Volume 3, Issue 1 • Fall 2013

Table of Contents

Forword

The Impact of Juvenile Mental Health Court on Recidivism Among Youth

Gender-Specific Mental Health Outcomes of a Community-Based Delinquency Intervention

Predicting Recidivism Among Juvenile Delinquents: Comparison of Risk Factors for Male and Female Offenders

Building Connections Between Officers and Baltimore City Youth: Key Components of a Police–Youth Teambuilding Program

Internet-Based Mindfulness Meditation and Self-regulation: A Randomized Trial with Juvenile Justice Involved Youth

Assessing Youth Early in the Juvenile Justice System

A Jury of Your Peers: Recidivism Among Teen Court Participants

Commentary: Place-Based Delinquency Prevention: Issues and Recommendations

Predicting Recidivism Among Juvenile Delinquents: Comparison of Risk Factors for Male and Female Offenders

Kristin C. Thompson, School Psychology Program, University of Arizona

Richard J. Morris, School Psychology Program, University of Arizona.

Acknowledgment: Preparation of this article was supported by the University of Arizona’s Jacqueline Anne Morris Memorial Foundation’s Children’s Policy and Research Project.

Correspondence concerning this article should be addressed to: Kristin C. Thompson, University of Arizona–School Psychology, 1430 East 2nd Street, PO Box 210069, Tucson, Arizona 85721; E-mail: thomkr01@email.arizona.edu

Keywords: gender, recidivism, education, juvenile delinquency

Abstract

Research examining risk factors for delinquency and risk factors that predict recidivism among delinquents has produced inconsistent results, due, in part, to the various methodologies and samples used in studies. The majority of studies have used all-male samples or have been limited to groups of offenders who have committed only minor offenses on the one hand, or severe offenses on the other. As the number of female offenders increases, more research is needed that controls for gender and methodology in an effort to clarify risk factors for both delinquency and recidivism among male and female juveniles.

This study examined risk factors for recidivism related to education, demographics, and offense patterns in a diverse sample of 3,287 male and female juvenile delinquents from Arizona. The study sought to determine whether differences existed between male and female offenders in regard to risk factors for recidivism, and to identify those that were predictive of recidivism in male versus female delinquents. Overall, this study found significant differences between risk factors, and that male and female delinquents differed with respect to which risk factors were predictive of recidivism. Academic achievement was not predictive of recidivism among females in this study, and contrary to the results of other studies, we found offense severity was not related to recidivism for either sex. Despite its relatively low frequency in the sample, we found emotional disabilities were predictive of recidivism for both sexes. Additional analyses found that juvenile delinquents with an emotional disability were at significantly greater risk for recidivism than were delinquents without an emotional disability.

Introduction

A major concern in the area of juvenile delinquency is the repeated arrest and incarceration of juveniles. Although the actual arrest rate of juveniles has declined over the past decade, recidivism has remained high and stable, with estimates of reoffending among juveniles ranging from 30% to 90% (e.g., McMackin, Tansi, & LaFratta, 2004; Trulson, Marquart, Mullings, & Caeti, 2005; van der Geest, 2008). This concern has led a number of studies to address risk factors related to recidivism among juvenile offenders. For example, in regard to offense patterns, studies consistently report that the earlier juveniles begin to commit crimes, the greater the likelihood that they will continue to reoffend (e.g., Barrett, Katsiyannis, & Zhang, 2010; Cottle, Lee, & Heilbrun, 2001; Trulson et al., 2005). In addition, studies have found that delinquents who commit crimes of greater severity are at an increased risk for reoffending (e.g., Cottle et al., 2001; Dembo et al., 1998; Myner, Santman, Cappelletty, & Perlmutter, 1998).

Researchers have also linked academic achievement with recidivism (e.g., Katsiyannis, Ryan, Zhang, & Spann, 2008). For example, Archwamety and Katsiyannis (2000) studied juvenile delinquents in remedial math and reading groups and found that they were twice as likely to recidivate as those in the control group who were not in need of remedial academic instruction. A literature review by Vacca (2008) that focused on reading achievement and delinquency suggested that recidivism would decrease if more time were spent teaching delinquents to read.

