Treatment-based Approaches to Juvenile Crime: Effects on Education Outcomes

Youth crime is recognized as a serious social problem.  Further, it is symptomatic of conduct disorder which is considered a serious mental health problem in youth.  One significant policy issue is how best policymakers should respond to mental health problems that express themselves through crimes.  Once delinquent youth are identified by the justice system, key sanctioning alternatives include minimal or no supervision, community treatment services, or incarceration.  These sanctioning alternatives embed potentially significant trade-offs.  Evidence-based mental health services that are provided in the community, such as functional family therapy or aggression replacement therapy, have been found to be cost-effective compared when compared to usual probation services.  However, these findings based on randomized trials do not address the full range of sanctions, nor do they examine treatment as offered or taken-up in practice. Further they do not address outcomes for offenders who are determined not to qualify for evidence-based programs, due to the nature of the youth’s offense or due to other family circumstances.  Finally, they do not examine impacts on human capital formation, particularly educational outcomes, leaving major gaps in our understanding of optimal approaches to delinquent youth. 

This study extends previous work and examines differences in outcomes between youth who receive specialized community treatment or remain untreated in the community with usual probation. Using unique, linked administrative data, both education and criminal justice outcomes are examined.

This study uses a unique multi-year merged data set of juvenile court administrative data, screening and assessment data, and education data which are linked at the person-level.  The screening data include rich information on juvenile’s family, school, employment, mental health, and substance use. The justice data include arrest, type of offense, and type of sanction, including dates.  The education data include completion of high school and GED, advancement, absences, drop-out status, cumulative GPA and selected test scores. 

We use two empirical approaches to compare those who received specialized treatment to those who do not.  One is a comparison of outcomes with extensive controls based on rich administrative and screening data. The other takes advantage of discontinuities in the program eligibility screening scores to construct comparison groups in the context of a Regression Discontinuity Design (RDD).  Eligibility cut-offs for services based on the assessment data allow us to compare youth just above and below cut-offs and identify the effect of alternative sanctions. To identify the effect of incarceration we use discontinuities in the determinate sanctioning grid imposed by state law.  We compare youth just below and above sanction cut-offs on the assumption that youth just above and below the grid are similar along many relevant dimensions, an assumption which we are able to test with respect to observable characteristics.