Thinking outside the (black) box: Measuring black and multiracial identification on surveys

To better understand the diversity of the multiracial population, compare multiracial data to single-race data, and evaluate the rigidity of racial boundaries, we must understand the single-race identification choices of multiracial respondents. Many studies assume that this pattern will be straightforward for multiracial respondents who choose a part-black identification, with virtually all choosing a "black" single-race identification. I investigate whether this assumption is justified by available survey data. Using the May 1995 Current Population Survey's Race and Ethnicity Supplement, I explore the single-race identifications of individuals who have chosen a part-black multiracial label on a survey. I find that single-race identification choices on forced-choice questions vary considerably across family heritage groups, with those who choose a "black-American Indian" identity extremely likely to select a black single-race identity, while other groups like "black-whites" have substantial variation in single-race identifications. Identification patterns vary significantly by age, family context and survey characteristics. (c) 2006 Elsevier Inc. All rights reserved.