OBJECTIVE We investigated the effect of race among Hispanic and non-Hispanic people on self-reported diabetes after adjusting for selected individual characteristics and known risk factors. METHODS Using the National Health Interview Survey 2000-2003, these analyses were limited to Hispanic and non-Hispanic people who self-identified as white or black/African American for a final sample of 117,825 adults, including 17,327 Hispanic people (with 356 black and 16,971 white respondents). RESULTS The overall prevalence of diabetes was 7.2%. After adjusting for selected covariates, Hispanic white and black respondents were 1.56 (95% confidence interval [CI] 1.32, 1.83) and 2.64 (95% CI 1.10, 6.35) times more likely to report having diabetes than non-Hispanic white respondents. The estimate for non-Hispanic black respondents was 1.45 (95% CI 1.29, 1.64). When compared to low-income non-Hispanic white respondents, low-income Hispanic white respondents (odds ratio [OR] 1.64; 95% CI 1.26, 2.19) and non-Hispanic black respondents (OR 1.71; 95% CI 1.38, 2.11) were more likely to report having diabetes. Hispanic black people born in the U.S. were 3.54 (95% CI 1.27, 9.82) times more likely to report having diabetes when compared to Hispanic white people born in the U.S. In comparison to non-Hispanic white respondents, the odds of reporting diabetes decreased for non-Hispanic black respondents, while the odds remained constant for Hispanic white respondents (p-value for interaction between survey year and race/ethnicity = 0.03). CONCLUSIONS This study suggests that race may be a proxy for unmeasured exposures among non-Hispanic and Hispanic people. Thus, given the importance of race on health and the racial heterogeneity among Hispanic people, race among Hispanic people should be investigated whenever the data allow it.