We wrote “Racial Fluidity and Inequality in the United States” (Saperstein and Penner 2012) with the aim of jump-starting a conversation about how race is best conceptualized in studies of stratification. Does assuming that people have static, often mutually exclusive, “races” help us understand disparities in the contemporary United States? Or are there inequality-sustaining mechanisms we might have missed by assuming a process of consensus through which racial categorizations are ascribed at birth and effectively fixed? That two teams of scholars took time to engage with our work is a positive sign that a much-needed conversation is happening. It is, of course, disheartening that Kramer et al. (2016) (hereafter KDH) and Alba et al. (2016) (hereafter AIL) come to the conclusion that we, at worst, misinterpreted our data and, at best, overstated our case. Nevertheless, we are grateful to have this opportunity to clarify our claims and intentions and to offer new evidence that returning to assumptions of rigid racial ascription is not the way forward.
Our reply addresses the three main points of empirical critique across the two comments: (1) that a relationship between social status and racial categorization of similar direction and magnitude could be produced by measurement error; (2) that our findings are neither as common nor as generalizable as we claimed; and (3) that we did not adequately demonstrate that stereotypes, operating through what the interviewer does or does not hear about the respondent, are a key causal mechanism. We either provide evidence that directly refutes each point, explain why it results from a misreading of our argument, or both. In the process, we underscore our earlier findings with additional evidence of how social factors shape categorization: through selective processes of “ethnic attrition” as well as what interviewers knew about the respondents’ use of crack cocaine.