Directly related to academic achievement, a limited number of studies have examined the relationships among delinquency, disability, and recidivism, as there is an overrepresentation of juveniles with such disabilities in the juvenile justice system. In fact, research has suggested that between 30% and 100% of delinquents have a disability as categorized under the Individuals with Disabilities Education Improvement Act (IDEIA) (IDEIA, 2004; Morris & Morris, 2006; Quinn, Rutherford, Leone, Osher, & Poirer, 2005), with delinquents having emotional disabilities (ED) being overrepresented. Some studies specifically examining the relationship between disability and reoffending suggest that juveniles with disabilities may be particularly vulnerable to recidivism (Barrett, et al., 2010; Zhang, Barrett, Katsiyannis, & Yoon, 2011; Zhang, Hsu, Katsiyannis, Barrett, & Ju, 2011), although research in this area is limited.

Inconsistent findings plague delinquency research, particularly when examining risk factors for recidivism. Qualitatively, a review of the literature shows nearly as many studies supporting various factors as being predictive of recidivism as studies failing to find any relationship. Specifically, although several studies have found academic achievement, disability, ethnicity, socioeconomic status, conduct problems, and offense patterns to be predictive of recidivism, several other studies have not (e.g., Calley, 2012; Cottle et al., 2001; Dembo et al., 1998; Duncan, Kennedy, & Patrick, 1995; Mulder, Vermunt, Brand, Bullens, & Marle, 2012; Myner et al.,1998; Tille & Rose, 2007). These inconsistencies are due, in part, to the various methodologies and samples used in studies, as the majority of studies utilize all-male samples or are limited to groups of juveniles who have committed either relatively minor or relatively severe offenses.

The differing ways in which studies define variables can also affect their results. For example, few studies examine the influence of a specific disability (e.g., emotional disability versus learning disability) on recidivism, instead using a generic category of “special education placement” despite supporting evidence that mental health issues (associated with an emotional disability) may be correlated with recidivism.

The majority of studies also utilize samples of primarily male delinquents, or combine male and female delinquents into one sample, despite available evidence suggesting that male and female adolescents may differ with regard to characteristics of delinquency and risk factors for recidivism (e.g., Tille & Rose, 2007; Trulson et al., 2005; Vitopoulos, Peterson-Badali, & Skilling, 2012). A study by Funk (1999), for example, found that risk factors of recidivism for males were different than risk factors of recidivism for females. A recent study by Steketee, Junger, and Junger-Tas (2013) examined male and female delinquents in 30 countries and concluded that there are significant differences in risk factors of recidivism for females versus males. These researchers asserted that traditional theories of delinquency do not apply to females as they do to males. It is notable, however, that this study (and many other studies) relied on self-reported delinquency versus an actual arrest history.

The results of the studies mentioned above highlight the importance of gender-responsive delinquency research, given that it is scholarly research that serves as the premise for many risk-assessment instruments and intervention planning for delinquent youth. Consequently, the purpose of the present study was twofold: first, to examine whether differences exist between female and male juvenile offenders with respect to educational and offense variables that research suggests may be predictive of recidivism; and second, to determine which are the best predictors of recidivism among male and female delinquents.

This study contained educational, demographic, and offense data for a large and diverse sample of youth arrested at least once over the 5 years between August 2006 and May 2011. The sample included male and female delinquents, and included youth with various offense histories. Youth in the sample had been arrested any-where from 1 to 54 times, and had committed offenses ranging from relatively minor status offenses and misdemeanor offenses, to more severe felony offenses. The educational variables examined include diagnosis of an emotional disability; diagnosis of a learning disability; and standardized academic achievement test scores in reading, writing, and math. Offense variables examined included their type and severity, adjudication status, and total days in detention. Demographic variables included in the analyses were socioeconomic status, dual involvement in the court system because of either delinquency or child welfare issues, and ethnicity. We formulated the following hypotheses: (a) based on the limited research suggesting that female delinquents differ considerably from their male counterparts in terms of risk factors for delinquency, differences would be observed between male and female delinquents in regard to risk factors for recidivism; and (b) variables included in the analysis would significantly predict recidivism among male and female delinquents.

Method

Participants

This study included 3,287 youth (2,134 males and 1,153 females) between the ages of 8 and 17 years. These juveniles were enrolled in a large public school district in Arizona and had been arrested at least once between August 2006 and May 2011. The most current data were used for participants who had been arrested multiple times during this time period. This project was approved by the University of Arizona’s Institutional Review Board, the participating school district, and the participating juvenile court center.

