, K. T., Piquero, A. R., & Epps, N. (2015). The relationship between adverse childhood experiences (ACE) and juvenile offending trajectories in a juvenile offender sample. Journal of Criminal Justice, 43, 229-241.

Baird, C., Healy, T., Johnson, K., Bogie, A., Dankert, E. W., & Scharenbroch, C. (2013). A com- parison of risk assessment instruments in juvenile justice. Madison, WI: National Council on Crime & Delinquency.

Barrett, D. E., Katsiyannis, A., Zhang, D., & Zhang, D. (2014). Delinquency and recidivism: A multicohort, matched-control study of the role of early adverse experiences, mental health problems, and disabilities. Journal of Emotional and Behavioral Disorders, 22, 3-15. doi:1063426612470514

Bellis, M. A., Lowey, H., Leckenby, N., Hughes, K., & Harrison, D. (2014). Adverse childhood experiences: Retrospective study to determine their impact on adult health behaviors and health outcomes in a UK population. Journal of Public Health, 36, 81-91. doi:10.1093/ pubmed/fdt038

Benda, B. B. (2005). Gender differences in life-course theory of recidivism: A survival analysis. International Journal of Offender Therapy and Comparative Criminology, 49, 325-342.

Bronfenbrenner, U., Moen, P., & Garbarino, J. (1984). Child, family, and community. In R. Parke (Ed.), Review of child development research (pp. 283-328). Chicago, IL: University of Chicago Press.

Cauffman, E. (2012). Aligning justice system processing with developmental science. Criminology & Public Policy, 11, 751-758.

Cauffman, E., Feldman, S., Waterman, J., & Steiner, H. (1998). Posttraumatic stress disor- der among female juvenile offenders. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 1209-1216.

Centers for Disease Control and Prevention. (2015). Injury prevention and control: Adverse Childhood Experiences (ACE) Study. Retrieved from http://www.cdc.gov/violencepreven- tion/acestudy/

Chartier, M. J., Walker, J. R., & Naimark, B. (2010). Separate and cumulative effects of adverse childhood experiences in predicting adult health and heath care utilization. Child Abuse & Neglect, 34, 454-464.

Chen, W. Y., Propp, J., deLara, E., & Corvo, K. (2011). Child neglect and its association with subsequent juvenile drug and alcohol offense. Child & Adolescent Social Work Journal, 28, 273-290.

Chesney-Lind, M. (2001). What about the girls? Delinquency programming as if gender mat- tered. Corrections Today, 63, 38-45.

Chesney-Lind, M., & Sheldon, R. G. (1992). Girls, delinquency and juvenile justice. Pacific Grove, CA: Brooks Cole.

Chung, C., Schmidt, P., & Witte, A. D. (1991). Survival analysis: A survey. Journal of Quantitative Criminology, 7, 59-98.

Cicchetti, D., & Manly, J. T. (2001). Operationalizing child maltreatment: Developmental pro- cesses and outcomes. Developmental Psychopathology, 13, 755-757.

Cottle, C. C., Lee, R. J., & Heilbrun, K. (2001). The prediction of criminal recidivism in juve- niles: A meta-analysis. Criminal Justice and Behavior, 28, 367-394.

28 International Journal of Offender Therapy and Comparative Criminology

Cox, D. R. (1972). Regression models and life-tables. Journal of the Royal Statistical Society, 34, 187-220.

Cullen, F. T., Vose, B. A., Jonson, C. L., & Unnever, J. D. (2007). Public support for early inter- vention: Is child saving a “habit of the heart.” Victims and Offenders, 2, 109-124.

Danese, A., & McEwen, B. S. (2012). Adverse childhood experiences, allostasis, allostatic load, and age-related disease. Physiology & Behavior, 106, 29-39.

Dean, C. H., Brame, R., & Piquero, A. R. (1996). Criminal propensities, discrete groups of offenders, and persistence in crime. Criminology, 34, 547-574.

DeLisi, M., & Piquero, A. R. (2011). New frontiers in criminal careers research, 2001–2011: A state-of-the-art review. Journal of Criminal Justice, 39, 289–301.

