As a result of the significant U.S. prison population build-up over the past several decades, a large number of inmates are now being released from prison and returned to the community. One mechanism for facilitating this transition to the community is for inmates to be conditionally released under parole supervision. Once on parole, a parolee is subject to certain rules and conditions that, if violated, can result in a return to prison, even if not a criminal act. These types of non-criminal parole violations are typically referred to as Technical Parole Violations (TPVs). Many states return a large number of TPVs to prison each year, and TPVs contribute significantly to the prison population in many states. However, there is virtually no existing research examining what impact returning TPVs to imprisonment has on their subsequent rates of re-offending. While a large body of literature examining the overall impact of incarceration on recidivism has mostly concluded that imprisonment has a null or even slightly criminogenic effect, this overall finding is not necessarily generalizable to all sub-populations within the prison population. Strong theoretical cases can be made each way, for the impact on recidivism of incarcerating TPVs.

This dissertation examines the impact on recidivism of sanctioning TPVs to imprisonment versus an alternative sanction, and also examines the dose-response impact on recidivism of varying lengths of stay in prison for a TPV, using a large sample of TPVs in one state (Pennsylvania). The bulk of the evidence supports the conclusion that recidivism rates are mostly lowered by using incarceration in response to first TPV violations. However, the evidence also suggests that the specific mechanism for lowering recidivism rates among incarcerated TPVs is largely attributable to aging and exposure time rather than to deterrence.

The findings on the dose-response impact of differential lengths of stay in prison for TPVs who are sanctioned to imprisonment are more mixed. Generally the evidence suggests somewhat lowered recidivism rates attributable to longer lengths of stay in prison for a TPV violation, yet the effect sizes are generally smaller and in some cases statistically insignificant. It again appears that the particular mechanism for reduced recidivism rates associated with longer lengths of stay in prison is associated with aging and exposure time rather than with traditionally formulated deterrence mechanisms.

A few contingencies of these findings are noted. First, the effect of imprisonment on recidivism among TPVs is likely highly contingent upon the swiftness, certainty, and perceived fairness of sanctioning, yet measures of these factors were not available for this study. Second, this dissertation only focuses on the first TPV violation instance after release from prison, and also is mostly limited to higher risk TPVs. Third, lower overall recidivism rates for TPVs sanctioned to imprisonment, and sanctioned for longer periods of time in prison, were influenced heavily by lower re-incarceration rates, whereas re-arrest rates did not significantly differ in any of the models. Since re-incarceration rates not only include new criminal activity but also new technical violations, it is unclear whether imprisonment for a first TPV reduces serious criminal behavior or rather mostly reduces additional technical violations and minor crimes. Future research must address these contingencies.