Risk and the Timing of Recidivism in Three Cohorts of Prison Releasees

Abstract

In recent years several methods have been developed to assess the timing of recidivism among offenders released from correctional programs. These methods have been used for some time in biomedical and engineering sciences, but applications to the study of criminal recidivism are relatively new. In this article, one of these methods, survival analysis, was used to examine the comparative performance of three groups of offenders released from a state prison system and followed for five years after release. The comparison of survival distributions showed that paroleesfailed at a slower pace than inmates released by conditional release or by the maximum expiration of the term. These findings were maintained when the a priori risk of recidivism was controlled, suggesting that selection effects were not wholly responsible for the more favorable performance of the parole group.

In the second part of the study, the survival distributions for the parole group were compared across risk levels. The findings suggested that if a critical period of probability of failure exists, the timing of this period varies with characteristics of the offender.

The implications of the findings for resource allocation, correctional population and program planning, parole policy, and program evaluation methodology are discussed.