Week 4 Discussion 1- Special Populations: A Challenge to Juvenile Justice

The Impact of Victimization and Mental Health Symptoms on Recidivism for Early System-Involved Juvenile Offenders

Lindsey E. Wylie University of Nebraska Omaha

Katrina A. Rufino University of Houston–Downtown and The Menninger Clinic,

Baylor College of Medicine

Although research has linked mental health symptoms and prior victimization to recidivism for youth on probation or in detention, little attention has been given to these risk factors for early system-involved youth. We conducted a survival/hazard model to estimate the impact of official records of abuse/neglect, crime victimization, and mental health issues (mood, anxiety, disruptive, and substance use disorders) on recidivism in a sample of 2,792 youth in a large Midwestern diversion program. Results indicated that youth with official records of abuse/neglect, person crime victimization, and property crime victimization were more likely to recidivate sooner than those without these victimization experiences (hazard ratio: 1.37, 1.42, and 1.52, respectively). Findings from the present study also demonstrated that substance use disorder was the only mental health cluster that predicted quicker time to recidivism. As one of the earliest points of entry into the juvenile justice system, diversion programs are in a unique position to address trauma from multiple types of victimization and adapt diversion programming to be responsive to each juvenile’s mental health needs.

Public Significance Statement Early system-involved youth referred to juvenile diversion had high levels of mental health symp- toms and many had prior experiences with various types of victimization that are based on official law enforcement records. Prior victimization significantly predicted whether a youth had future contact with the juvenile or adult criminal justice system, even while considering other factors, such as risk level and youth characteristics.

Keywords: juvenile recidivism, juvenile diversion, mental health, victimization

In 2016, there were approximately 856,130 juvenile arrests in the United States—many for nonviolent offenses such as larceny–theft, other assaults, drug abuse violations, liquor law violations, vandalism, disorderly conduct, and curfew/loitering (OJJDP, 2016). As such, the juvenile justice system is often tasked with how to address youth who commit less serious offenses. One approach is to divert them away from formal juvenile justice system involvement through diversion programs. As the gateway to the juvenile justice system, diversion programs are in a unique position to address the needs of early system-involved youth, including needs related to victimization and mental health symptoms, to reduce future involvement in the juvenile or adult criminal justice system.

Developmental models of antisocial behavior propose that “delin- quency is marked by a reliable developmental sequence of experi-

ences,” in which childhood experiences and social environment put children at risk for social maladjustment and criminal behavior (Pat- terson, Debaryshe, & Ramsey, 1989, p.263). Specifically, studies find that experiences with victimization, broadly defined as maltreatment, adverse childhood experiences, and general crime victimization, are related to mental health issues (e.g., Abram et al., 2004; Kilpatrick et al., 2000) and that both victimization and mental health issues are related to juvenile justice involvement (e.g., Barrett, Katsiyannis, Zhang, & Zhang, 2014; Fazel, Doll, & Långström, 2008). Although the association of victimization and mental health symptoms within juvenile justice populations are well-documented, especially within samples of serious juvenile offenders (e.g., adjudicated or incarcer- ated), fewer studies have examined these risk factors in a sample of early system-involved youth. The purpose of this study is to examine the relationship between prior victimization, as obtained from official law enforcement records, and mental health symptoms on time to recidivism in a sample of early system-involved youth in a juvenile diversion program.

Operationalizing Victimization

Researchers operationalize victimization using multiple defini- tions. Most studies measure victimization as child maltreatment, utilizing official data obtained from social service agencies or

This article was published Online First November 1, 2018. Lindsey E. Wylie, School of Criminology and Criminal Justice, Juvenile

Justice Institute, University of Nebraska Omaha; Katrina A. Rufino, De- partment of Social Sciences, University of Houston–Downtown and The Menninger Clinic, Baylor College of Medicine.

Correspondence concerning this article should be addressed to Lindsey E. Wylie, Juvenile Justice Institute, University of Nebraska Omaha, 941 O Street, Suite 706, Lincoln, NE 68508. E-mail: [email protected]

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Law and Human Behavior © 2018 American Psychological Association 2018, Vol. 42, No. 6, 558 –569 0147-7307/18/$12.00 http://dx.doi.org/10.1037/lhb0000311

558

child protective services (e.g., Barrett et al., 2014; English, Wi- dom, & Brandford, 2002; Smith, Ireland, & Thornberry, 2005), or self-report data obtained from caregivers or youth (e.g., Conrad, Tolou-Shams, Rizzo, Placella, & Brown, 2014). Other studies include broader definitions of victimization, usually measured with self-report data, including adverse childhood experiences (ACEs), such as abuse/neglect, parental divorce, and family violence (e.g., Wolff, Baglivio, & Piquero, 2015; Kilpatrick et al., 2003) or general crime victimization, such as theft or assault (e.g., Finkel- hor, Ormrod, & Turner, 2009; Manasse & Ganem, 2009). Research employing these broader definitions of victimization typically have not included data from official agency records.

As such, the current study expands previous research using a broader definition of victimization, to include abuse/neglect, sex- ual assault, property crimes, and person crimes, utilizing reported incidents of victimization data obtained from official law enforce- ment records. Although official records are likely an underestima- tion of abuse/neglect (Swahn et al., 2006) or general crime trends (see Loftin & McDowall, 2010) because of failure to report or other system-wide factors, using this definition has practical im- plications for programmatic interventions because this information may be readily available to diversion programs, and may produce different findings than studies using self-report data.

Victimization and Mental Health Symptoms

Research demonstrates that victimization as a child or adoles- cent is associated with later mental health problems in both lon- gitudinal studies with representative samples (e.g., Finkelhor et al., 2009; Kilpatrick et al., 2000; Manasse & Ganem, 2009) and retrospective studies with justice-involved samples (e.g., Barrett et al., 2014; Dierkhising et al., 2013; Ford, Grasso, Hawke, & Chap- man, 2013). For instance, in a national random sample of non- justice-involved children ages 2 to 17, Finkelhor and colleagues (2009) examined the Developmental Victimization Survey (DVS) to assess the range of childhood victimizations across five victim- ization types, including conventional crime (e.g., assaults and property crimes), child maltreatment, peer and sibling victimiza- tion, sexual assault, and indirect victimization (e.g., witnessing violence). Overall, 79.6% of the sample reported lifetime victim- ization and analysis revealed a strong association between lifetime poly victimization (the total number of different types of victim- izations) and mental health symptoms (anger, depression, and anxiety).

When examining samples of justice-involved youth, a large proportion of youth report exposure to some type of potentially traumatic event. Dierkhising and colleagues (2013) reported up to 90% of justice-involved youth experienced one or more potentially traumatizing events (e.g., traumatic loss, impaired caregiver, do- mestic violence, school violence), with an average of 4.9 lifetime events. In measuring the link between victimization and mental health problems, Ford and colleagues (2013) found that 58% of the sample endorsed one of the 19 potentially traumatic events (e.g., being in a bad accident, witnessing violence, sexual assault), which was related to posttraumatic stress symptoms, emotional and be- havioral problems, suicide risk, and alcohol and drug use prob- lems.

Victimization and Delinquency

Within criminological perspectives and strain theory, Agnew (1992) argued that criminal victimization is “among the types of strain that are most likely to lead to delinquency” (Agnew, 1992, p.306) because it is perceived as unjust and traumatic, which evokes anger and resentment, and contributes to deviance as a mechanism to cope with strain (Agnew, 1992; Hay & Evans, 2006). Studies have examined whether victimization increases the risk for later delinquency, including general delinquency and vio- lent delinquency, in both representative samples (i.e., longitudinal studies of nonjustice-involved children) and justice-involved sam- ples (i.e., retrospective studies of youth who are justice-involved).

Studies including representative samples found that those with a history of child maltreatment were significantly more likely to have contact with the police as a juvenile or adult than those without a history of child maltreatment (Smith & Thornberry, 1995; Smith et al., 2005; Zingraff, Leiter, Myers, & Johnson, 1993). Although the link between maltreatment and delinquency is consistent across longitudinal studies, the impact of maltreatment on future violence may be less predictive than other factors, including antisocial peers, substance abuse, and family socioeco- nomic status (Hawkins et al., 2000). Furthermore, research dem- onstrates that general crime victimization is associated with delin- quency. For example, using a measure of general crime victimization that included theft, assault, parental physical abuse, attack by a weapon, and property damage, Manasse and Ganem (2009) found that for every one point on their victimization mea- sure, the odds of engaging in delinquency increased by 19%.

