The sample consisted of 529 offenders released on parole. An experimental design with random assignment to one of three groups was employed. The Experimental Group received frequent, random drug testing with instant results, immediate sanctions, and referral for substance abuse treatment. Control Group I received frequent, random drug testing and treatment referral, but did not receive immediate test results or immediate sanctions. Control Group II followed standard parole practice. Members of this group were not tested on a random basis and did not receive immediate sanctions. Repeated measures ANOVA and survival analysis techniques were used to explore group differences. Frequent monitoring of drug use with randomized testing protocols, immediate feedback, and certain consequences is effective in lowering rates of relapse and recidivism. The effectiveness is particularly salient in the short term during the period of exposure to testing conditions. The findings lend support to the use of randomized testing with swift and certain sanctions with parolees. Additional quality evidence is necessary to generalize and refine findings from this study and others that focus on sanction certainty. Future replications must consider the immediacy of test result and sanction execution as well as the length of exposure to randomized testing periods. Abstract published by arrangement with Springer.