Eighty-seven studies published since 1964 through 1982 on training self-help skills to severely and profoundly mentally retarded persons were analyzed according to 19 parameters reflecting their methodological details. The results showed a steady interest in this research area over time, but 63% of the studies focused on toileting and feeding with fewer studies looking at other self-help skills. Package treatments composed primarily of accelerative techniques were most frequently used to train these skills. Methodologically, it was found that these studies typically involved profoundly mentally retarded people (33% of studies) who were trained by residential staff (69% of studies) in institutional settings (63% of studies). The results also indicated an increase over time in the number of studies rated acceptable on the reliability and design parameters. Finally, very few studies reported assessments of generalization, maintenance, or social validity. It was concluded that, (a) researchers need to broaden their interests in terms of settings, trainers, and behaviors studied to best meet the needs of this population, (b) the experimental quality of this literature is improving, and (c) the social impact of observed behavior changes has yet to be fully explored.