Variables

The independent variables were as follows:

  1. Presence of an emotional disability, as defined by the IDEIA (2004).
  2. Presence of a learning disability, as defined by IDEIA (2004). The diagnoses of emotional and learning disabilities are two of the most common IDEIA diagnoses found among delinquents.
  3. Presence of a speech or language impairment, as defined by IDEIA (2004). This variable was included because it was the second most prevalent IDEA diagnosis in the sample of participants.
  4. Ethnicity.
  5. Socioeconomic status, determined by students’ participation in the school lunch program. This program makes students eligible to receive free or reduced lunch based on the income level of their parents or guardians. Those receiving no free lunch or reduced lunch fall above a predetermined income; those receiving free lunch fall into the lowest income category. The three categories for the variable of socioeconomic status were: no free or reduced lunch, reduced lunch, and free lunch.
  6. Dual involvement of youth in the court system, which included youth who were involved with the courts because of delinquency offenses, as well as child welfare issues.
  7. The participants’ Arizona Instrument to Measure Standards (AIMS) standardized achievement test scores in the areas of reading performance, writing performance, and math performance. Standardized achievement scores were determined from the scores available from the students’ most recent AIMS standardized test results in these areas. Reliability coefficients for the AIMS tests are above 0.90 for all grade levels (Arizona Department of Education, 2008).
  8. Adjudication status (guilty vs. not guilty) of participants arrested. Adjudication status was defined according to whether the participant was found guilty of at least one of the offenses for which he or she was arrested.
  9. Severity of offense. Offense severity was determined by the local juvenile court center in four categories: status offenses, obstruction, misdemeanor, and felony. Given that some youth are likely to have committed more than one offense during a particular school year, the most severe offense was used in all statistical analyses. Because status offense is the only category in which a crime is illegal solely because of the juvenile’s age, this was rated to be the least severe offense type. Obstruction, which consists of offenses such as violation of conditions of release or probation violations, was ranked second in terms of severity because violations are deemed criminal acts regardless of an individual’s age. Misdemeanors and felonies ranked third and fourth, respectively, in terms of severity.
  10. Type of offense committed. Type of offense was determined by the local juvenile court center and included a variety of categories, such as status offense, drug possession or sales, misdemeanor against property, misdemeanor against persons, felony against property, and felony against persons.

Recidivism served as the dependent variable and was defined as the total number of arrests present in a student’s lifetime record with the local juvenile court center, including probation violations.

Procedure

Data for this study were obtained through the University of Arizona Juvenile Delinquency Project (UAJDP) database. This extensive database, which we created for this project, consists of offense history and educational data for students 8 to 17 years of age who have been arrested in an Arizona county. New data are obtained yearly through an intergovernmental data-sharing agreement between an Arizona juvenile court center, an Arizona school district, and the University of Arizona. As of 2012, the UAJDP had amassed data for 8,997 youths. This database is comprehensive in that it contains information for all youth arrested in a large school district (n = 60,000) each year. Consequently, it represents male and female offenders, first-time and repeat offenders, minor status offenders, juveniles completing probation violations or misdemeanor offenses, and offenders who have committed serious felonies. Some youth in the database have been detained and/or incarcerated, while others received only paper arrests. All data went through the juvenile court center before being given to researchers at the University of Arizona. To maintain confidentiality, each student in the database was first assigned a random identification number by information technology staff at the juvenile court center, none of whom were affiliated with this study or informed about its specific purpose or hypotheses. The identification numbers of youth who had been arrested over multiple years of data collection were flagged for researchers to avoid duplication.

Demographics of Sample

The sample for this study consisted of 64.9% male delinquents and 35.1% female delinquents. These percentages are consistent with nationwide data on gender and delinquency. In regard to ethnicity, 54.0% of the study population was Hispanic, 27.9% Caucasian, 9.9% African American, 6.0% Native American, and 1.3% Asian American. The ethnic representation of the sample was consistent with that of the original population of youth in the school district. See Table 1 for a comparison of male and female juveniles on all variables.

