Although randomized clinical trials are the most scientifically rigorous design for evaluating a clinical intervention, the methodology involved is intricate, particularly when elderly subjects are involved. Special attention must be paid to framing precise primary and secondary study hypotheses, adjusting for the potential confounding effects of comorbidity and disability, and analyzing multiple outcomes. When the sample or study cohort is chosen, there will be an inevitable trade-off between internal and external validity. Because elderly study cohorts are more prognostically heterogeneous, consideration should be given to enrolling an "at-risk" cohort and possibly to stratified randomization.