A Plea for Ignoring Race and Including Insured Status in American Research Reports on Social Science and Medicine
My argument is that the term 'African-American' is too broad a category to be useful in social science research. Rather, researchers should (if this is a relevant exercise) carefully define and measure the ethnicity of their respondents or subjects. Only when specific hypotheses about ethnic or cultural group differences in beliefs about illness, symptom presentation and availability and use of treatment modalities are explored. should ethnicity itself be defined and measured. A good example of such an exercise is the study (recently published in Social Science & Medicine) by Bates and Rankin-Hill  on "control, culture and chronic pain".