Research That Matters (January 17 - 20, 2008)

The proliferation of legislative changes easing the process of transferring juveniles to the criminal court have raised important questions concerning the public safety implications of transferring juvenile offenders. Employing a variety of techniques to reduce selection bias, studies of the recidivism of juveniles transferred to the criminal court compared to those retained in the juvenile court have been very consistent in showing that transferred juveniles are more likely to recidivate, recidivate faster, and are more likely to commit serious offenses than youth retained in the juvenile court (e.g., Fagan et al., 2003).

A key limitation of this research is that it utilizes variables measured at the time of the sentence and does not incorporate subsequent information such as amount of prison time served and age at release. Myers (2003) illustrates the potential importance of these variables by incorporating sentence length into his recidivism model. Two key findings emerged from this study: (1) youth retained in the juvenile court were less likely to recidivate than transferred youth; and, (2) among transferred youth, those with longer prison sentences were less likely to recidivate than those with shorter sentences or those receiving probation sentences. While these findings led Myers to recommend keeping youth in the juvenile court, they also led him to speculate that longer prison sentences might serve as a deterrent. A limitation of Myers' study is that it relied on sentence length and not actual time served or age at release, a significant limitation because time served varies substantially across individuals based on a variety of factors besides sentence length.

This paper uses administrative data from the Michigan Department of Corrections to examine the factors that influence the recommitment of juveniles to adult prisons. Using survival analysis, we find that youth with more extensive prior histories in the juvenile system and youth entering prison for less serious committing offenses are more likely to be recommitted. We also find that when independently entered, time served and age at release are both significantly associated with a lower likelihood of recommitment. When entered simultaneously, however, only age at release remained a significant predictor.

Other analyses of Michigan data indicate that the type of offender most likely to be recommitted is the type of offender increasingly being committed to adult prisons. Taken together, these findings demonstrate a need to rethink whether current policies and practices are the most effective means for dealing with serious and chronic juvenile offenders. Viewed in conjunction with studies indicating that juveniles retained in juvenile court are less likely to recidivate than transferred youth, these findings suggest that transfer should be reserved for youth who are both chronic and serious offenders, whereas first-time serious offenders or chronic juvenile offenders with a less serious current offense may be better served and present less likelihood of recidivism when retained in the juvenile court.