Victimization also impacts future reoffending after initial justice involvement. Wolff and colleagues (2015) tested whether exposure to ACEs, measured using a sum of 10 binary absence/presence indicators, significantly predicted time to recidivism following community-based treatment. Overall, having a greater number of ACEs was related to a shorter time to recidivism, even while controlling for youth demographics, and risk factors such as sub- stance abuse, criminal history, deviant peers, and school behavior. Others found that gender may also moderate the relationship between victimization and delinquency. Specifically examining the association between prior sexual abuse and delinquency in a sample of 454 juveniles referred by a judge for a mental health evaluation, Conrad and colleagues (2014) indicated that being sexually abused as a child was the strongest predictor of recidivism for girls but not boys, even while controlling for prior legal involvement and conduct problems.

Mental Health Symptoms

There is growing attention on the high prevalence of mental health problems in the juvenile justice system and researchers have consistently found higher rates of mental health problems in justice-involved youth, than youth in the general population (Abram et al., 2004; Dierkhising et al., 2013; Fazel et al., 2008; Teplin, Abram, McClelland, Dulcan, & Mericle, 2002). A meta- analysis by Fazel and colleagues (2008) examined the results of 25 studies that included interviews with detained juveniles and found that juveniles in detention and correctional facilities were signifi- cantly more likely to have mental health disorders (conduct dis- order, psychosis, attention-deficit/hyperactivity disorder, and ma- jor depression) than age-equivalent juveniles in the general

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

559VICTIMIZATION AND MENTAL HEALTH ON RECIDIVISM

population. Overall prevalence rates, however, appear to differ by gender. In a large random sample of youth interviewed during detention intake, Teplin and colleagues (2002) estimated that ap- proximately two thirds of males and three quarters of females met criteria for one or more psychiatric disorders, including disruptive disorders (attention deficit disorders, oppositional defiant disor- der), substance use disorders, and affective disorders.

Although the high prevalence of mental health issues within justice-involved youth has been well-documented, previous re- search has been mixed with respect to whether mental health is predictive of reoffending, and if so, what types of disorders may be most predictive. Cottle, Lee, and Heilbrun’s (2001) meta-analysis of 23 studies that measured recidivism in juveniles indicated that conduct problems (e.g., conduct disorder) and nonsevere psycho- pathology (e.g., stress, anxiety) elevated the risk for recidivism, but severe psychopathology (e.g., psychosis, suicidality) and a history of psychiatric treatment did not increase risk for recidi- vism. On the other hand, in a study of youth referred to probation, the results indicated that while anxiety and mood disorders did not relate to recidivism, both substance use disorders and disruptive behaviors did increase the risk of recidivism (McReynolds, Schwalbe, & Wasserman, 2010).

Trauma-exposed youth who are justice-involved are at risk for several mental health disorders, including posttraumatic stress disorder, major depressive disorder, and substance abuse (Abram et al., 2004; Dierkhising et al., 2013; Ford, Hartman, Hawke, & Chapman, 2008). When testing the role of mental health in the victimization-delinquency link, participants who endorsed depres- sive symptoms were more likely to respond to general crime victimization (i.e., theft, assault, parental physical abuse, attack by a weapon, and property damage) with delinquent behavior. There was a moderating effect of gender, such that males who experi- enced depressive symptoms were 50 times more likely to respond to victimization with delinquency than males without depressive symptoms; but there were no differences for females with depres- sive symptoms (Manasse & Ganem, 2009).

A limitation of these studies, however, is that they only provide support for an association between mental health problems and risk of delinquency, but do not explain the mechanism by which mental health problems increase this risk. Recent research has sought to “disentangle shared risk” to evaluate whether having a mental health problem explained delinquency outcomes above criminogenic risk factors and whether mental health issues mod- erate the relationship between criminogenic risk and delinquency outcomes (Guebert & Olver, 2014; Schubert, Mulvey, & Glasheen, 2011). These studies demonstrated that even though the presence of mental health problems was related to reoffending in bivariate analyses, after controlling for criminogenic risk markers and de- mographics, mental health issues did not uniquely contribute to reoffending above these other factors (Guebert & Olver, 2014; Schubert et al., 2011).

The Current Study

The present study utilized a sample of early system-involved youth referred to a juvenile diversion program in a large Midwest- ern city. The purpose of this study was to examine reoffending for youth with reported experiences of victimization, as well as mental health symptoms at the time of diversion intake. Although research

has examined the recidivism trajectory of youth at the deeper end of the juvenile justice system, fewer studies have linked victim- ization and mental health problems to recidivism in a sample of early system-involved youth. Juveniles in the diversion program are typically first-time offenders referred because of minor of- fenses (e.g., shoplifting, possession of marijuana, status offenses) and assessed as low to moderate risk. The present research con- tributes to the larger body of literature by examining whether the association between victimization, mental health problems, and recidivism is similar for early system-involved youth to better inform diversion efforts. Furthermore, the present study extends prior research by including a broader measure of victimization that includes abuse/neglect, sexual assault, property crime, and person crimes that have been reported to law enforcement.

Method

Participants

Participants included 2,792 justice-involved juveniles referred for diversion in a large Midwestern city. The mean age was 15.08 (SD � 1.64) and the majority were male (59.7%, n � 1,668). Approximately half identified as White (48.4%, n � 1,352), fol- lowed by Black (34.1%, n � 953), Hispanic/Latino (14.6%, n � 407), Asian/Pacific Islander (1.1%, n � 31), Native American/ Alaskan Native (1.1%, n � 31), and other or multiple races (0.6%, n � 17). Most participants were referred to diversion for drug or alcohol-related offenses such as possession of marijuana or para- phernalia (35.2%, n � 984) and property offenses, such as shop- lifting and theft (35.5%, n � 990). Other offenses included disor- derly conduct (11.7%, n � 326), crimes against others such as third-degree assault (9.5%, n � 265), traffic offenses such as driving without a license (2.5%, n � 69), and other offenses such as vandalism, curfew violation, providing false information to the police, and obstructing an officer (5.7%, n � 158). Although youth may be referred to diversion for truancy in this county, this sample does not include those youth, because truancy diversion is a separate program that utilizes a different assessment process. If youth successfully complete the diversion program, the county attorney does not file their case and they are not adjudicated delinquent.

Study Design and Procedure

Data were obtained from the juvenile diversion program’s case management system as part of a statutorily required statewide evaluation of juvenile justice-related programs that receive fund- ing from the state. The data included identifying information (e.g., name and date of birth) so that we could compute recidivism, as required under statute for the statewide program evaluation. Insti- tutional review board approval was obtained by the University of Nebraska Medical Center as part of the program evaluation.

Between July 1, 2012 and June 30, 2015, a total of 3,934 juveniles were referred to the juvenile assessment center for pos- sible participation in juvenile diversion following a law violation. Once a juvenile receives a law violation, if eligible based on evidentiary factors and the type of offense, the county attorney will refer the case to the assessment center to determine whether the youth should participate in the diversion program. Of the youth

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

560 WYLIE AND RUFINO

referred during this time, 2,792 were eligible and decided to participate in diversion, which is the total sample for the study. Youth may be referred, but not enroll for various reasons includ- ing: receiving a warning letter if they are screened as lower risk, they were deemed not eligible by the diversion program, the youth or family refused to participate, or procedural reasons (e.g., out of jurisdiction, recommend nolle pros, the youth received a new charge while awaiting assessment, or the county attorney withdrew the referral).

During the assessment process, each juvenile completes several assessments and the assessment specialist creates a diversion plan based on each juvenile’s risk and needs. If the youth successfully completes the diversion plan, their case is dismissed and not filed in juvenile court. If a youth is not successful, either because he or she did not complete the diversion plan requirements or receives a new law violation while on diversion, then the case is filed and the youth goes through the traditional juvenile court process.

