A variety of approaches for addressing drug use and drug-related crime among the nearly 5 million offenders on community supervision in the United States has been tried and evaluated, but questions remain about which policies or programs are most effective. The authors use a large dataset to assess the impact of residential and nonresidential drug treatment on recidivism. Propensity score matching is used to establish equivalent treatment and comparison groups and to enable comparisons of treatment type. Survival analysis is used to determine the extent to which each treatment modality and numerous covariates were associated with time until recidivism. Compared to those receiving no treatment, those receiving nonresidential treatment took longer to fail or recidivate. However, those receiving residential treatment did not differ from those who received no treatment in time to failure. In the treatment-only model, nonresidential treatment participants took longer to fail than their matched residential treatment counterparts. (Published Abstract)