Data on research participants and populations frequently include race, ethnicity, and gender as categorical variables, with the assumption that these variables exert their effects through innate or genetically determined biologic mechanisms. There is a growing body of research that suggests, however, that these variables have strong social dimensions that influence health. Socioeconomic status, a complicated construct in its own right, interacts with and confounds analyses of race/ethnicity and gender. The Academy recommends that research studies include race/ethnicity, gender, and socioeconomic status as explanatory variables only when data relevant to the underlying social mechanisms have been collected and included in the analyses.