On the causal interpretation of race in regressions adjusting for confounding and mediating variables

摘要: We consider several possible interpretations of the "effect race" when regressions are run with race as an exposure variable, controlling also for various confounding and mediating variables. When adjustment is made socioeconomic status early in a person's life, we discuss under what contexts regression coefficients can be interpreted corresponding to extent which racial inequality would remain if distributions life across groups could equalized. adult status, note how overall decomposed into portion that eliminated by equalizing even were stronger interpretation effect (stronger terms assumptions) involving joint effects race-associated physical phenotype (eg, skin color), parental phenotype, genetic background, cultural context such variables thought hypothetically manipulable adequate control possible. some challenges interpretation. Further discussion given use selected populations examining disparities additionally complicate effects.