Summary:Research data describing the prevalence and patterns of behaviors that place persons at increased risk of HIV infection are extremely limited. The scarcity of data has constrained potential applications of surveillance data, research on specific high-risk behaviors, and epidemiological studies. This article critically reviews available research for four population groups: intravenous drug users, homosexual males, and sexually active adolescents and adults. The fallacies inherent in estimating risk group size underscore the need for population-based research that can provide detailed data on sexual and drug use behaviors.