Assessing disparity using measures of racial and educational isolation

Abstract

Background: We previously developed a local, spatial measure of racial isolation (RI). Here, we extend the isolation concept to educational attainment, which has been consistently linked to health and wellbeing.Methods: Using 2010 Census data, we develop a local, spatial measure of educational isolation (EI) for census tracts across the continental United States (US). We also calculate RI at census tracts across the continental US. We develop a novel measure of local correlation to understand how the relationship between RI and EI varies across space. We also evaluate how local RI-EI correlations differ by urbanicity.Results: Patterns of EI and RI differ substantially across the US. EI tends to be lower (higher) in urban (rural) areas. In contrast, RI is relatively low across much of the US but is high across the Southeast, in cities along and east of the Mississippi River, and in West Coast cities.The global EI-RI correlation (0.21) masks substantial local variations in heterogeneity. In the South, there are strong local RI-EI correlations: high RI tracts in the South are often also educationally isolated. High RI-EI correlations are observed elsewhere (e.g., Northeast, West Coast), but these high correlations more often reflect low levels of both isolation measures.We observe greater heterogeneity in EI in tracts with low RI than in tracts with high RI. This is most pronounced for suburban and rural tracts. In rural tracts with RI < 0.10, EI values range from 0.30 to 0.95; for rural tracts with RI > 0.50, corresponding EI values are > 0.75. Thus, residents of tracts with high RI are more likely to be both educationally and racially isolated.Conclusion: RI and EI are important to consider jointly given their different geographic distributions. Future work should assess the importance of these measures as predictors of health outcomes/disparities.