A meta-analysis was conducted on 73 studies of social skills training in four adult psychiatric populations: developmentally disabled, psychotic, nonpsychotic, and legal offenders. Findings from this analysis showed that patients participating in social skills training programs broadened their repertoire of skills, continued to demonstrate these skills several months after treatment, and showed diminished psychiatric symptoms related to social dysfunctions. Although results from an ANCOVA comparing effect sizes across the four populations (with design quality as a covariate) were nonsignificant, consistent trends suggested that social skills training had the greatest effect on developmentally disabled groups and the least effect on offender groups. In addition, social skills training was found to be relatively more effective in outpatient than inpatient settings.