Recidivism Measures

Definitions of recidivism in past studies have varied. Generally, recidivism has been measured as arrest, conviction, or imprisonment during a specified period of time following release from correctional custody. Variations in the definition of recidivism affect the results. Re-arrests, for example, are more likely to occur than re-conviction or re-imprisonment since rearrest may not lead to re-conviction and re-conviction may not lead to re-imprisonment. Also, rearrest and re-imprisonment can be due to technical violations, which will increase their numbers in comparison with re-conviction for a new offense. Traditionally researchers use a limited amount of post prison release follow-up time, ranging from a few months to a number of years. Consistent with most contemporary recidivism research, this study analyzes the length of time to recidivism and the patterns in that time to failure for the treatment and control groups.

Following the FDOC’s recidivism measurements and rate analysis, this study employs two recidivism measures, re-offense and re-imprisonment. Re-offense is measured as the number of months from prison release to the first felony offense. Re-imprisonment is measured as the number of months from prison release to first re-admission to prison for an intervening offense. Both measures require a conviction for a new offense to have occurred and that the conviction result in either a prison or supervision sentence to the department.

The length of time to re-offense or re-imprisonment can be artificially long for those inmates who return to prison for a technical violation of post-release supervision. The FDOC accounts for this “not at risk time” by reducing the follow-up time by the number of months spent back in prison for a technical violation. Likewise, the recidivism measures used in this study are adjusted to exclude not at risk time. For the subset of inmates who had post release time in prison for technical violations, both the follow-up time and the time to failure represent the true number of months at risk for recidivating.