Dierkhising, C. B., Ko, S. J., Woods-Jaeger, B., Briggs, E. C., Lee, R., & Pynoos, R. S. (2013). Trauma histories among justice-involved youth: Findings from the National Child Traumatic Stress Network. European Journal of Psychotraumatology, 4, 1-12.

Dilulio, J. J., & United States. Bureau of Justice Statistics. (1992). Rethinking the criminal justice system: Toward a new paradigm (pp. pp-1). US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

Dong, M., Anda, R. F., Felitti, V. J., Dube, S. R., Williamson, D. F., Thompson, T. J., . . . Giles, W. H. (2004). The interrelatedness of multiple forms of childhood abuse, neglect, and household dysfunction. Child Abuse & Neglect, 28, 771-784.

Dowden, C., & Brown, S. L. (2002). The role of substance abuse factors in predicting recidi- vism: A meta-analysis. Psychology, Crime, & Law, 8, 243-264.

Dube, S. R., Anda, R. F., Felitti, V. J., Croft, J. B., Edwards, V. J., & Giles, W. H. (2001). Growing up with parental alcohol abuse: Exposure to childhood abuse, neglect, and house- hold dysfunction. Child Abuse & Neglect, 12, 1627-1640.

Dube, S. R., Felitti, V. J., Dong, M., Giles, W. H., & Anda, R. F. (2003). The impact of adverse childhood experiences on health problems: Evidence from four birth cohorts dating back to 1900. Preventive Medicine, 37, 268-277.

Dube, S. R., Williamson, D. F., Thompson, T., Felitti, V. J., & Anda, R. F. (2004). Assessing the reliability of retrospective reports of adverse childhood experiences among adult HMO members attending a primary care clinic. Child Abuse & Neglect, 28, 729-737.

Duke, N. N., Pettingell, S. L., McMorris, B. J., & Borowsky, I. W. (2010). Adolescent vio- lence perpetration: Associations with multiple types of adverse childhood experiences. Pediatrics, 125, 778-786.

Edleson, J. L. (1999). Problems associated with children’s witnessing domestic violence. Available from www.vaw.umn.edu

Evans-Chase, M. (2014). Addressing trauma and psychosocial development in juvenile justice- involved youth: A synthesis of the developmental neuroscience, juvenile justice and trauma literature. Laws, 3, 744-758.

Farrington, D. P., Loeber, R., & Howell, J. C. (2012). Young adult offenders: The need for more effective legislative options and justice processing. Criminology & Public Policy, 11, 729-750.

Farrington, D. P., Piquero, A. R., & Jennings, W. G. (2013). Offending from childhood to late middle age: Recent results from the Cambridge study in delinquent development. New York, NY: Springer.

Farrington, D. P., Ttofi, M. M., Crago, R. V., & Coid, J. W. (2014). Prevalence, frequency, onset, desistance and criminal career duration in self-reports compared with official records. Criminal Behaviour and Mental Health, 24, 241-253.

Wolff et al. 29

Felitti, V.J., Anda, R.F., Nordenberg, D., Williamson, D.F., Spitz, A.M., Edwards, V. (1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study. American Journal of Preventive Medicine, 14, 245–258.

Florida Department of Juvenile Justice. (2011). 2010 comprehensive accountability report. Tallahassee, FL: Florida Department of Juvenile Justice.

Florida Department of Juvenile Justice. (2012). 2011 comprehensive accountability report. Tallahassee, FL: Florida Department of Juvenile Justice.

Florida Department of Juvenile Justice. (2013). 2012 comprehensive accountability report. Tallahassee, FL: Florida Department of Juvenile Justice.

Florida Department of Juvenile Justice. (2014). 2013 comprehensive accountability report. Tallahassee, FL: Florida Department of Juvenile Justice.

Florida Department of Juvenile Justice. (2015a). 2014 Comprehensive Accountability Report. Tallahassee, Fl: Florida Department of Juvenile Justice.

