Social and biomedical scientists, journal editors, and public health officials continue to debate the merits of the use of race and ethnicity in health-related research. As biomedical research focuses on issues of racial or ethnic health disparities, it remains unclear how biomedical scientists investigate race or ethnicity and health. This paper examines how biomedical researchers construct and analyze race or ethnicity in their studies and what conclusions they make about difference and health. Using content analysis of 204 biomedical research journal publications, which were supported by grants won from the National Cancer Institute of the National Institutes of Health in the USA, I demonstrate that although authors tended to see race or ethnicity as important and significant in their research, they rarely defined or operationalized the concepts adequately. Moreover, when presenting findings of racial or ethnic difference, authors generally did not provide explanations of the difference. I argue that this under-theorized and unspecified use of race or ethnicity and the biological conclusions drawn about health and difference have the potential to reify "race" and to limit our thinking about what these biomedical differences suggest about health disparities and inequalities in general.