Measuring Segregation in a Multiracial Era: The Impact of Federal Racial Reporting Changes on Estimates of Public School Segregation

Background Despite accounting for 3% of the student population, multiracial students are the fastest growing racial/ethnic group in the United States. Aligning with new federal guidelines, in 2008, the National Center for Education Statistics (NCES) revised its single-race reporting scheme for the Common Core of Data (CCD), permitting students to identify as two or more races (i.e., multiracial). Study Objective The addition of a multiracial reporting category to the NCES CCD permits students to more accurately represent their racial identity. By creating categories that are not strictly comparable to their historical counterparts, however, it may problematize longitudinal analysis of trends in the racial composition and segregation of schools, which are of perennial scholarly and public interest. In this study, we examine the extent to which the reporting change affects estimates of segregation over time. Research Design We compute annual values of racial/ethnic segregation (i.e., isolation and dissimilarity) among elementary schools for 5,357 public school districts using data from the NCES CCD Public Elementary/Secondary School Universe Survey from 2000–01 to 2013–14. Before 2008, we compute segregation for the five monoracial groups tracked by the CCD. After 2008, we compute segregation using the two analytical approaches to the treatment of multiracial students that have been employed by scholars of segregation. We estimate the effect of the racial/ethnic reporting change on estimates of segregation via a series of multilevel longitudinal fixed-effects models. Results We found that the NCES CCD's addition of a multiracial reporting category has problematized longitudinal analyses of segregation, particularly racial/ethnic isolation. Models revealed abrupt changes in most dimensions of segregation in the year immediately following the reporting change. Moreover, the reporting change has complicated interpretation of segregation trends since 2008, because seemingly minor analytical decisions may lead to contradictory conclusions: When multiracial students are excluded from calculations, isolation appears to be increasing for all racial/ethnic groups; however, when multiracial students are included in calculations, isolation appears to be decreasing for all racial/ethnic groups. The reporting change has a weaker, but still significant, effect on certain dimensions of racial/ethnic dissimilarity. Conclusions In this study, we found that the NCES CCD's rec