Epidemiologic studies of fatal diseases often require that information be sought from relatives or friends of deceased or disabled patients. The authors have evaluated the ability of several types of surrogate respondents to provide information on the smoking, occupational, medical history, and demographic characteristics of their next of kin in three recent case-control studies involving interviews with 2606 individuals. The ability of the surrogates to provide this information varied by topic, degree of detail requested, race, sex, age, and study area, but was most affected by the type of respondent. Sibs were best able to respond to questions about the subject's immediate family or events that occurred during early life, while spouses and offspring were best able to describe events that occurred during adult life. Several recommendations are made to improve the design of future interview studies.