Table 1. Sample Characteristicsa

  Males Females

Ethnicity

 

 

African American

9.7

10.5

Asian

1.1

1.6

Caucasian

28.0

27.7

Hispanic

54.6

53.0

Native American

5.8

6.2

Special Education

33.7

17.5

LD

18.5

9.2

ED

8.3

3.5

Speech-Language Impairment

10.3

6.0

Demographics

 

 

Free/Reduced Lunch

69.0

69.1

Dually Involved

10.9

11.3

Academic Achievementb

 

 

Reading

42.1

53.5

Math

50.4

57.5

Writing

42.8

66.5

Offense Severityc

 

 

Felony 

33.8

16.4

Misdemeanor

55.0

64.2

Obstruction

2.5

1.0

Status Offense

8.5

18.2

Adjudicated

12.7

4.3


a: Values represent percentages
b: Percentage with passing scores on state standardized achievement tests
c: Most severe offense reported

Data Analyses

We conducted chi-square analyses to determine whether there were differences among independent variables for male and female delinquents. We then used standard multiple regression analysis to determine which factors best predicted recidivism among male and female juvenile delinquents. These analyses allowed us to determine which independent variables best predicted recidivism for each sex, and whether risk factors differed between males and females.

Due to the categorical nature of ethnicity and offense type, we did not include these variables in the linear regression model. Instead, we conducted a one-way analysis of variance (ANOVA) to determine whether the number of referrals differed significantly among the various ethnic groups, as well as among the types of offenses committed.

Results

Hypothesis I: Comparison of Risk Factors for Male and Female Offenders

We conducted chi-square analyses to determine whether significant differences existed between males and females for the variables that were to be included in the prediction model for recidivism. While no significant associations were observed between males and females on ethnicity, socioeconomic status, or dual involvement in the juvenile court system, we observed differences between the sexes in all other areas. In regard to educational variables, more females than males were likely to have passed standardized achievement tests in the areas of reading χ2 (1, N = 3098) = 37.15, p < .001, phi = .11; writing χ2(1, N = 3287) = 168.81, p < .001, phi = .227; and math χ2(1, N = 3005) = 13.681, p < .001, phi = .067. It is noteworthy, however, that the effect size was small for both reading and math. In addition, a significantly larger proportion of males than females were enrolled in special education programs χ2(1, N = 3287) = 97.58, p < .001, phi = .172. In regard to specific disabilities, a smaller proportion of females than males were diagnosed with emotional disabilities χ2(1, N = 3287) = 28.26, p < .001, phi = -.093, but again, the effect size was small. Fewer females than males were observed with a diagnosis of a learning disability χ2(1, N = 3287) = 50.29, p < .001, phi = -.124. We observed similar results for speech and language impairments χ2(1, N = 3287) = 32.89, p < .001, phi = -.072.

We categorized offense patterns as follows: (a) 1 offense; (b) 2 to 5 offenses; and (c) 6 or more offenses. We observed significant differences in offense patterns between male and female delinquents χ2(2, N = 3280) = 19.52, p < .001, Cramer’s V = 0.78. Specifically, more females than males had only one offense, while a greater proportion of males than females had six or more offenses. We also observed differences in offense severity, χ2(3, N = 3279) = 160.61, p < .001, Cramer’s V = .221, with a greater proportion of females than males having status offenses, and significantly more males than females having felony offenses. Males were also more likely than females to have been adjudicated χ2(1, N = 3280) = 67.06, p < .001, phi = .143.

Hypothesis II: Predicting Recidivism Among Male and Female Juvenile Offenders

To test hypothesis II, we conducted standard regression analysis for both the male and female samples. For both male and female delinquents, initial analyses found that offense severity, learning disabilities, and a speech-language impairment had extremely low correlations (all <0.10) with the dependent variable (recidivism) and consequently were not included in the final regression analysis. The final variables included in the analysis were as follows: socioeconomic status, dual involvement, adjudication status, total time spent in detention, emotional disability, and standardized achievement in reading, writing, and math. Given the large sample size for both males and females, the number of cases per variable is well above the suggested number of cases needed to ensure a reliable equation in multiple regression (Tabachnick & Fidell, 2007).

