Study
Jordan (2012) conducted a quasi-experimental design to study the impact of juvenile transfer on recidivism among Pennsylvania juveniles. The study included 308 males and females, charged with a violent crime and use of a deadly weapon, from three Pennsylvania counties (Philadelphia, Allegheny, and Dauphin), between the ages of 15 and 17, who were followed for 8 years.
Juveniles who are waived to adult court under Pennsylvania Act 33 are able to request a hearing to be decertified and sent back to the juvenile courts. Of the 308, 135 were decertified (i.e., returned to juvenile courts) and 173 were non-decertified (i.e., transferred to adult courts). Propensity score matching was used to counter any impact of selection bias. There were no significant differences between the treatment and comparison groups on baseline characteristics.
Data for recidivism was collected from two sources. First, the Center for Juvenile Justice Training and Research provided rearrest data prior to the person’s 18th birthday. Second, the Pennsylvania Commission on Crime and Delinquency provided state adult criminal records for the 8-year follow-up period.
Study
Myers (2003) used a quasi-experimental design to analyze the statewide data of juveniles to compare recidivism between juveniles transferred to adult courts with juveniles retained in the juvenile courts, prior to the March 1996 revision of Pennsylvania’s legislative waiver law.
The study included 494 male juveniles, between the ages of 15 and 18, who had been arrested for robbery, aggravated assault, or both, with the use of a deadly weapon. Of the study participants, 79 were transferred to adult courts (treatment group) while the remaining 415 were retained in juvenile courts (comparison group). There were no significant differences between the treatment and comparison groups on baseline characteristics.
The data used in this study was collected soon after the modified Act 33 went into effect. Therefore, cases processed under the new law could not be used as those cases were still being processed. This study was not a direct test of the effectiveness of Act 33; instead, it was instrumental in defining the population to be studied.
Case information on the juveniles was obtained through the Center for Juvenile Justice Training and Research’s Statistical Analysis Center, which compiles data and publishes annual reports on the activities of all juvenile courts in Pennsylvania. Recidivism measures were obtained through official arrest, case dispositions, and criminal records checks. This data included all cases where “transfer to criminal court” was the final disposition. The study looked at general recidivism, which included any arrest during the follow-up period; and violent felony recidivism, which included any violent felony arrest (offenses such as rape, involuntary deviate sexual intercourse, aggravated assault, robbery, robbery of a motor vehicle, aggravated indecent assault, kidnapping, voluntary manslaughter, or an attempt, conspiracy, or solicitation to commit any of these offenses) during the follow up period.
Study authors noted that, due to the lack of random assignment, there was a possible threat of selection bias. Therefore, a logistic regression model was used to control for these possible threats.