A multiple-baseline design across subjects and an alternating-treatments design were used. Hypothetical social dilemmas were presented in individual sessions to the five youths participating in the program. The effects of intervention on the youths' self-recorded personal problemsolving was also assessed. Training involved the acquisition of a problemsolving strategy for the hypothetical dilemmas. Self-evaluation was later introduced through a checklist procedure. All youths showed improvement during the training of problemsolving on the hypothetical dilemmas. The self-evaluation procedure had little effect on performance. Some youths generalized their use of the problemsolving strategy to the untrained problems during the alternating-treatments condition. The maintenance of treatment gains occurred for all youths according to followup data. Also, the generalization of improved skills to personal problemsolving occurred for some youths. Anecdotal information suggested that there were general improvements in youths' social behavior outside of therapy sessions, and showed improved rates of progress through the institutions behavioral level system. 2 figures and 18 references. (Author abstract modified)