Most functional analyses of problem behavior include a common condition (play or noncontingent reinforcement) as a control for both positive and negative reinforcement. However, test-specific conditions that control for each potential source of reinforcement may be beneficial occasionally. We compared responding during alone, ignore, play, and differential reinforcement of other behavior (DRO) control conditions for individuals whose problem behavior was maintained by positive or negative reinforcement. Results showed that all of the conditions were effective controls for problem behavior maintained by positive reinforcement; however, the DRO condition was consistently ineffective as a control for problem behavior maintained by negative reinforcement. Implications for the design of functional analyses and future research are discussed.