We administered the Rose Questionnaire to 1442 black, white, and Latino patients (approximately equal numbers) who sought care for acute chest pain at two medical centers. Of these, 718 subjects were enrolled at a large public hospital serving a low-socioeconomic status population and 724 at a large health maintenance organization hospital serving a middle-class clientele. Using the standard definition of Rose angina, multivariate logistic regression analysis identified five factors that contributed to the relative risk of a positive response: family history of myocardial infarction (2.48), history of peripheral vascular disease (1.41), history of high blood pressure (1.29), history of high cholesterol (1.26), and low-socioeconomic status hospital (0.78). Inquiring about shortness of breath as a substitute for chest pain or an alternative complaint in set one of the Rose Questionnaire did not increase the number of positive responses or differentiate between the socioeconomic groups or race-ethnic subgroups. Having a prior history of self-reported risk factors clearly defined a group with greater likelihood of a positive response to the Rose Questionnaire. Receiving care at a large public hospital (ie, being in a low-socioeconomic status group) was associated with reduced likelihood of having "typical" angina in comparison to receiving care at a health maintenance organization (middle socioeconomic status) for white subjects but not for Latinos and blacks.