Florida Department of Juvenile Justice. (2015b). 2013-2014 delinquency profile. Tallahassee, FL. Retrieved from http://www.djj.state.fl.us/research/delinquency-data/delinquency-pro- file/delinquency-profile-dashboard

Ford, J. D., Chapman, J. F., Hawker, J., & Albert, D. (2007). Trauma among youth in the juve- nile justice system: Critical issues and new directions. National Center for Mental Health and Juvenile Justice. Retrieved from http://www.ncmhjj.com/trauma-among-youth-in-the- juvenile-justice-system-critical-issues-and-new-directions/

Fox, B. H., Perez, N., Cass, E., Baglivio, M. T., & Epps, N. (2015). Trauma changes every- thing: Examining the relationship between adverse childhood experiences and serious, vio- lent and chronic juvenile offenders. Child Abuse & Neglect, 46, 163-173. doi:10.1016/j. chiabu.2015.01.011

Gatti, U., Tremblay, R. E., & Vitaro, F. (2009). Iatrogenic effect of juvenile justice. Journal of Child Psychology & Psychiatry, 50, 991-998.

Gibson, C. L., & Krohn, M. D. (2012). Raising the age: Issues in emerging adulthood and adult court referral of youthful offenders. Criminology & Public Policy, 11, 759-768.

Gilbert, R., Widom, C. S., Browne, K., Fergusson, D., Webb, E., & Janson, S. (2009). Burden and consequences of child maltreatment in high-income countries. The Lancet, 33, 68-81.

Graham v. Florida,130 S. Ct. (2011). Grattet, R., Lin, J., & Petersilia, J. (2011). Supervision regimes, risk, and official reactions to

parolee deviance. Criminology, 49, 371-400. Hamburger, M. E., Leeb, R. T., & Swahn, M. H. (2008). Childhood maltreatment and early alco-

hol use among high risk adolescents. Journal of Studies on Alcohol and Drugs, 69, 291-295. Herrenkohl, T. I., Mason, W. A., Kosterman, R., Lengua, L. J., Hawkins, J. D., & Abbott, R. D.

(2004). Pathways from physical childhood abuse to partner violence in young adulthood. Violence and Victims, 19, 123-136.

Herrenkohl, T. I., Sousa, C., Tajima, E. A., Herrenkohl, R. C., & Moylan, C. A. (2008). Intersection of child abuse and children’s exposure to domestic violence. Trauma, Violence & Abuse, 9, 84-99.

Herrera, V. M., & McCloskey, L. A. (2001). Gender differences in the risk for delinquency among youth exposed to family violence. Child Abuse and Neglect, 25, 1037-1051.

Hertzig, M. (1983). Temperament and neurological status. In M. Rutter (Ed.), Developmental neuropsychiatry (pp. 164-180). New York, NY: Guilford Press.

30 International Journal of Offender Therapy and Comparative Criminology

Holleran, L. K., & Jung, S. (2005). Acculturative stress, violence, and resilience in the lives of Mexican-American youth. Stress, Trauma, and Crisis, 8, 107-130.

Hosmer, D. W., Lemeshow, S., & May, S. (2008). Model development. Applied Survival Analysis: Regression Modeling of Time-to-Event Data, Second Edition, 132-168. New York, NY: Wiley

Howell, J. C. (2012). Gangs in America’s Communities. Thousand Oaks, CA: Sage. Howell, J. C. (2009). Preventing and reducing juvenile delinquency: A comprehensive frame-

work. Thousand Oaks, CA: Sage. Howell, J. C., & Lipsey, M. W. (2012). Research guidelines for juvenile justice programs.

Justice Research and Policy, 14, 17-34. Howell, J. C., Lipsey, M. W., & Wilson, J. J. (2014). A handbook for evidence-based juvenile

justice systems. Lanham, MD: Lexington Books. Jackson, v. Hobbs, No. 10-9647 (2012). Krohn, M. D., Thornberry, T. P., Rivera, C., & LeBlanc, M. (2001). Later delinquency careers.

Child delinquents: Development, intervention, and service needs. 67-93. Thousand Oaks, CA: Sage.

Krohn, M. D., Ward, J. T., Thornberry, T. P., Lizotte, A. J., & Chu, R. (2011). The cascading effects of adolescent gang involvement across the life course. Criminology, 49, 991–1028.

Langan, P. A., & Levin, D. (2002). Recidivism of prisoners released in 1994. Washington, DC: Bureau of Justice Statistics.

Lattimore, P. K., Visher, C. A., & Linster, R. L. (1995). Predicting rearrest for violence among serious youthful offenders. Journal of Research in Crime & Delinquency, 32, 54-83.

