Statistical quality control procedures such as acceptance sampling for a manufacturing process can be considered reactive. In contrast, in social settings, in which process may not be well-defined or directly or easily controlled, testing of a well-defined output can provide a means for constructively affecting the process. This approach can be useful in relation to the use of illegal drugs, which has been a major focal point for policy-makers and decision-makers for more than a decade. The model uses Bayesian acceptance sampling techniques to test a probation population for drug use. These techniques appear to improve on the ad hoc approaches to drug testing used in the past. Acceptance sampling allows the decision-maker to enforce a decision rule when the number of individuals who test positive with a sample of the population is excessive. A Bayesian sampling plan also gives the decision-maker more complete information from which to plan a drug testing program and thus considers all the subjective and objective costs associated with drug testing and drug use. Using this testing approach can maintain deterrence and monitoring effects while reducing the expected cost of testing. Tables, footnotes, and 26 references