Measures

Demographics. We included age, gender, and race/ethnicity as demographic controls. Age was measured continuously as the youth’s age at the time of referral. Gender was dichotomous (0 � male, 1 � female). Race/ethnicity was measured using a dichot- omous variable with 0 � White and 1 � non-White.

Successful completion of diversion. To control for success- ful completion of the diversion program, which may influence future reoffending, we included a dichotomous variable for suc- cessful discharge (0) or unsuccessful discharge (1) as measured within the juvenile diversion program’s case management system. Overall, 83.5% of youth successfully completed their diversion program requirements.

Risk-level. To measure each juvenile’s level of risk, the Youth Level of Service/Case Management Inventory 2.0 (YLS/ CMI 2.0; Hoge & Andrews, 2011) was used. The YLS/CMI 2.0 is a 42-item checklist designed to be completed by a mental health professional or a probation officer utilizing interviews and/or record reviews. The measure provides a total score and a score for each of eight subscales including: Offense History, Family, Edu- cation, Substance Abuse, Leisure/Recreation, Peer Relations, Per- sonality/Behavior, and Attitudes/Orientation. Each item is coded as present or absent for a total score ranging from 0 to 42. These total scores determine if the juvenile is low (0 through 8), moderate (9 through 22), high (22 through 34), or very high (35 through 42) risk for recidivism. The information for each item is gathered from the youth and family directly, and through collateral information (e.g., the school, other agencies). The assessment tool is for youth aged 12 to 18 years old, therefore youth in the sample younger than 12 were not assessed via the YLS/CMI. The manual provides evidence of strong reliability and validity (Hoge & Andrews, 2011) and was validated elsewhere (Onifade et al., 2008; Vincent, Guy, & Grisso, 2012). On average, the sample was a low to moderate risk group with a mean YLS/CMI-total score of 7.46 (SD � 4.21). Most of the sample scored in the low risk range (65.7%) or moderate risk range (34.2%), with only three partici- pants scoring in the high-risk range (0.1%) and none in the very high-risk range.

Official records of victimization. As part of the diversion program’s case management system, data is automatically pulled

from law enforcement any time the youth is listed as a victim. As such, the measure of victimization includes only reported incidents that involved law enforcement. Each incident of victimization was recoded into one of four categories: sexual assault, abuse/neglect, property crime (e.g., burglary, theft, robbery) victimization, and person crime (e.g., assault) victimization. Property and person crimes were classified based on the FBI Uniform Crime Reporting Program (U.S. Department of Justice, 2016), with person crimes including those where the individual is the direct victim and property crimes including instances where the object is to obtain money or some other benefit. Overall, 16.3% of the sample had at least one occurrence of reported victimization, 1.9% had two occurrences, and 0.3% had three occurrences of reported victim- ization. The prevalence of each victimization type was: abuse/ neglect (6.9%), sexual assault (1.7%), victim of a person crime (9.9%), and victim of a property crime (2.6%).

Mental health symptoms. The DISC Predictive Scales (DPS; Lucas et al., 2001) was used to measure each juvenile’s presenting symptoms at the time of their assessment. The DPS was developed as an efficient diagnostic screening tool for juveniles and identifies youth who are highly likely to meet diagnostic criteria (McReyn- olds, Wasserman, Fisher, & Lucas, 2007). The DPS was validated in a community and mixed sample of “troubled” 10- to 18-year-old youth of both genders (Lucas et al., 2001) and with justice- involved juveniles (McReynolds et al., 2007). The DPS is a com- puterized self-report tool that uses audio to read each question. The number of questions for each module varies depending on the rule-out criteria and follow-up questions based on flagged re- sponses. The DPS derives from the most sensitive questions con- tained in the Diagnostic Interview Schedule for Children-2.3 (DISC; Shaffer, Fisher, Lucas, Dulcan, & Schwab-Stone, 2000) to determine if symptoms are “present,” “possible,” or “absent” within the last year.

For the purposes of these analyses, we combined “possible” and “absent” for a dichotomous measure of each symptom as either present (1) or possible/absent (0). For each of the symptoms, we organized them into four clusters similar to previous studies (McReynolds et al., 2007, 2010; Wasserman, McReynolds, Lucas, Fisher, & Santos, 2002). The clusters utilized in the analyses were: disruptive behaviors (attention deficit hyperactivity, oppositional defiant, and conduct disorder), substance use (alcohol use, mari- juana use, and other substance use), anxiety (posttraumatic stress disorder, agoraphobia, social phobia, general anxiety, obsessive– compulsive disorder, specific phobia, and panic disorder), and mood (depression and mania). Overall, 63.8% of the juveniles in this sample endorsed one or more mental health symptom clusters: 41% endorsing the disruptive cluster, 35.3% the anxiety cluster, 21.7% the substance use cluster, and 19.6% the mood cluster.

Recidivism. Recidivism data were obtained from the state’s trial court case management system and was defined as any of- fense that was filed in court following discharge from diversion, excluding cases that were eventually dismissed. Data included all juvenile and adult misdemeanor and felony cases between July 1, 2012 and December 31, 2015, including sealed records. These dates allowed at least a 6-month recidivism period for juveniles enrolling in diversion at the end of the study period (June 30, 2015). The time at risk from discharge to the end of the study period ranged from 180 days to 1,271 days, with a mean of 850.83 days (SD � 389.60 days, Mdn � 859.00 days).

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

561VICTIMIZATION AND MENTAL HEALTH ON RECIDIVISM

Using probabilistic record linkage software, we matched youth in the sample to the recidivism records using first name, middle name, last name, and date of birth, which accommodates mis- spelled names or typos with dates of birth. We included any offenses, including status offenses, that may bring a youth back in to the juvenile justice system or diversion, which in this state, includes truancy offenses eligible to be filed on in court (20 or more absences that are not medically related). Of those who recidivated (n � 839, 30.1%), the most common offenses included drug or alcohol-related offenses (n � 255, 30.4%) and property offenses (n � 229, 27.3%); the remaining offenses included dis- orderly conduct (7.2%, n � 60), crimes against others such as assault and robbery (7.7%, n � 67), traffic offenses such as driving without a license (2.5%, n � 21), truancy (10.7%, n � 90), and other offenses such as vandalism, curfew violation, providing false information to the police, and obstructing an officer (13.9%, n � 117).

Data Analysis and Hypotheses

First, several bivariate analyses were conducted to compare juveniles who endorsed victimization and those who did not. The victimization types, gender, and reason for discharge were com- pared to each other and with each mental health cluster using a chi-square test due to the dichotomous nature of the variables. Differences for gender and victimization on age and risk level were compared using analysis of variance. To test the relationship of the independent variables while controlling for demographic factors and risk level, data were analyzed using a hazard Cox regression analysis to predict time to failure (i.e., recidivating event), which provides more details than a simple dichotomous indicator of recidivism because it specifically targets the amount of time to reoffending (Wolff et al., 2015). Prior research has indi- cated that the time to the recidivating event is an important outcome to consider because the individual characteristics of ju- veniles who recidivate sooner may be different than the charac- teristics of juveniles who recidivate later (Maltz, 1984; Schmidt & Witte, 1989).

The model specifically included the four victimization types and the four mental health symptom clusters across gender and race/ ethnicity, while controlling for age, risk level, and successful discharge. For the purposes of the final multivariate model, we removed the Substance Abuse subscale from our measure of risk level because of the overlap between the YLS/CMI Substance Use subscale and the mental health substance use cluster measured with the DISC Predictive Scales. Preliminary analyses confirmed the overlap between the YLS/CMI-total and the substance use mental health cluster. Specifically, when including the YLS/CMI- total and the substance use mental health cluster in the model, the substance use mental health cluster was not significant; however, when including the modified risk level score without the Substance Use subscale, the substance use cluster remained a significant predictor of time to recidivism. As such, the measure of risk level in the multivariate model includes the modified YLS/CMI without the Substance Use subscale and is identified as the modified risk score in analyses.