Females

We conducted preliminary analyses to ensure no violation of the assumptions of normality, linearity, multicollinearity, or homoscedasticity. As previously mentioned, we removed three variables from the final analysis because of low correlations with the dependent variable. We found a significant model that explained 36% of the variance in recidivism among females, F(9, 1057) = 67.31, p <.001. Variables that significantly predicted recidivism among females included socioeconomic status, adjudication status, dual involvement, emotional disability, and total time in detention (see Table 2). We conducted a one-way ANOVA to determine whether ethnicity was significantly related to recidivism among females; results indicated that the number of referrals did not differ significantly from one ethnic group to another, F(4,1133) = 0.671; p > .05. We also conducted a one-way ANOVA among groups of females who had committed different types of offenses to determine the effect of offense type on recidivism. We found a significant difference among females for offense type, F(6, 1129) = 5.98, p < .001. However, despite reaching statistical significance, the actual difference in mean number of arrests among the groups of females by offense type was small. The effect size, calculated using eta squared, was 0.001. Consequently, the statistical significance was likely due to the large sample size and, therefore, no post hoc analyses are reported here because upon examination these results were clinically insignificant.

Table 2. Variables Predicting Recidivism in Adolescent Females (N =1,153 )

Variable β

Total time in detention

0.324

Dually Involved

0.281

ED

0.175

Adjudication Status

0.169

Socioeconomic Status

-0.110

Males

We conducted preliminary analyses to ensure no violation of the assumptions of normality, linearity, multicollinearity, or homoscedasticity. We found a significant model that explained 35% of the variance in recidivism among males, F(9, 1921) = 116.58, p <.001. Variables that significantly predicted recidivism among males included socioeconomic status, adjudication status, dual involvement, diagnosis of an emotional disability, total time in detention, writing achievement, and math achievement (see Table 3).

Table 3. Variables Predicting Recidivism in Adolescent Males (N = 2,134)

Variable β

Total time in detention

0.342

Adjudication Status

0.196

Dually Involved

0.187

ED

0.155

Socioeconomic Status

-0.149

Math

-0.110

Writing

-0.072

We conducted a one-way ANOVA to determine whether ethnicity was significantly related to recidivism; results indicated no significant differences among ethnic groups in regard to the number of arrests, F(4,2107) = 1.791; p > .05. We also conducted a one-way ANOVA between groups to explore the impact of offense type on recidivism. Results indicated a significant difference for offense type, F(6, 2075) = 2.97, p < .01. However, despite reaching statistical significance, the actual difference in mean scores among the groups by type of offense was small. The effect size, calculated using eta squared, was 0.01. Consequently, the statistical significant was likely due to the large sample size and, therefore, no post hoc analyses are reported here.

Emotional Disability

Although the presence of an emotional disability was relatively small in this sample, particularly for females, an emotional disability was a significant predictive factor for recidivism for both male and female delinquents. This is important given that no educational variables were predictive of recidivism among female delinquents, including standardized achievement scores in reading, writing, and math or the presence of a learning disability or speech-language impairment. Consequently, we conducted further analyses to examine differences between samples with and without emotional disabilities.

Among female delinquents, we observed significant associations between recidivism, and an emotional disability by ethnic group, with significantly more Caucasian and African American females being diagnosed with an emotional disability than would be expected, χ2(4, N = 1141) = 10.12, p < .05, Cramer’s V = .094. We observed the same pattern and associations among male delinquents, χ2(4, N = 2116) = 39.79, p < .001, Cramer’s V = .137.

Offense type also differed significantly with diagnosis of an emotional disability. Females with an emotional disability committed fewer drug offenses and misdemeanors against property than would be expected, and more misdemeanors against persons, χ2(9, N = 1150) = 20.43, p < .05, Cramer’s V = .133. We observed a similarly significant pattern for male delinquents with an emotional disability, who also had a substantially greater number of felonies against persons compared to males without an emotional disability, χ2(9, N = 2130) = 67.218, p < .05, Cramer’s V = .18. We conducted an independent sample t-test to determine whether the amount of time spent in detention among youth with and without emotional disabilities differed; results indicated a significant difference between males with an emotional disability (M = 2.31, SD = 7.76) and males without an emotional disability (M = 1.06, SD = 6.42); t(2128) = -2.41, p < .05. We found no significant differences in the amount of time spent in detention for females with and without an emotional disability.

