Clinical investigators face formidable methodologic and ethical challenges in the study of elderly people. To define problems specific to geriatric research the fundamental elements of clinical research are outlined and considerations important to a geriatric study population are identified. Because of the underestimated biologic variation of elderly people, the evaluation of new interventions in one older population may be difficult to generalize to all older people. Inasmuch as older people represent a study population with a lifetime of cumulative exposures for various diseases such as cancer, historical cohort studies are particularly suitable research designs. Despite these advantages, the retrospective study of certain interventions research designs. Despite these advantages, the retrospective study of certain interventions in older people, for example cardiopulmonary resuscitation, requires thoughtful attention to avoid strong and often subtle selection biases. Issues of informed consent and the presence of multiple, comorbid illnesses, especially relevant to clinical trials, are problematic in studying geriatric conditions. Recently, instruments measuring function and disability have been used extensively by investigators in geriatric research, but the precise way to quantify these observations remains controversial. Testing artifacts may be considerably greater in older persons than in many of the populations in which the instruments have been validated. Finally, certain statistical assumptions and inferences may be inappropriate in the study of older people.