On the basis of previous research on the role of victimization in juvenile delinquency, we hypothesized that juveniles who had been victimized would recidivate sooner than juveniles who had

not been victimized (Wolff et al., 2015) for all four types of victimization. Consistent with previous research on the relation- ship between mental health symptoms on offending and recidi- vism, we also hypothesized that mental health issues would predict recidivism at the bivariate level, especially for the disruptive behavior and substance use clusters (McReynolds et al., 2010; Vermeiren, Schwab-Stone, Ruchkin, De Clippele, & Deboutte, 2002); however, it was hypothesized that once we controlled for other factors such as risk level, then mental health symptoms would be less predictive (Guebert & Olver, 2014; Schubert et al., 2011).

Results

Bivariate Relationships

First, we compared the bivariate relationship of those who had experienced a reported incident of victimization, compared with those who had not, to gender, race/ethnicity, and discharge reason for each victimization type (see Table 1). With respect to gender, the only victimization type that was significantly different by gender was sexual assault, �2(1) � 37.68, p � .001, as females have 4.9 times greater odds of reporting sexual assault than males. There were no gender differences for abuse/neglect, �2(1) � 0.17, p � .68, property crime, �2(1) � 0.94, p � .33, or person crime, �2(1) � 0.08, p � .79. There were differences by race/ethnicity across three of the four victimization types. Although White youth were 1.71 times more likely to report property victimization inci- dents than non-White youth, �2(1) � 4.75, p � .029, non-White youth were 1.63 times more likely to report victimization incidents of person crimes, �2(1) � 14.20, p � .001, and 1.92 times more likely to report abuse/neglect, �2(1) � 17.57, p � .001, than White youth. There were no differences for race/ethnicity and sexual assault, �2(1) � 1.53, p � .22. For discharge reason, youth with a reported instance of abuse/neglect, �2(1) � 13.46, p � .001, were 1.87 times more likely to be unsuccessfully discharged than those without a reported abuse/neglect incident. Those with a reported instance of a person crime, �2(1) � 18.68, p � .001, were 1.08 times more likely to be unsuccessfully discharged than those without a reported instance of a person crime. There were no differences for discharge reason for sexual assault victimization, �2(1) � 0.17, p � .68, or property crime, �2(1) � 3.82, p � .05.

Similar bivariate comparisons were conducted for the mental health symptom clusters. Female youth were 2.93 times more likely to endorse the anxiety cluster, �2(1) � 165.89, p � .001, 3.57 times more likely to endorse the mood cluster, �2(1) � 163.80, p � .001, and 1.47 times more likely to endorse the disruptive cluster, �2(1) � 22.99, p � .001, than male youth; however, there were no differences for the substance use cluster, �2(1) � 3.18, p � .08. For race/ethnicity, there were differences between White and non-White youth for the disruptive cluster, �2(1) � 30.02, p � .001, substance use cluster, �2(1) � 88.81, p � .001, and anxiety cluster, �2(1) � 22.56, p � .001. More specif- ically, White youth were 2.49 times more likely to endorse sub- stance use symptoms than non-White youth, and non-White youth were 1.55 times more likely to endorse disruptive disorder symp- toms and 1.48 times more likely to endorse anxiety symptoms than White youth. There were no racial/ethnic differences for the mood cluster, �2(1) � 0.23, p � .63. With respect to successfully

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

562 WYLIE AND RUFINO

completing the diversion program, youth endorsing the disruptive disorder cluster, �2(1) � 14.33, p � .001, were 1.5 times less likely to be successfully discharged and youth endorsing the sub- stance use cluster, �2(1) � 9.17, p � .01, were 1.44 times less likely to be successfully discharged from the program. On the other hand, youth with anxiety symptoms, �2(1) � 0.88, p � .35, and mood disorders, �2(1) � 1.27, p � .26, were equally as likely to complete the program successfully.

Table 2 presents mean ages and YLS/CMI risk scores for each victimization type and mental health cluster. Overall, youth with reported victimization were approximately the same age as youth without reported victimization for sexual assault, F(1, 2789) � 0.52, p � .47, abuse/neglect, F(1, 2789) � 3.40, p � .07, and person crime, F(1, 2789) � 2.58, p � .11. However older youth were more likely to have experienced a reported property crime incident as compared to those younger, F(1, 2789) � 10.767, p � .001. For most types of victimization, youth had higher YLS/CMI risk-level scores than youth without a reported instance of victim- ization. More specifically, those reporting sexual assault, F(1, 2668) � 7.97, p � .01, abuse/neglect, F(1, 2668) � 14.07, p � .001, and person crime, F(1, 2668) � 13.99, p � .001, scored higher on the YLS/CMI than those without reported experiences. There were no differences on YLS/CMI risk-level scores between youth with a reported property crime victimization and youth without a reported property crime victimization, F(1, 2668) � 0.01, p � .97.

Next, we conducted several bivariate comparisons to compare juveniles who recidivated (using a binary yes/no variable) and those who did not (see Table 3). Results demonstrated there were significant gender differences, �2(1) � 28.81, p � .001, with males 1.59 more likely to recidivate than females; and significant racial/ethnic differences, �2(1) � 38.93, p � .001, with non-White youth 1.52 more likely to recidivate than White youth. It should be noted, however, that racial/ethnic differences are not necessarily indicative of greater delinquency in non-White youth, but could stem from system-wide issues of disproportionate minority con- tact/racial ethnic disparities. Those who were unsuccessfully dis-

charged were 1.91 times more likely to recidivate than those who were successfully discharged, �2(1) � 38.93, p � .001. For the victimization types, results revealed significant differences in re- cidivism rates for abuse/neglect, �2(1) � 7.97, p � .01, person crime victimization, �2(1) � 16.16, p � .001, and property crime victimization, �2(1) � 7.29, p � .01, showing that victims of abuse/neglect were 1.54 times, victims of person crimes were 1.68 times, and victims of property crimes were 1.90 times more likely to recidivate. There was, however, no significant difference in recidivism for participants who were the victims of sexual assault, �2(1) � 0.21, p � .65. With respect to mental health variables, there were not significant differences in recidivism for juveniles who endorsed anxiety symptoms, �2(1) � 3.69, p � .06, or disruptive behavior symptoms, �2(1) � 0.59, p � .44. On the other hand, there were significant differences for both mood symptoms, �2(1) � 4.15, p � .05, and substance use symptoms, �2(1) � 4.18, p � .05. More specifically, while juveniles endorsing the sub- stance use cluster were 1.22 times more likely to recidivate, juveniles endorsing the mood cluster were 1.25 times less likely to recidivate.

In comparing the age at intake to diversion of those who recidivated (M � 15.10, SD � 1.49) to those who did not (M � 15.08, SD � 1.70), there were no age differences between the two groups, F(1, 2789) � .16, p � .69. Moreover, there were signif- icant differences by risk level measured with the YLS/CMI total score, F(1, 2668) � 81.73, p � .001, on recidivism, as those who recidivated scored higher on the YLS/CMI (M � 8.56, SD � 4.42) than youth who did not recidivate (M � 6.98, SD � 4.02).

Modeling Variables on Time to Recidivism

Last, we conducted a survival/hazard analysis with a stepwise Cox regression to examine if time to recidivism (i.e., days to failure) differed based on a history of reported victimization and mental health symptom clusters. A hazard ratio (HR) greater than 1 indicates a shorter time to failure. In other words, juveniles with that characteristic or risk factor recidivated more quickly than

Table 1 Gender, Race, and Discharge Type by Victimization Type and Mental Health Clusters (in Percentages)

Variable

Gender Race Discharge

Male Female White Non-White Successful Unsuccessful

Victimization Sexual assault .5%a 3.6%b 1.4% 2.0% 1.7% 1.9% Abuse/neglect 6.7% 7.1% 4.8%a 8.8%b 6.1%a 10.8%b

Person crime 10.0% 9.7% 7.7%a 12.0%b 8.8%a 15.4%b

Property crime 2.8% 2.2% 3.3%a 1.9%b 2.3% 3.9% Mental health

Anxiety 25.4%a 49.9%b 30.8%a 39.6%b 34.9% 37.3% Mood 11.5%a 31.6%b 19.2% 20.0% 19.2% 21.6% Disruptive 37.3%a 46.7%b 35.7%a 46.2%b 39.5%a 49.4%b

Substance use 20.6% 23.5% 29.5%a 14.4%b 20.7%a 27.3%b

Note. Percentages are presented as the presence of victimization and mental health within each level of gender, race, and discharge type. Significantly different bivariate comparisons among gender, race, and discharge type are represented with different superscripts. For example, for gender (e.g., male vs. female), the superscripts indicate that there was a significant difference for sexual assault, anxiety, mood, and disruptive. All differences were significant at p � .05.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

563VICTIMIZATION AND MENTAL HEALTH ON RECIDIVISM

juveniles in the comparison group, after controlling for the other variables in the model (Chung, Schmidt, & Witte, 1991; Wolff et al., 2015). Several iterative regression models were estimated to evaluate the variables of interest on recidivism. The results of this analysis are presented in Table 4, and the HR function presented by each type of victimization is presented in Figure 1.

