Four boys with autistic-like behavior were treated for self-stimulatory behavior with three different treatment procedures--time out, differential reinforcement of other behavior (DRO), and overcorrection. All four boys showed a rapid response to the overcorrection procedure. Three boys demonstrated some evidence of decrement in responding with time-out. During the DRO procedure, one showed a modest decrease, two showed no change, but one exhibited a consistent increase in responding under this condition. A multiple baseline applied to one of the subjects failed to reveal any generalization of suppression from one setting to another. A strong but not perfect relationship was found between a frequency and a duration measure of self-stimulation. There was some evidence of negative side effects for one boy during overcorrection and for another during time-out. None of these negative side effects was enduring. There was also some indirect evidence that overcorrection facilitated appropriate play.