Empirical studies that have used DSM and ICD criteria with people who are mentally retarded were reviewed. They included both studies using clinical interviews and case notes for diagnosis and those using behavioral checklists. The merits of these methods, which have hitherto developed relatively independently, were critically evaluated. Areas for future development include: the application of the psychometric rigor, hitherto associated with behavioral checklists, to diagnosis based on clinical interviews and case notes; the development of broader and associated measures of the integrity of the diagnostic process; and the development and piloting of new criteria for some disorders for use with people with severe and profound mental retardation where current diagnostic criteria are inappropriate with this population.