Researchers who study differential outcomes based on racial classifications should acknowledge the sociopolitical forces that influence the concept of race. In this article, logistic regression findings based on a traditional methodological approach to race are compared to an approach that used a composite variable incorporating race/ethnicity and immigration status. Participants were 258 ethnically diverse low-income women drawn via convenience sampling from the Women, Infants, and Children's Program (WIC) in or near a northeastern city. The PrimeMD Patient Health Questionnaire was used to identify whether the women had subthreshold or major depressive syndrome. The analyses using the composite variable better account for the findings than the analyses using race as a separate variable. Researchers should strive to identify and utilize various dimensions of participants' social positions (e.g., immigration status, social economic status and language) that help explicate differential outcomes.