To assist soon-to-be released prisoners in their re-entry process, the Community Mediation Maryland Re-Entry (CMM) program focuses on strengthening pre-existing pro-social relationships and assisting in the development of re-entry plans. Prior evaluations suggest this program is successful at reducing recidivism, but given the varied nature of the treatment, important questions remain. This thesis investigates potential treatment heterogeneity based on the subjects’ selection into different formats of the mediation process (i.e., selection of outside participants, focusing on emotional support, and reaching an agreement). The investigations into potential treatment heterogeneity (e.g., propensity score models) reveal null results. In the context of the prior evaluations, it is possible that the null results reflect that CMM is a generally effective program. Future research should consider capturing a larger sample of subjects mediating in the program so that subsequent evaluations can investigate important distinctions in the treatment received, which may be valuable in guiding future implementation of the re-entry mediation model.