A multiple baseline design across subjects and tasks and an alternating-treatments design were used. Five youths participated in training which involved the acquisition of a problem-solving strategy for one set of problems and increasing impulse-control through self-instructions for the second set of problems. Two different forms of hypothetical social problems were presented in individual sessions. The results showed improvement in both problem-solving and impulse-control responses as a result of treatment. The alternating-treatments design showed that some of the youths generalized the use of training procedures to untrained problems. Follow-up data suggested that maintenance of treatment gains occurred for most of the youths. General improvements in the youths' social behavior outside of the therapy sessions were addressed in terms of anecdotal information. (Author abstract)