Lipsey, M. W., & Howell, J. C. (2012). A broader view of evidence-based programs reveals more options for state juvenile justice systems. Criminology & Public Policy, 11, 515-523.

Lipsey, M. W., Howell, J. C., Kelly, M., Chapman, G., & Carver, D. (2010). Improving the effectiveness of juvenile justice programs: A new perspective on evidence-based practice. Washington, DC: Center for Juvenile Justice Reform, Georgetown Public Policy Institute, Georgetown University.

Loeber, R., & Farrington, D. P. (1998). Serious & violent juvenile offenders: Risk factors and successful interventions. Thousand Oaks, CA: Sage.

Loeber, R., & Farrington, D. P. (Eds.). (2001). Child Delinquents: Development, interventions, and service needs. Thousand Oaks, CA: Sage.

Loeber, R., & Farrington, D. P. (2012). From juvenile delinquency to adult crime: Criminal careers, justice policy, and prevention. New York, NY: Oxford University Press.

Loeber, R, Farrington, D.P., & Petechuk, D. (2003). Child delinquency: Early intervention and prevention. Child Delinquency Bulletin Series. Washington, DC: Office of Juvenile Justice and Delinquency Prevention.

Lösel, F., & Farrington, D. P. (2012). Direct protective and buffering protective factors in the development of youth violence. American Journal of Preventive Medicine, 43(Suppl.), S8- S23.

Loughran, T. A., Mulvey, E. P., Schubert, C. A., Fagan, J., Piquero, A. R., & Losoya, S. H. (2009). Estimating a dose-response relationship between length of stay and future recidi- vism in serious juvenile offenders. Criminology, 47, 699-740.

Maltz, M. D. (1984). Recidivism. Orlando, FL: Academic Press. Mass, C., Herrenkohl, T. I., & Sousa, C. (2008). Review of research on child maltreatment and

violence in youth. Trauma, Violence, & Abuse, 9, 56-67. Maxfield, M., & Widom, C. S. (1996). The cycle of violence: Revisited six years later. Archives

of Pediatrics & Adolescent Medicine, 150, 390-395.

Wolff et al. 31

McGrath, A., & Thompson, A. P. (2012). The relative predictive validity of the static and dynamic domain scores in risk-need assessment of juvenile offenders. Criminal Justice and Behavior, 39, 250-263.

Melde, C., & Esbensen, F. (2011). Gang membership as a turning point in the lifecourse. Criminology, 49, 513–552.

Mercy, J. A., & Saul, J. (2009). Creating a healthier future through early interventions for chil- dren. Journal of the American Medical Association, 301, 2262-2264.

Miller v. Alabama, No. 10-9646 (2012). Miner, M. H. (2002). Factors associated with recidivism in juveniles: An analysis of serious sex

offenders. Journal of Research in Crime & Delinquency, 39, 421-436. Moffitt, T. E. (1993). Adolescent-limited and life-course persistent antisocial behavior: A

developmental taxonomy. Psychological Review, 100, 674-701. Monahan, K., Steinberg, L., & Piquero, A. R. (2015). Juvenile justice policy and prac-

tice: A developmental perspective. In M. Tonry (Ed.), Crime and justice: A review of research,557-619. Chicago, IL: University of Chicago Press.

Moylan, C. A., Herrenkohl, T. I., Sousa, C., Tajima, E. A., Herrenkohl, R. C., & Russo, M. J. (2010). The effects of child abuse and exposure to domestic violence on adolescent inter- nalizing and externalizing behavior problems. Journal of family Violence, 25(1), 53-63.

Mulvey, E. P., Schubert, C. A., & Chung, H. L. (2007). Service use after court involvement in a sample of serious adolescent offenders. Children and Youth Services Review, 29, 518-544.

Mulvey, E. P., Steinberg, L., Fagan, J., Cauffman, E., Piquero, A. R., Chassin, L., . . . Losoya, S. H. (2004). Theory and research on desistance from antisocial activity among serious adolescent offenders. Youth Violence and Juvenile Justice, 2, 213-236.

Nagin, D. S., Piquero, A. R., Scott, E. S., & Steinberg, L. (2006). Public preferences for reha- bilitation versus incarceration for juvenile offenders: Evidence from a contingent valuation survey. Criminology and Public Policy, 5, 627-651.

