Data concerning familial history of psychiatric disorders are often used to assist in diagnosis, to examine the role of genetic or nongenetic familial factors in etiology, or to develop new methods of classification. Information concerning familial prevalence may be collected by two different methods: the family history method (obtaining information from the patinet or a relative concerning all family members), and the family study method (interviewing directly as many relatives as possible concerning their own present or past symptomatology). This study compares these two methods. In general, the family study method is preferred since information is likely to be more accurate. The family history method leads to significant underreporting, but this can be minimized through the use of diagnostic criteria. This study reports on an instrument that has been developed for collecting information concerning family history and that provides criteria for 12 diagnoses--the Family History-Research Diagnostic Criteria. Using diagnostic criteria leads to greater sensitivity, but underreporting remains a major problem of the family history method.