How and When to Make Causal Claims Based on Race or Ethnicity.Abstract. 2012.
Causal inference is considered the gold standard in social science research. Making causal claims about ``immutable characteristics'' such as race, however, has been strongly discouraged. In contrast to previous literature, which assumes a fixed conception of race, we propose a different framework that in some cases reconciles race and causation. First, we distinguish those units of analysis in which intrinsic problems of race and causality can be avoided. Second, we demonstrate that race can be defined as a composite measure that has some mutable elements. These extensions allow us to synthesize two areas where causal claims about race may be permissible: (1) studies that measure the effect of exposing an entity to a racial cue and (2) studies that disaggregate race into constituent pieces and measure the causal effect of some mutable element. We demonstrate these techniques via examples from contemporary scholarship