Prediction of Recidivism in Adult First-Time Offenders

Patricia M. Harris, University of Texas at San Antonio

Much research on criminal careers consists of longitudinal studies of onset and development of criminality in youthful populations, generally using birth cohorts. Typically, researchers resort to those same longitudinal studies to examine the criminal careers of adults, reflecting the generally accepted view that most adult offenders began their careers as juveniles. This approach to understanding criminal careers has resulted in an emphasis on criminal history variables in formal instruments for assessing the likelihood of reoffending. A limitation of such tools emerges in the context of community corrections, where large proportions of adult offenders have no criminal history other than the instant offense. Though these first-time offenders generally score low on risk assessments, they nonetheless contribute substantially to rates of rearrest, revocation and rules violations. This paper introduces a empirically-based typology of adult first time offenders, using a large (n=1,736) cohort of first time offenders sentenced to felony probation in Texas during a one month period. It then develops and compares the utility of alternate models for predicting recidivism in these subjects, roughly one-third of whom were rearrested and roughly one-fifth of whom were revoked over a subsequent three year follow-up.

(Return to Program Resources)

Updated 05/20/2006