Evaluations of prison-based drug treatment programs typically focus on one or two dichotomous outcome variables related to recidivism. In contrast, this paper uses multiple measures of outcomes related to crime and drug use to examine the impact of prison treatment. Crime variables included self-report data of time to first illegal activity, arrest type, and number of months incarcerated. Days to first reincarceration and type of reincarceration are based on official records. Drug use variables included self-report data of the time to first use and drug testing results. Prisoners randomly assigned to treatment performed significantly better than controls on: days to first illegal activity, days to first incarceration, days to first use, type of reincarceration, and mean number of months incarcerated. No differences were found in type of first arrest or in drug test results. Subjects who completed both prison-based and community-based treatment performed significantly better than subjects who received lesser amounts of treatment on every measure. Survival analysis suggested that subjects were most vulnerable to recidivism in the 60 days after release. Although the overall results from the analyses presented support the effectiveness of prison-based treatment, conclusions about the effectiveness of a treatment program may vary depending on which outcomes are selected. The results of this study argue for including more rather than fewer outcomes in assessing the impact of prison-based substance abuse treatment. (Contains 7 tables and 7 figures.)