Diabetes mellitus is an important public health problem that disproportionately affects minorities. Using a cross sectional, convenience sample, we estimated the prevalence of self-reported diabetes for Whites (n = 212), Arabs (n = 1,303), Chaldeans (n = 828), and Blacks (n = 789) in southeast Michigan. In addition, using a logistic regression model, we estimated odds ratios and 95% confidence intervals for the association between ethnicity and diabetes before and after adjusting for demographic, socioeconomic status, health care, chronic conditions, and health behavior variables. The overall age- and sex-adjusted prevalence of diabetes was 7.0%. Estimates were highest for Blacks (8.0%) followed by Arabs and Whites (7.0% for each group) and Chaldeans (6.0%). In the fully adjusted model, the association between ethnicity and diabetes was not statistically significant. Future studies should collect more detailed socioeconomic status, acculturation and health behavior information, which are factors that may affect the relationship between race/ethnicity and diabetes. [ABSTRACT FROM AUTHOR]