The privatization of prisons is one of the most hotly contested topics in the public debates over the increasing presence of the criminal justice system in the lives of disadvantaged Americans. At the same time, there has been little research into the wellbeing of inmates housed in private prisons compared to their public counterparts. With a massive correctional population and overstretched state and federal budgets, however, there is an urgent need to inform correctional policies with reliable evidence. In this study, I contribute empirically to these policy debates by pursuing three goals. First, I use propensity score matching with the data from the 2005 Census of State and Federal Adult Correctional Facilities to examine the differences in access to health care resources and victimization between private and public prisons. Second, I apply the random forest algorithm to the same data to analyze facility-level predictors of the two sets of outcomes in private and public facilities. Third, I discuss the policy implications of the lack of inmate-level data from private prisons and the potentially effective alternatives to studying the impact of prison privatization on inmate health.