The Quality of Data on “Race” and “Ethnicity”: Implications for Health Researchers, Policy Makers, and Practitioners

Abstract

Implicit assumptions about the quality of data on “race” and “ethnicity” underlie the design of much of today’s research on health disparities. Health researchers, policy makers, and practitioners tend to take it for granted that racial/ethnic categories are clearly and consistently defined; that individual race/ethnicity can be easily, validly, and reliably determined; and that categories capture population groups that are so inherently different from each other that any reported racial/ethnic difference can automatically be generalized to the US population as a whole. This article outlines a series of issues that challenge these assumptions about the quality of race/ethnicity data. While race/ethnicity classifications can approximate socially constructed identities for some groups of people under some circumstances, these classifications are inherently too imprecise to allow meaningful statements to be made about underlying biological or genetic differences between groups. Findings of racial/ethnic differences should be reported with appropriate caveats and interpreted with caution. Particular caution should be exercised in hypothesizing genetic differences between groups in the absence of convincing genetic evidence.