Although some scholars suggest that racial/ethnic concordance between physicians and patients will do much to eliminate disparities in medical care, the evidence for concordance effects is mixed. Using nationally representative data with an oversample of blacks and Latinos, this study examines a variety of topics, including beliefs about and preferences for concordance, the effects of concordance on patient experiences, and interactions between expectations and experiences. The results point to the limited effects of concordance in general but illuminate for whom concordance matters most. The results encourage more nuanced and contingent theories. They suggest that racial/ethnic concordance holds little salience in the minds of most black and Latino patients and that discordance has little effect. Nevertheless, there is some evidence that concordance has a positive effect among those who prefer concordance-thus the apparent effects of concordance might reflect the effects of patient choice more than concordance per se. The conclusion sketches policy implications, including the merits of promoting concordance among targeted groups of patients, even in the absence of overall effects on disparities.