Osgood, D. W., Wilson, J. K., O’malley, P. M., Bachman, J. G., & Johnston, L. D. (1996). Routine activities and individual deviant behavior. American Sociological Review, 635- 655.

Petersilia, J. (1993). Measuring the performance of community corrections. In G. P. Alpertet al. (Eds.), Performance measures of the criminal justice system (pp. 61-84). Washington, DC , U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

Peterson-Badali, M., Skilling, T., & Haqanee, Z. (2015). Examining implementation of risk assessment in case management for youth in the justice system. Criminal Justice and Behavior, 42, 304-320.

Pew Center on the States. (2011). State of recidivism: The revolving door of America’s prisons. Washington, DC: The Pew Charitable Trusts.

Piquero, A. R. (2015). Understanding race/ethnicity differences in offending across the life course: Gaps and opportunities. Journal of Developmental and Life Course Criminology, 1, 21-32.

Piquero, A. R., Cullen, F. T., Unnever, J. D., Piquero, N. L., & Gordon, J. A. (2010). Never too late: Public optimism about juvenile rehabilitation. Punishment & Society, 12, 187-207.

Piquero, A. R., Jennings, W. G., & Barnes, J. C. (2012). Violence in criminal careers: A review of the literature from a developmental life-course perspective. Aggression and Violent Behavior, 17, 171-179.

Reid, J. A., & Piquero, A. R. (2013). Applying general strain theory to youth commercial sexual exploitation. Crime & Delinquency. doi:10.1177/0011128713498213

Roper v. Simmons, 543 U.S. 551 (2005).

32 International Journal of Offender Therapy and Comparative Criminology

Rosenfeld, R., White, H. R., & Esbensen, F. (2012). Special categories of serious and vio- lent offenders: Drug dealers, gang members, homicide offenders, and sex offenders. In R. Loeber & D. P. Farrington (Eds.), From juvenile delinquency to adult crime: Criminal careers, justice policy, and prevention (pp. 118-149). New York, NY: Oxford University Press.

Rutter, M. (1983). Statistical and personal interactions: Facets and perspectives. In D. Magnusson & V. Allen (Eds.), Human development: An interactional perspective (pp. 295-320). New York, NY: Academic Press.

Ryan, J. P., Abrams, L. S., & Huang, H. (2014). First-time violent juvenile offenders: Probation, placement, and recidivism. Social Work Research, 38, 7-18. doi:10.1093/swr/svu004

Schmidt, P., & Witte, A. D. (1988). Predicting recidivism using survival models. New York, NY: Springer.

Schmidt, P., & Witte, A. D. (1989). Predicting criminal recidivism using “split population” survival time models. Journal of Econometrics, 40, 141-159.

Schwalbe, C. S. (2007). Risk assessment for juvenile justice: A meta-analysis. Law and Human Behavior, 31, 449-462.

Shalev, I., Moffitt, T., Sugden, K., Williams, B., Houts, R. M., Danese, A., . . . Caspi, A. (2013). Exposure to violence during childhood is associated with telomere erosion from 5 to 10 years of age: A longitudinal study. Molecular Psychiatry, 18, 576-581.

Smith, C., & Thornberry, T. P. (1995). The relationship between childhood maltreatment and adolescent involvement in delinquency. Criminology, 33, 451-477.

Sternberg, K. J., Lamb, M. E., Greenbaum, C., Cicchetti, D., Dawud, S., Cortes, R. M., . . . Lorey, F. (1993). Effects of domestic violence on children’s behavior problems and depres- sion. Developmental Psychology, 29, 44-52.

Sweeten, G., Pyrooz, D. C., & Piquero, A. R. (2013). Disengaging from gangs and desistance from crime. Justice Quarterly, 30, 469-500.

Taxman, F. S., Caudy, M. S., & Pattavina, A. (2013). Risk-Need-Responsivity (RNR): Leading towards another generation of the model. In F. S. Taxman & A. Pattavina (Eds.), Simulation strategies to reduce recidivism (pp. 285-307). New York, NY: Springer.

Teague, R., Mazerolle, P., Legosz, M., & Sanderson, J. (2008). Linking childhood exposure to physical abuse and adult offending: Examining mediating factors and gendered relation- ships. Justice Quarterly, 25, 313-348.