Academically, both male and female delinquents with an emotional disability performed significantly more poorly on standardized measures of reading, writing, and math than their counterparts without an emotional disability. Specifically, for reading, although 43.9% of delinquents passed standardized tests, only 21.6% of male delinquents with an emotional disability passed, χ2(1, N = 2000 = 29.33 p < .001, Cramer’s V = .123. Among females, 54.4% of girls without an emotional disability passed, while 26.3% with an emotional disability passed, χ2(1, N = 1098) = 11.66, p < .001, Cramer’s V = .103. In regard to writing, only 17.5% of male youth with an emotional disability passed standardized tests, and 22.5% of girls with an emotional disability passed, χ2(1, N = 2134) = 49.23, p < .001, Cramer’s V = .15; χ2(1, N = 1153) = 34.04, p < .001, Cramer’s V = .17. For state standardized math tests, only 8.6% of male juveniles with an emotional disability passed, and 7.9% of female delinquents with an emotional disability passed; χ2(1, N = 1938) = 42.36, p < .001, Cramer’s V = .15; χ2(1, N = 1067) = 13.55, p < .001, Cramer’s V = .11.

Discussion

The purpose of this study was to determine whether male and female delinquents differ on educational and offense variables predictive of recidivism and, if so, to identify which risk factors are predictive of recidivism in both populations. Overall, this study found significant differences between risk factors, and that male and female delinquents differed with respect to which risk factors are predictive of recidivism.

Our results concur with those of other studies (e.g., Steketee et al., 2013), finding that females committed fewer and less severe offenses than males. Specifically, females in our study were more likely than males to commit status offenses. In fact, females had committed more than twice as many status offenses as males. Males, on the other hand, had committed more than twice as many felonies as females.

We observed differences in risk factors for recidivism among males and females primarily in the area of academic achievement. Specifically, females performed significantly better than their male counterparts (although they were still below average) in the areas of reading, writing, and math. Higher achievement in these areas was not predictive of recidivism among females, although low achievement in writing and math was predictive of recidivism among males. Interestingly, despite evidence suggesting that poor reading skills and learning disabilities are predictive of delinquency and recidivism (e.g., Barrett et al., 2010; Archwamety & Katsiyannis, 2000; Vacca, 2008), neither of these was significant in this study.

We found a diagnosis of an emotional disability to be a strong predictor of recidivism among both male and female delinquents. This is of particular importance considering the small incidence of emotional disability within the sample. This variable precluded many other variables in the regression model. In addition, given the lack of empirical research examining the relationship between emotional disability, special education status, and recidivism, further investigation regarding this relationship is needed. Learning disabilities were not found to be a significant predictor of recidivism, despite the fact that nearly 55% of the sample receiving special education services were diagnosed with a learning disability.

Additional analyses found that delinquents with emotional disabilities differed significantly from their non-emotionally disabled counterparts in several areas, including number of arrests, academic achievement, and types of offenses committed. Interestingly, we found no relationship between offense severity and recidivism, and we observed no differences in offense severity between youth with and without emotional disabilities—despite the fact that youth with an emotional disability committed significantly more offenses than youth without an emotional disability. This finding has direct implications for how these youth are being dealt with in both the school system and the juvenile justice system. Early intervention for youth with emotional disabilities, which may include school administrators and personnel finding alternative strategies for discipline, is critical. Finding ways of dealing with these children before turning them over to the juvenile justice system seems imperative, since they appear to not respond well to current procedures. Delinquents in this study who were diagnosed with an emotional disability spent significantly more time in detention than those without an emotional disability, even though the former group was not committing more severe offenses than the latter.

The discussions focusing on why delinquents with disabilities are overrepresented in the juvenile justice system are interminable. Yet, despite special education reforms and progress in providing services to students with disabilities within the general education system, more needs to be done within the juvenile justice system. Results of the present study have specific implications for the development of transition services for juveniles with disabilities, specifically those with emotional disabilities. Transition services are a required component of the services students are entitled to by the IDEIA (2004); however, the extent to which these services are provided in juvenile correctional facilities is minimal (Griller-Clark, & Mathur, 2010; Nelson, Jolivette, Leone, & Mathur, 2010). Given recent evidence suggesting that delinquents with disabilities who are provided transition services upon release are less likely to recidivate than those without access to such services (e.g., Griller-Clark, Mathur, & Helding, 2011), it is crucial that more be done to facilitate the development of prevention, intervention, and transition services for these youth.

