This study compared self-regulation and external regulation procedures in the treatment of children's disruptive dassroom behavior. After baseline data were collected, three of the four most disruptive children in each of 10 first-and second-grade dassrooms received reinforcement for achieving low rates of disruptive behavior. The fourth child served as a control subject throughout the experiment. Two of the three experimental subjects were then taught to self-observe their own disruptive behavior. In the final reinforcement period, these subjects were given control over dispensing reinforcers to themselves, based on their self-collected behavioral data while subjects in the other experimental group continued with the externally managed reinforcement In extinction, reinforcement was discontinued for all subjects, but one of the self-regulation subjects in each classroom continued overtly to self-observe. Results indicated that both reinforcement programs reduced disruptive behavior. The self-regulation procedures were slightly more effective in reducing disruptiveness than was the external regulation procedure, and this advantage persisted into extinction. These results suggest that selfregulation procedures provide a practical, inexpensive, and powerful alternative in dealing with disruptive behavior in children.Disruptive classroom behavior has often been the target of behavior modification technology. Many studies have demonstrated that rates of disruptive behavior can be substantially reduced by the systematic application of externally managed contingencies (e.g., Allen, Hart, Buell, Harris, and Wolf, 1964;Patterson, 1965;Homme, DeBaca, Devine, Steinhorst, and Rickert, 1963;Schmidt and Ulrich, 1969;Wasik, Senn, Welch, and Cooper, 1969;O'Leary, Becker, Evans, and Saudargas, 1969;Thomas, Becker, and Armstrong, 1968 (1969) demonstrated the potential of self-regulation for increasing a student's academic response rate. They found that higher academic rates occurred when the pupil arranged the contingency requirements than when the teacher specified them. In another classroom study, Glynn (1970) found that selfdetermined reinforcement was as effective as experimenter-determined reinforcement in increasing academic response rate and that differential token reinforcement experience influenced subsequent rates of self-determined reinforcement. Several other studies conducted in laboratory settings have further suggested the potential utility of self-monitoring and self-reinforcement in the modification of behavior. The results of studies on self-administered reinforcement have consistently demonstrated that behavior may be modified and maintained as well with a self-administered token reinforcement system as with an externally managed reinforcement system. Marston and Kanfer (1963)