The first model with only the youth characteristics was signif- icant, �2(3) � 40.35, p � .001; race/ethnicity and gender signif- icantly predicted recidivism, but age was not a significant predic- tor. Specifically, juveniles who were non-White and male recidivated sooner than juveniles who were White and female. In Model 2, we added discharge type and risk level measured with the modified YLS/CMI total score (i.e., without the Substance Use subscale); the model was significant, �2(5) � 117.11, p � .001. More specifically, juveniles with higher risk scores and who were unsuccessfully discharged recidivated sooner than juveniles with lower risk scores and those who successfully completed the pro- gram. Gender and race/ethnicity remained significant, and age became a significant predictor in Model 2.

Next, the four victimization variables were added to Model 3, �2(9) � 143.08, p � .001, and it remained significant. Results revealed that victims with reported incidents of abuse/neglect, person crimes, and property crimes were more likely to recidivate

at a faster rate, but there was not a significant difference for victims of reported sexual assault. The pattern for youth charac- teristics were similar to the previous model. The final model, which added the four mental health clusters, significantly pre- dicted time to recidivism, �2(13) � 153.89, p � .001, though the substance use cluster was the only mental health variable to predict time to recidivism. This full model revealed that juveniles who recidivated sooner were more likely to be male (HR � .75; CI [.64, .88]), non-White (HR � 1.21; CI [1.04, 1.41]), unsuccessfully discharged (HR � 1.31; CI [1.09, 1.57]), have a higher modified YLS/CMI risk-level score (HR � 1.06; CI [1.04, 1.09]), have a reported incident of abuse/neglect (HR � 1.37; CI [1.07, 1.75]), person crime victimization (HR � 1.41; CI [1.14, 1.74]), and property crime victimization (HR � 1.52; CI [1.06, 2.19]), as well as substance use symptoms (HR � 1.12; CI [1.02, 1.43]). As the HRs indicate, the three significant victimization variables were the strongest predictors of time to recidivism, relative to the other significant variables of race/ethnicity, gender, modified risk score, and discharge reason.

Table 2 Age and YLS/CMI Risk Level for the Presence and Absence of Victimization Type and Mental Health Clusters

Variable

Age YLS

M (SD) M (SD)

Victimization Sexual assault

Present 14.92 (1.29) 9.26 (4.53)a

Absent 15.09 (1.64) 7.43 (4.20)b

Abuse/neglect Present 14.88 (1.68) 8.59 (4.17)a

Absent 15.10 (1.63) 7.38 (4.20)b

Person crime Present 14.93 (1.70) 8.40 (4.67)a

Absent 15.10 (1.63) 7.36 (4.14)b

Property crime Present 15.71 (1.42)a 7.48 (3.39) Absent 15.07 (1.64)b 7.46 (4.23)

Mental health Anxiety

Present 14.78 (1.65)a 8.20 (4.36)a

Absent 15.31 (1.51)b 7.10 (4.07)b

Mood Present 14.92 (1.56)a 8.69 (4.28)a

Absent 15.17 (1.59)b 7.19 (4.14)b

Disruptive Present 14.68 (1.60)a 8.89 (4.49)a

Absent 15.43 (1.49)b 6.51 (3.71)b

Substance use Present 15.64 (1.24)a 8.93 (4.74)a

Absent 14.98 (1.64)b 7.07 (3.95)b

Note. Significantly different bivariate comparisons between the presence and absence of victimization and mental health are represented with different superscripts. For example, for property crime, the superscripts indicate that there was a significant difference in age for those with and without a reported incident of property crime (e.g., present vs. absent). YLS/CMI � Youth Level of Service/Case Management Inventory. All differences were significant at p � .05.

Table 3 Bivariate Comparisons of Demographics, Victimization Types, and Mental Health Clusters on Recidivism (in Percentages)

Variable Recidivated

Gender Male 33.9%a

Female 24.4%b

Race/ethnicity White 25.6%a

Non-White 34.4%b

Discharge Successful discharge 27.6%a

Unsuccessful discharge 42.2%b

Victimization type Sexual assault

Presence 27.1% Absence 30.1%

Abuse/neglect Presence 39.1%a

Absence 29.4%b

Person crime Presence 40.6%a

Absence 28.9%b

Property crime Presence 44.4%a

Absence 29.7%b

Mental health Anxiety

Presence 28.0% Absence 31.6%

Mood Presence 26.6%a

Absence 31.2%b

Disruptive Presence 31.2% Absence 29.8%

Substance use Presence 33.8%a

Absence 29.4%b

Note. Significantly different bivariate comparisons within each variable represented with different superscripts. For example, for gender (e.g., male vs. female), the superscripts indicate that there was a significant difference for gender on recidivism. All differences were significant at p � .05.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

564 WYLIE AND RUFINO

Discussion

The juvenile justice system is built on the notion that young people are malleable and have a likelihood of rehabilitation, which may be especially true for lower risk youth who come into contact with law enforcement for less serious offenses. Assessing the extent to which specific risk factors are associated with continued reoffending after initial juvenile justice involvement, is important to consider in the development of theoretical frameworks and the implementation of evidence-based practices aimed at reducing future system involvement. We investigated the impact of two risk factors, prior victimization and mental health symptoms, using official law enforcement records for both victimization and recid- ivism.

Approximately one in six of the juveniles in this sample was the victim of a reported crime, either abuse/neglect, sexual assault, or the victim of property or person crime offenses. Although this is probably an underestimation, as studies that use self-report meth- ods have found higher rates of self-reported victimization (e.g., Abram et al., 2004), there may be something unique about reported incidents as compared to those that go unreported—a pattern that likely differs by victimization type. For instance, one explanation may be that reported incidents of property or person crimes could

be more serious because they came to the attention of law enforce- ment, as opposed to less serious incidents of these types of crimes. On the other hand, it may be that more serious incidents of child maltreatment are less likely to be reported (Brown, Cohen, John- son, & Salzinger, 1998; Swahn et al., 2006). Reported incidents could, alternatively, be the result of youth or familial differences. For example, Swahn and colleagues (2006) compared self-reported child maltreatment to official court records in a sample of youth in detention and found that of those who self-reported maltreatment, females were more likely to have an official court record than males, and African Americans were more likely to have an official court record than Whites.

Overall, the most frequent reported type of victimization was person crimes, followed by abuse/neglect and property offenses, with fewer reported incidents of sexual assault. We attempted to compare the rates of victimization in this sample to other studies, however because victimization has been operationalized in multi- ple ways, comparisons were not meaningful. Future research should compare whether early system-involved youth experience similar rates of victimization using official records and varying types of victimization to adjudicated and/or detained youth. In examining gender differences, the only type of victimization that differed by gender was sexual assault, in which females were more likely to experience sexual assault than males. Results from pre- vious research on gender differences in victimization experiences have been mixed. For instance, although Abram and colleagues (2004) found that males were more likely to report a traumatic event (defined broadly by witnessing or experiencing violence) than females, Conrad and colleagues (2014) found no differences by gender for a history of child sexual abuse. These data also suggest there may be some differences based on race/ethnicity for reported incidents of victimization. Non-White youth were signif- icantly more likely to have reported incidents of both abuse/ neglect and person crimes, but White youth were more likely to have reported incidents for property offenses. Because our victim- ization data are limited by official records, it is not clear whether non-White youth actually have higher rates of victimization, or whether these youths are just more likely to have system involve- ment following victimization.