Thornberry, T. P., Krohn, M. D., Lizotte, A. J., & Chard-Wierschem, D. (1993). The role of juvenile gangs in facilitating delinquent behavior. Journal of Research in Crime and Delinquency, 30, 55–87.

Thornberry, T. P., Krohn, M. D., Lizotte, A. J., Smith, C. A., & Tobin, K. (2003). Gangs and Delinquency in Developmental Perspective. New York: Cambridge University Press.

Tolan, P. H. (1987). Implications of age of onset for delinquency risk identification. Journal of Abnormal Child Psychology, 15, 47–65.

Tolan, P. H., & Thomas, P. (1995). The implications of age of onset for delinquency risk: II. Longitudinal data. Journal of Abnormal Child Psychology, 23, 157–181.

Unnever, J. D., & Gabbidon, S. L. (2011). A theory of African American offending: Race, rac- ism, and crime. New York, NY: Taylor & Frances.

Visher, C. A., Lattimore, P. K., & Linster, R. L. (1991). Predicting the recidivism of serious youthful offenders using survival models. Criminology, 29, 329-366.

Vose, B., Lowenkamp, C. T., Smith, P., & Cullen, F. T. (2009). Gender and the predictive valid- ity of the LSI-R: A study of parolees and probationers. Journal of Contemporary Criminal Justice, 25, 459-471.

Wolff et al. 33

Widom, C. S. (1989). Child abuse, neglect, and violent criminal behavior. Criminology, 27, 251-271.

Widom, C. S., & Maxfield, M. G. (2001). An update on the “cycle of violence.” Washington, DC: National Institute of Justice, U.S. Department of Justice.

Wilson, J. J., & Howell, J. C. (1993). A comprehensive strategy for serious, violent, and chronic juvenile offenders. Washington, DC: U.S. Department of Justice, Office of Juvenile Justice and Delinquency Prevention.

Winokur-Early, K., Hand, G. A., & Blankenship, J. L. (2012). Validity and reliability of the Florida Positive Achievement Change Tool (PACT) risk and needs assessment instru- ment: A three-phase evaluation (validation study, factor analysis, inter-rater reliability). Tallahassee, FL: Justice Research Center.

Wood, J., Foy, D. W., Layne, C., Pynoos, R., & James, C. B. (2002). An examination of the rela- tionships between violence exposure, posttraumatic stress symptomatology, and delinquent activity: An “ecopathological” model of delinquent behavior among incarcerated adoles- cents. Journal of Aggression, Maltreatment, Trauma, 6, 127-147.

Woolard, J. L. (2012). Young adults: How different is “different enough” for social science, law, and policy? Criminology & Public Policy, 11, 723-728.

Wright, K. A., & Rodriguez, N. (2014). A closer look at the paradox: Examining immigration and youth reoffending in Arizona. Justice Quarterly, 31, 882-904.

Yu-Ling Chiu, Y. L., Ryan, J. P., & Herz, D. C. (2011). Allegations of maltreatment and delin- quency: Does risk of juvenile arrest vary substantiation status? Children and Youth Services Review, 33, 855-860.

Author biography

Kevin T. Wolff is an assistant professor at John Jay College of Criminal Justice in New York City. He earned his PhD from the College of Criminology and Criminal Justice at Florida State University. His research interests include the spatial patterning of crime, juvenile justice, crimi- nological theory, and quantitative methods.

Michael T. Baglivio is Director of Research and Program Development for G4S Youth Services, LLC, tasked with examining the effectiveness of juvenile treatment programming on short- and long-term performance measures and outcomes. For the past ten years he has evalu- ated the effectiveness of juvenile justice reform initiatives throughout Florida. His research interests include criminological theory, risk assessment, and life-course criminology.

Alex R. Piquero is Ashbel Smith Professor of Criminology and Associate Dean of Graduate Programs in the School of Economic, Political and Policy Sciences at the University of Texas at Dallas. His research interests include criminal careers, criminological theory, and quantitative research methods. He has received several research, teaching, and service awards and is Fellow of both the American Society of Criminology and the Academy of Criminal Justice Sciences. In 2014, he received The University of Texas System Regents’ Outstanding Teaching Award.