Criminological research has long explored the impact of incarceration on recidivism. However this line of research typically combines jail and prison as a single type of incarceration sentence and does not distinguish between these two different sentences. Judges often can choose between a prison or jail sentence for certain categories of offenders, but little research offers any perspective on potentially differential impacts of the two types of sentences on recidivism. The lack of understanding about the consequences of prison and jail sentences hampers the efforts to improve the criminal justice system. Using data from Pennsylvania, the current study examines the impacts of prison and jail, as two separate types of sentences, on recidivism. Pennsylvania offers a unique sample of offenders where jail inmates may be serving up to five years’ incarceration and prison inmates may be serving a sentence shorter than 1 year , compared to most other states where jail is typically a 1-year sentence or less and prison serves for sentences longer than 1 year. Propensity score matching was utilized to compare those with jail and prison sentences who are otherwise comparable, thus allowing balance on observables. In particular, the sentencing guidelines were utilized to ensure the comparability of offenders regarding the seriousness of the current offense and the prior record and to focus on offenders who could have received either prison or jail sentence based on judicial discretion. Results from a statewide sample indicate that within a short follow-up period following release from incarceration, there is largely a null effect of prison on rearrest. However, given longer recidivism follow-up periods, prison is shown to increase the probability of rearrest compared to jail. Although, using a county sample reveals a negative effect of prison on rearrest within a year of release, once reincarceration is accounted for and longer follow-up periods are used, there is largely a null effect of prison on recidivism. Several subsamples were examined revealing largely null effects. Findings indicated racial differences in recidivism based on confinement in prisons and jail. Additional tests were conducted to attempt to understand the differences within facilities that could affect recidivism by comparing across jails, although the results were largely null. Findings from this study reveal key policy implications regarding judicial decision making. In the long-term, once reincarceration is taken into account, there is no strong evidence to suggest that a difference in recidivism exists between prison and jail, suggesting that judge sentencing decisions can be guided by budgetary concerns and facility capacity.