In both the bivariate comparisons and multivariate model ex- amining time to recidivism, abuse/neglect and person crime vic- timization were related to both discharge from diversion and recidivism (property offense victimization was only related to recidivism). It is unclear from this data whether victimization in these types of incidents directly relates to difficulties in completing diversion or future reoffending, or whether these rates are an artifact of exposure to system involvement because of victimiza- tion. Moreover, the relationship between being the victim of a person or property crime and recidivism, could be the product of delinquent peers or engaging in delinquent lifestyles. Within this literature, scholars have proposed several avenues for this connec- tion. The first is related to Routine Activities Theory, which suggests that engaging in criminal behavior puts individuals at higher risk of victimization because they find themselves in situ- ations where they may be more likely to be victimized (Cohen & Felson, 1979). Others have proposed that exposure to victimization serves as a learned experience within certain subcultural environ- ments (Fagan, Piper, & Cheng, 1987), whereas general strain theory (Agnew, 1992) asserts that the negative experiences of

Table 4 Results From a Hazard Cox Regression Predicting Recidivism

Model Variable Exp(B) 95% CI

1 Age 1.04 .99–1.09 Race 1.40��� 1.20–1.61 Gender .71��� .61–.82

2 Age 1.07� 1.02–1.12 Race 1.20� 1.04–1.40 Gender .70��� .61–.82 Modified risk score 1.06��� 1.04–1.08 Discharge 1.37�� 1.14–1.63

3 Age 1.06� 1.01–1.12 Race 1.18� 1.02–1.37 Gender .71��� .61–.82 Modified risk score 1.06��� 1.04–1.08 Discharge 1.33�� 1.11–1.59 Sexual assault .84 .47–1.49 Abuse/neglect 1.37� 1.07–1.75 Person crime 1.40�� 1.14–1.73 Property crime 1.53� 1.07–2.20

4 Age 1.05 .99–1.10 Race 1.21� 1.04–1.41 Gender .75��� .64–.88 Modified risk score 1.06��� 1.04–1.09 Discharge 1.31�� 1.09–1.57 Sexual assault .84 .47–1.50 Abuse/neglect 1.37� 1.07–1.75 Person crime 1.41�� 1.14–1.74 Property crime 1.52� 1.06–2.19 Anxiety cluster .92 .77–1.10 Mood cluster .84 .68–1.03 Disruptive cluster 1.00 .85–1.17 Substance use cluster 1.12� 1.02–1.43

Note. Gender: male � 0; race: White � 0; victimization: nonvictim � 0; mental health clusters: 0 � not endorsed. The modified risk score included the YLS-total score with the substance use subscale values removed from the total score to reduce multicollinearity with the substance use mental health cluster. Exp(B) � exponentiation of the B coefficient. � p � .05. �� p � .01. ��� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

565VICTIMIZATION AND MENTAL HEALTH ON RECIDIVISM

victimization create strain, which contribute to delinquency (Ma- nasse & Ganem, 2009).

Our results demonstrate that two thirds of these early system- involved juveniles experience one or more mental health symp- toms, with one in five reporting substance use or mood symp- toms, one in three reporting anxiety symptoms, and almost half reporting disruptive disorder symptoms. In comparing these proportions to research using the same mental health assess- ment tool involving juveniles referred to probation (McReyn- olds et al., 2007, 2010) and incarcerated juveniles (Wasserman et al., 2002), it appears that early system-involved juveniles may experience similar mental health symptoms as deeper end justice-involved juveniles. In examining the symptom clusters specifically, a higher proportion of our sample endorsed anxi-

ety, mood, and disruptive symptoms, but fewer endorsed sub- stance use symptoms when compared to McReynolds et al. (2010) and Wasserman et al. (2002). Even though this sample differed from these previous studies in specific proportions, across all studies, disruptive disorders are most common among justice-involved youth and mood disorders are least common (Shufelt & Cocozza, 2006). As such, there may be few differ- ences between early justice-involved juveniles and later justice- involved juveniles in terms of how mental health symptoms are presented.

Furthermore, the results of the present study show the impact of victimization and mental health on early justice-involved youth. In support of the first hypothesis, juveniles with a history of victim- ization recidivated sooner than juveniles without a history of

Figure 1. Hazard function for each victimization type.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

566 WYLIE AND RUFINO

victimization (Cottle et al., 2001; Vidal et al., 2017). Although previous literature on mental health and recidivism in juvenile justice is mixed, the bivariate analyses and multivariate analyses demonstrated that participants endorsing substance use were more likely to recidivate, which is supported by previous research (Wolff et al., 2015; Wierson & Forehand, 1995). The bivariate comparisons also found that those who endorsed the mood disor- der symptoms were less likely to recidivate, which differs from studies that have not found a significant relationship between mood disorders and recidivism (McReynolds et al., 2010). One possibility is the mood cluster acts as a protective factor against recidivism, as juveniles who are feeling depressed are less likely to engage in social activities with peers, where they would be more likely to encounter peer pressure or engage in risky shift behaviors. Furthermore, juveniles who are feeling depressed may be experi- encing anhedonia and a lack of energy.

As partially hypothesized, however, once the mental health clusters were included in the multivariate model with demographic variables, the modified risk level score, and victimization, only the substance use mental health cluster predicted time to recidivism, while the mood mental health clusters did not. Although substance use remained a significant predictor of time to recidivism even while controlling for criminogenic risk factors, youth in this sam- ple with mental health issues did exhibit higher criminogenic risk and needs profiles as measured with the YLS/CMI compared with those without mental health issues as previous research has indi- cated (Guebert & Olver, 2014; Schubert et al., 2011). One caveat, however, is that the risk/needs profile may not be independent of mental health symptoms because there is overlap between items on the YLS/CMI and mental health issues. Although we were able to partially address this by removing the Substance Use subscale from the YLS/CMI and creating a modified YLS/CMI risk score without substance use, there are additional items on the YLS/CMI that have overlap with mental health symptoms that we were not able to separate. For instance, the Personality/Behavior YLS/CMI subscale measures short attention span, poor frustration tolerance, and tantrums, which may overlap with symptoms of disruptive disorders. The Leisure/Recreation YLS/CMI subscale measures personal interests, which could be affected by depression symp- toms. Future research should continue to explore these domains and examine measurement tools that can more accurately parse out these risk factors for delinquency.

The results of the current study demonstrate the importance of addressing mental health concerns in juvenile diversion programs, which are one of the first points that youth may touch in the juvenile justice system. While research guided by the Risk-Needs- Responsivity model has generally found that only treating mental health issues is mostly ineffective (e.g., Andrews & Bonta, 2010), research has demonstrated that addressing mental health issues and criminogenic needs can be effective (Ashford, Wong, & Stern- bach, 2008) because it adheres to the responsivity principle. The data from this study demonstrate there is a relationship between mental health needs and discharge reason—namely, that youth endorsing substance use and disruptive disorders were less likely to successfully complete the diversion program. Perhaps by spe- cifically attending to these mental health needs, diversion pro- grams can adapt programming for youth with these issues, which may contribute to more positive outcomes such as successful

completion of the program and reduced recidivism (Andrews & Bonta, 2010).

Limitations and Future Directions

This study has several limitations that should be considered when interpreting the results. The present sample was comprised entirely of early system-involved juvenile offenders who were deemed eligible for diversion and who chose to be in diversion rather than go through traditional juvenile court. As such, the sample may be limited to those with eligible offenses (minor offenses or status offenses) but would not include more serious first-time offenders. There may also be differences between the types of youth who choose diversion and those who choose tradi- tional juvenile court, including youth in more need of services (who may not otherwise have access to services) or youth whose families choose traditional court because it is often less time consuming than committing to a juvenile diversion plan. Although this sample was diverse, with approximately half identifying as White, one third as Black, and about 14% as Hispanic, this diversity does not necessarily represent the population of the city from which the data were collected, with Black youths overrepre- sented in this sample. This overrepresentation may be due to any number of related variables, including disproportionate minority contact that could impact the variables of interest and recidivism outcomes. Although the sample did include a small number of other races and ethnicities (e.g., Asian, Native Americans), these groups were too small to make meaningful comparisons. Future research may consider a more ethnically diverse sample.