As previously mentioned, contrary to the findings of other studies, offense severity was not significantly predictive of recidivism in this study, either for males or females. In fact, the correlation between severity and recidivism was so low that the variable was not retained in the final regression analysis. This finding may relate to the impact of sampling biases in other studies; a diverse, broad sample was a strength of our study, as participants were recruited from more than one setting. This implies that research may need to further examine the use of severity as a covariate when conducting analyses, particularly given that females are more likely to have status offenses than are males.

Limitations of the Study

Our study had several limitations. First, because this study used a snapshot of information, research based on a more thorough offense history is needed to determine whether offense history continues to be predictive of recidivism. Second, the present study did not include youth who committed crimes that required them to be transferred to adult court. Such youth, if included in a similar study, may have an effect on the results, since their offenses are likely to be more serious. Third, although the sample was representative of the local population, there were a higher percentage of both females and Hispanic youth in our study than is found in the general population of delinquents nationwide; consequently, our sample may not be representative of delinquent youth across the United States. In addition, this study did not address the varied emotional and familial characteristics that studies have indicated may be disproportionately related to female, compared with male, delinquency (e.g., Steketee et al., 2013; Tille & Rose, 2007).

This study highlights the need to treat male and female offenders as two distinct populations when conducting research, as each population is characterized by distinct offense and academic patterns. This separation of male and female delinquents in research samples is particularly important if the goal of the research is to develop effective prevention or intervention programs, as the treatments must be modified to the etiology of delinquency.

The age of first offense is an important variable to include in a regression model, since research consistently shows that juveniles who begin offending at a younger age are likely to commit a greater number of, and more serious, offenses than those who begin to offend later in their teenage years (e.g., Jones, Harris, Fader, & Grubstein, 2001; Trulson et al., 2005). It would also be important for future research to compare the influence of age at first offense to other variables, such as diagnosis of an emotional disability, which the current model found to be an important predictor of recidivism in this population.

About the Authors

Kristin C. Thompson, PhD, is an assistant professor of practice and teaches graduate courses in school psychology at the University of Arizona. A nationally certified school psychologist and licensed psychologist, Dr. Thompson also works with children and adolescents in private practice and has experience working in juvenile corrections.

Richard J. Morris, PhD, is a professor of school psychology at the University of Arizona. A fellow of the American Psychological Association, Dr. Morris has written or edited more than 13 books on child behavioral disorders and psychotherapy.

References

Archwamety, T., & Katsiyannis, A. (2000). Academic remediation, parole violation, and recidivism rates among delinquent youths. Remedial and Special Education, 21, 161–170. doi:10.1177/074193250002100306

Arizona Department of Education (2008). Arizona’s instrument to measure standards results. Retrieved January 17, 2009, from http://www.ade.state.az.us/researchpolicy/AIMSresults

Barrett, D. E., Katsiyannis, A., & Zhang, D. (2010). Predictors of offense severity, adjudication, incarceration and repeat referrals for juvenile offenders: A multi-cohort replication study. Remedial and Special Education, 31, 261–275. doi:10.1177/0741932509355990

Calley, N. G. (2012). Juvenile offender recidivism: An examination of risk factors. Journal of Child Sexual Abuse: Research, Treatment, & Program Innovations for Victims, Survivors, & Offenders, 21, 257–272.

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. doi:10.1177/0093854801028003005

Dembo, R., Schmeidler, J., Nini-Gough, B., Sue, C. C., Borden, P., & Manning, D. (1998). Predictors of recidivism to a juvenile assessment center: A three year study. Journal of Child and Adolescent Substance Abuse, 7, 57–77. doi:10.1300/J029v07n03_03

Duncan, R., Kennedy, W., & Patrick, C. (1995). Four-factor model of recidivism in male juvenile offenders. Journal of Clinical Child Psychology, 24, 250–257. doi:10.1207/s15374424jccp2403_1

Funk, S. J. (1999). Risk assessment for juveniles on probation: A focus on gender. Criminal Justice and Behavior, 26, 44-68.

