The purpose of this article was to determine whether risk factors for four dimensions of disability differ and whether it is legitimate to use aggregated disability measures in risk factor analyses. Using data from the baseline Duke Established Populations for Epidemiologic Studies of the Elderly survey (n = 4,162), the authors examined four measures of disability--basic activities of daily living (ADLs), household ADLs, advanced ADLs, and mobility--and an aggregated measure consisting of these four measures summed. Sociodemographic risk factors were examined using stagewise multivariate regression analysis for the five measures of disability. Weighted least squares with an arbitrary distribution function estimator were used to determine differences in each risk factor's performance across the unaggregated measures. Risk factors varied in strength, presence, and direction of impact across the four dimensions of disability; as a result, analyses using an aggregated measure were misleading.