Although previous research has consistently found various types of victimization do predict recidivism, better understanding how the types of victimization differ by sample characteristics and how different types of victimization impact recidivism warrants further investigation. The victimization variable was limited by using official law enforcement reports, which was likely an underesti- mation of victimization incidents in general. Official reports of victimization may also result in an underestimation of certain types of victimization that go unreported to law enforcement (e.g., child abuse/neglect, mutual assault). Future research may consider uti- lizing a range of victimization measures, including self-report data and official law enforcement records. Finally, although we know that victimization and mental health symptoms are related to recidivism, at least in bivariate comparisons, our data does not explain why these risk factors contribute to recidivism. Research- ers may consider specifically testing theoretical frameworks aimed at measuring for the underlying relationships between victimiza- tion, mental illness, and reoffense.

Conclusion

Juvenile diversion programs embrace the mission of the juvenile justice system, namely that youth can be rehabilitated by linking them to services and that juveniles who commit less serious offenses should not be formally processed through “the system.” This study provides preliminary evidence for some of the risk factors of juveniles most likely to recidivate after being connected to those services and diverted from the system. The three victim- ization type variables were the strongest predictors of recidivism, even while controlling for risk level and other juvenile character-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

567VICTIMIZATION AND MENTAL HEALTH ON RECIDIVISM

istics. As such, programs should specifically focus on trauma- informed programming that addresses the form of victimization a youth has experienced. Moreover, there is a need for early inter- vention tailored to the needs of abused, neglected, and victimized youths before they interact with the justice system or when they first enter the juvenile justice system. Early interventions should provide victimized youth with resources to increase resilience and teach positive and proactive coping strategies to minimize the effects that victimization may have on mental health functioning, and subsequent justice involvement.

References

Abram, K. M., Teplin, L. A., Charles, D. R., Longworth, S. L., McClelland, G. M., & Dulcan, M. K. (2004). Posttraumatic stress disorder and trauma in youth in juvenile detention. Archives of General Psychiatry, 61, 403– 410. http://dx.doi.org/10.1001/archpsyc.61.4.403

Agnew, R. (1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47– 88. http://dx.doi.org/10.1111/j.1745- 9125.1992.tb01093.x

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 Medi- cine, 39, 93–98. http://dx.doi.org/10.1016/j.amepre.2010.03.015

Andrews, D. A., & Bonta, J. (2010). Rehabilitating criminal justice policy and practice. Psychology, Public Policy, and Law, 16, 39 –55. http://dx .doi.org/10.1037/a0018362

Ashford, J. B., Wong, K. W., & Sternbach, K. O. (2008). Generic correc- tional programming for mentally ill offenders: A pilot study. Criminal Justice and Behavior, 35, 457– 473. http://dx.doi.org/10.1177/009 3854807313356

Baglivio, M. T., Epps, N., Swartz, K., Huq, M. S., Sheer, A., & Hardt, N. S. (2014). The prevalence of adverse childhood experiences (ACE) in the lives of juvenile offenders. Journal of Juvenile Justice, 3, 1. Retrieved from https://nicic.gov/prevalence-adverse-childhood-experiences-ace- lives-juvenile-offenders

Baglivio, M. T., Wolff, K. T., Piquero, A. R., Bilchik, S., Jackowski, K., Greenwald, M. A., & Epps, N. (2016). Maltreatment, child welfare, and recidivism in a sample of deep-end crossover youth. Journal of Youth and Adolescence, 45, 625– 654. http://dx.doi.org/10.1007/s10964-015- 0407-9

Barrett, D. E., Katsiyannis, A., Zhang, D., & Zhang, D. (2014). Delin- quency 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. http://dx.doi .org/10.1177/1063426612470514

Brown, J., Cohen, P., Johnson, J. G., & Salzinger, S. (1998). A longitudinal analysis of risk factors for child maltreatment: Findings of a 17-year prospective study of officially recorded and self-reported child abuse and neglect. Child Abuse & Neglect: The International Journal, 22, 1065–1078. http://dx.doi.org/10.1016/S0145-2134(98)00087-8

Chung, C. F., Schmidt, P., & Witte, A. D. (1991). Survival analysis: A survey. Journal of Quantitative Criminology, 7, 59 –98. http://dx.doi .org/10.1007/BF01083132

Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44, 588 – 608. http://dx.doi.org/10.2307/2094589

Conrad, S. M., Tolou-Shams, M., Rizzo, C. J., Placella, N., & Brown, L. K. (2014). Gender differences in recidivism rates for juvenile justice youth: The impact of sexual abuse. Law and Human Behavior, 38, 305–314. http://dx.doi.org/10.1037/lhb0000062

Cottle, C. C., Lee, R. J., & Heilbrun, K. (2001). The prediction of criminal recidivism in juveniles: A meta-analysis. Criminal Justice and Behavior, 28, 367–394. http://dx.doi.org/10.1177/0093854801028003005

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, 20274. http://dx.doi.org/10.3402/ejpt .v4i0.20274

English, D. J., Widom, C. S., & Brandford, C. (2002). Childhood victim- ization and delinquency, adult criminality, and violent criminal behav- ior: A replication and extension, final report. Retrieved from https:// www.ncjrs.gov/App/Publications/abstract.aspx?ID�192291

Fagan, J., Piper, E. S., & Cheng, Y. (1987). Contributions of victimization to delinquency in inner cities. The Journal of Criminal Law & Crimi- nology, 78, 586 – 613. http://dx.doi.org/10.2307/1143570

Fazel, S., Doll, H., & Långström, N. (2008). Mental disorders among adolescents in juvenile detention and correctional facilities: A systematic review and metaregression analysis of 25 surveys. Journal of the Amer- ican Academy of Child & Adolescent Psychiatry, 47, 1010 –1019. http:// dx.doi.org/10.1097/CHI.ObO13e31817eecf3

Finkelhor, D., Ormrod, R. K., & Turner, H. A. (2009). Lifetime assessment of poly-victimization in a national sample of children and youth. Child Abuse & Neglect, 33, 403– 411. http://dx.doi.org/10.1016/j.chiabu.2008 .09.012

Ford, J. D., Grasso, D. J., Hawke, J., & Chapman, J. F. (2013). Poly- victimization among juvenile justice-involved youths. Child Abuse & Neglect: The International Journal, 37, 788 – 800. http://dx.doi.org/10 .1016/j.chiabu.2013.01.005

Ford, J. D., Hartman, J. K., Hawke, J., & Chapman, J. F. (2008). Traumatic victimization, posttraumatic stress disorder, suicidal ideation, and sub- stance abuse risk among juvenile justice-involved youth. Journal of Child & Adolescent Trauma, 1, 75–92. http://dx.doi.org/10.1080/193 61520801934456

Guebert, A. F., & Olver, M. E. (2014). An examination of criminogenic needs, mental health concerns, and recidivism in a sample of violent young offenders: Implications for risk, need, and responsivity. Interna- tional Journal of Forensic Mental Health, 13, 295–310. http://dx.doi .org/10.1080/14999013.2014.955220

Hay, C., & Evans, M. M. (2006). Violent victimization and involvement in delinquency: Examining predictions from general strain theory. Journal of Criminal Justice, 34, 261–274. http://dx.doi.org/10.1016/j.jcrimjus .2006.03.005

Hawkins, J. D., Herrenkohl, T. I., Farrington, D. P., Brewer, D., Catalano, R. F., Harachi, T. W., & Cothern, L. (2000). Predictors of youth violence. Juvenile Justice Bulletin. Department of Justice, Washington, DC: Office of Juvenile Justice and Delinquency Prevention.

