OBJECTIVES: To determine prevalence and correlates of urinary (UI), fecal (FI), and dual (DI) incontinence in community-dwelling older adults. DESIGN: Cross-sectional, population-based in-home survey. SETTING: Three rural and two urban Alabama counties (1999-2001). PARTICIPANTS: Stratified random sample of 1,000 Medicare beneficiaries aged 65 and older: 25% African-American men, 25% white men, 25% African-American women, 25% white women. MEASUREMENTS: UI defined as involuntary urine loss at least monthly; FI defined as "In the past year, have you had any loss of control of your bowels, even a small amount that stained the underwear?" Independent variables were sociodemographics, medical history, and activities of daily living (ADLs) excluding toileting. Multinomial logistic regression models were constructed using UI, FI, and DI as dependent variables. RESULTS: UI, FI, and DI prevalences were 27%, 6%, and 6%, respectively. White women had higher prevalence of UI (41.1% vs 24.6%, P<.001) and DI (18.5% vs 8.0%, P=.07) but not FI (4.4% vs 3.8%, P=.83) than African-American women. Men had no racial differences in prevalence (UI 22.2%, FI 7.7%, DI 4.6%). In women, UI, FI, and DI correlates included higher Charlson comorbidity score and depressive symptoms and greater ADL difficulty for UI and DI. In men, UI was associated with prostate disease and depressive symptoms; FI with chronic diarrhea, history of transient ischemic attack (TIA), and poor self-perceived health; and DI with history of TIA, foot and leg swelling, and depressive symptoms. CONCLUSION: Women had racial differences as well as similar correlates of incontinence subtypes, whereas men did not.