Griller-Clark, H., & Mathur, S. R. (2010). Practices in transition for youth in the juvenile justice system. In D. Cheney (Ed.), Transition of secondary students with emotional or behavioral disabilities: Current approaches for positive outcomes (2nd ed., pp. 275–295). Arlington VA: Council for Children with Behavioral Disorders/Division of Career Development and Transition.

Griller-Clark, H., Mathur, S. R., & Helding, B. (2011). Transition services for juvenile detainees with disabilities: Findings on recidivism. Education and Treatment of Children, 34, 511–529.

Individuals with Disabilities Education Improvement Act (2004). P.L. 108–446.

Jones, P. R., Harris, P. W., Fader, J., & Grubstein, L. (2001). Identifying chronic juvenile offenders. Justice Quarterly, 18, 479–507. doi:10.1080/07418820100094991

Katsiyannis, A., Ryan, J., Zhang, D., & Spann, A. (2008). Juvenile delinquency and recidivism: The impact of academic achievement. Reading & Writing Quarterly: Overcoming Learning Difficulties, 24, 177–196. doi:10.1080/10573560701808460

McMackin, R. A., Tansi, R., & LaFratta, J. (2004). Special section: Studies in the rehabilitation of juvenile offenders, recidivism among juvenile offenders over periods ranging from one to twenty years following residential treatment. Journal of Offender Rehabilitation, 28, 1–15. doi:10.1300/J076v38n03_01

Morris, K., & Morris, R. J. (2006). Disability and juvenile delinquency: Issues and trends. Disability and Society, 21, 613–627.

Mulder, E., Vermunt, J., Brand, E., Bullens, R., & Marle, H. (2012). Recidivism in subgroups of serious juvenile offenders: Different profiles, different risks? Criminal Behaviour and Mental Health, 22, 122–135.

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

Nelson, C. M., Jolivette, K., Leone, P. E., & Mathur, S. R. (2010). Meeting the needs of adjudicated youth with behavioral challenges: The promise of juvenile justice. Behavioral Disorders, 36, 70–80.

Quinn, M. M., Rutherford, R. B., Leone, P. E., Osher, D. M., & Poirer, J. M. (2005). Youth with disabilities in juvenile corrections: A national survey. Exceptional Children, 71, 339–345.

Steketee, M., Junger, M., Junger-Tas, J. (2013). Sex differences in the predictors of juvenile delinquency: Females are more susceptible to poor environments; males are influenced more by low self-control. Journal of Contemporary Criminal Justice, 29, 88–105.

Tabachnick, B. G. & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Boston: Pearson Education.

Tille, J. E., & Rose, J. C. (2007). Emotional and behavioral problems of 13-to-18-year-old incarcerated female first-time offenders and recidivists. Youth Violence and Juvenile Justice, 5, 426–435. doi:10.1177/1541204007300355

Trulson, C., Marquart, J., Mullings, J., & Caeti, T. (2005). In between adolescence and adulthood: Recidivism outcomes of a cohort of state delinquents. Youth Violence and Juvenile Justice, 3, 355–387. doi:10.1177/1541204005278802

Vacca, J. S. (2008). Crime can be prevented if schools teach juvenile offenders to read. Children and Youth Services Review, 30, 1055–1062. doi:10.1016/j.childyouth.2008.01.013

van der Geest, B. C. (2008). Personal, background and treatment characteristics associated with offending after residential treatment: A 13-year follow up in adolescent males. Psychology, Crime, & Law, 14, 159-176.

Vitopoulos, N. A., Peterson-Badali, M., & Skilling, T. A. (2012). Male and female youth: examining the RNR principles in practice. Criminal Justice and Behavior, 39, 1025–141. doi:10.1177/0093854812442895

Zhang, D., Barrett, D. E., Katsiyannis, A., & Yoon, M. (2011). Juvenile offenders with and without disabilities: Risks and patterns of recidivism. Learning and Individual Differences, 21, 12–18. doi:10.1016/j.lindif.2010.09.006

Zhang, D., Hsu, H. Y., Katsiyannis, A., Barrett, D. E., & Ju, S. (2011). Adolescents with disabilities in the juvenile justice system: Patterns of recidivism. Exceptional Children, 77, 283–28. doi:10.1016/j.lindif.2010.09.006

OJJDP Home | About OJJDP | E-News | Topics | Funding
Programs | State Contacts | Publications | Statistics | Events