Abstract
The purpose of this study was to compare three conditions in a single-subject reversal design: no contract and no contingencies, a contract without contingencies, and a contract with contingencies. There were eight subjects who remained long enough in the halfway house environment to provide sufficient data for comparison across conditions. The results indicate that managerial behaviors (such as keeping a daily log and using a sign-in sheet) came under experimental control only during the “contract with contingencies” condition. Drug use did not come under control during any condition, apparently because adequate competing reinforcers were unavailable in this study.