Hoge, R. D., & Andrews, D. A. (2011). Youth Level of Service/Case Management Inventory 2.0 (YLS/CMI 2.0): User’s manual. Toronto, Ontario, Canada: Multi-Health Systems. https://www.mhs.com/MHS- Publicsafety?prodname�yls-cmi2

Kilpatrick, D. G., Acierno, R., Saunders, B., Resnick, H. S., Best, C. L., & Schnurr, P. P. (2000). Risk factors for adolescent substance abuse and dependence: Data from a national sample. Journal of Consulting and Clinical Psychology, 68, 19 –30. http://dx.doi.org/10.1037/0022-006X .68.1.19

Kilpatrick, D. G., Ruggiero, K. J., Acierno, R., Saunders, B. E., Resnick, H. S., & Best, C. L. (2003). Violence and risk of PTSD, major depres- sion, substance abuse/dependence, and comorbidity: Results from the National Survey of Adolescents. Journal of Consulting and Clinical Psychology, 71, 692–700. http://dx.doi.org/10.1037/0022-006X.71.4 .692

Loftin, C., & McDowall, D. (2010). The use of official records to measure crime and delinquency. Journal of Quantitative Criminology, 26, 527– 532. http://dx.doi.org/10.1007/s10940-010-9120-8

Lucas, C. P., Zhang, H., Fisher, P. W., Shaffer, D., Regier, D. A., Narrow, W. E., . . . Friman, P. (2001). The DISC Predictive Scales (DPS): Efficiently screening for diagnoses. Journal of the American Academy of

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

568 WYLIE AND RUFINO

Child & Adolescent Psychiatry, 40, 443– 449. http://dx.doi.org/10.1097/ 00004583-200104000-00013

Maltz, M. D. (1984). Recidivism. Orlando, FL: Academic Press. Manasse, M. E., & Ganem, N. M. (2009) Victimization as a cause of

delinquency: The role of depression and gender. Journal of Criminal Justice, 37, 371–378. http://dx.doi.org/10.1016/j.jcrimjus.2009.06.004

McReynolds, L. S., Schwalbe, C. S., & Wasserman, G. A. (2010). The contribution of psychiatric disorder to juvenile recidivism. Criminal Justice and Behavior, 37, 204 –216. http://dx.doi.org/10.1177/009 3854809354961

McReynolds, L. S., Wasserman, G. A., Fisher, P., & Lucas, C. P. (2007). Diagnostic screening with incarcerated youths: Comparing the DPS and Voice DISC. Criminal Justice and Behavior, 34, 830 – 845. http://dx.doi .org/10.1177/0093854807299918

Myner, J., Santman, J., Cappelletty, G. G., & Perlmutter, B. F. (1998). Variables related to recidivism among juvenile offenders. International Journal of Offender Therapy and Comparative Criminology, 42, 65– 80. http://dx.doi.org/10.1177/0306624X98421006

OJJDP Statistical Briefing Book. (2016). Juvenile enforcement & juvenile crime: Juvenile arrests. Retrieved from https://www.ojjdp.gov/ojstatbb/ crime/qa05101.asp?qaDate�2016

Onifade, E., Davidson, W., Campbell, C., Turke, G., Malinowski, J., & Turner, K. (2008). Predicting recidivism in probationers with the youth level of service case management inventory (YLS/CMI). Criminal Jus- tice and Behavior, 35, 474 – 483. http://dx.doi.org/10.1177/0093 854807313427

Patterson, G. R., DeBaryshe, B. D., & Ramsey, E. (1989). A developmental perspective on antisocial behavior. American Psychologist, 44, 329 – 335. http://dx.doi.org/10.1037/0003-066X.44.2.329

Schmidt, P., & Witte, A. D. (1989). Predicting criminal recidivism using ‘split population’ survival time models. Journal of Econometrics, 40, 141–159. http://dx.doi.org/10.1016/0304-4076(89)90034-1

Schubert, C. A., Mulvey, E. P., & Glasheen, C. (2011). Influence of mental health and substance use problems and criminogenic risk on outcomes in serious juvenile offenders. Journal of the American Academy of Child & Adolescent Psychiatry, 50, 925–937. http://dx.doi.org/10.1016/j.jaac .2011.06.006

Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab-Stone, M. E. (2000). NIMH Diagnostic Interview Schedule for Children Version IV (NIMH DISC-IV): Description, differences from previous versions, and reliability of some common diagnoses. Journal of the American Acad- emy of Child & Adolescent Psychiatry, 39, 28 –38. http://dx.doi.org/10 .1097/00004583-200001000-00014

Shufelt, J. L., & Cocozza, J. J. (2006). Youth with mental health disorders in the juvenile justice system: Results from a multi-state prevalence study. National Center for Mental Health and Juvenile Justice. Re- trieved from https://www.ncmhjj.com/resources/youth-mental-health- disorders-juvenile-justice-system-results-multi-state-prevalence-study/

Smith, C. A., Ireland, T. O., & Thornberry, T. P. (2005). Adolescent maltreatment and its impact on young adult antisocial behavior. Child Abuse & Neglect: The International Journal, 29, 1099 –1119. http://dx .doi.org/10.1016/j.chiabu.2005.02.011

Smith, C., & Thornberry, T. P. (1995). The relationship between childhood maltreatment and adolescent involvement in delinquency. Criminology, 33, 451– 481. http://dx.doi.org/10.1111/j.1745-9125.1995.tb01186.x

Swahn, M. H., Whitaker, D. J., Pippen, C. B., Leeb, R. T., Teplin, L. A., Abram, K. M., & McClelland, G. M. (2006). Concordance between self-reported maltreatment and court records of abuse or neglect among high-risk youths. American Journal of Public Health, 96, 1849 –1853. http://dx.doi.org/10.2105/AJPH.2004.058230

Teplin, L. A., Abram, K. M., McClelland, G. M., Dulcan, M. K., & Mericle, A. A. (2002). Psychiatric disorders in youth in juvenile deten- tion. Archives of General Psychiatry, 59, 1133–1143. http://dx.doi.org/ 10.1001/archpsyc.59.12.1133

U.S. Department of Justice. (2016). Crimes against person, property and society. Retrieved from https://ucr.fbi.gov/nibrs/2015/resource-pages/ crimes_against_persons_property_and_society-2015_final.pdf

Vermeiren, R., Schwab-Stone, M., Ruchkin, V., De Clippele, A., & De- boutte, D. (2002). Predicting recidivism in delinquent adolescents from psychological and psychiatric assessment. Comprehensive Psychiatry, 43, 142–149. http://dx.doi.org/10.1053/comp.2002.30809

Vidal, S., Prince, D., Connell, C. M., Caron, C. M., Kaufman, J. S., & Tebes, J. K. (2017). Maltreatment, family environment, and social risk factors: Determinants of the child welfare to juvenile justice transition among maltreated children and adolescents. Child Abuse & Neglect: The International Journal, 63, 7–18. http://dx.doi.org/10.1016/j.chiabu.2016 .11.013

Vincent, G. M., Guy, L. S., & Grisso, T. (2012). Risk assessment in juvenile justice: A guidebook for implementation. Retrieved from http:// www.nysap.us/Risk%20Guidebook.pdf

Wasserman, G. A., McReynolds, L. S., Lucas, C. P., Fisher, P., & Santos, L. (2002). The voice DISC-IV with incarcerated male youths: Preva- lence of disorder. Journal of the American Academy of Child & Ado- lescent Psychiatry, 41, 314 –321. http://dx.doi.org/10.1097/00004583-20 0203000-00011

Widom, C. S. (1989). Does violence beget violence? A critical examination of the literature. Psychological Bulletin, 106, 3–28. http://dx.doi.org/10 .1037/0033-2909.106.1.3

Wierson, M., & Forehand, R. (1995). Predicting recidivism in juvenile delinquents: The role of mental health diagnoses and the qualification of conclusions by race. Behaviour Research and Therapy, 33, 63– 67. http://dx.doi.org/10.1016/0005-7967(94)E0001-Y

Wolff, K. T., Baglivio, M. T., & Piquero, A. R. (2015). The relationship between adverse childhood experiences and recidivism in a sample of juvenile offenders in community-based treatment. International Journal of Offender Therapy and Comparative Criminology, 61, 1210 –1242. http://dx.doi.org/10.1177/0306624X15613992

Zingraff, M. T., Leiter, J., Myers, K. A., & Johnson, M. C. (1993). Child maltreatment and youthful problem behavior. Criminology, 31, 173– 202. http://dx.doi.org/10.1111/j.1745-9125.1993.tb01127.x

Received February 9, 2018 Revision received September 24, 2018

Accepted September 26, 2018 �

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

569VICTIMIZATION AND MENTAL HEALTH ON RECIDIVISM