It is important for genetic, epidemiologic, and nosological studies to determine accurate rates of lifetime psychiatric diagnoses in patient and nonpatient populations. As part of a case-control family study of major depression, lifetime psychiatric diagnoses were made for 1,878 individuals. Sources of information used in making diagnostic estimates included direct interview, medical records, and family history data systematically obtained from relatives. Diagnostic estimates were made by trained interviewers, experienced clinicians, and by computer program. The results indicate that it is possible to make lifetime best estimate diagnoses reliably among both interviewed and noninterviewed individuals for most diagnostic categories and that diagnoses based on interview data alone are an adequate substitute for best estimate diagnoses based on all available information in a limited number of diagnostic categories.