Midwest Social Sciences Journal


The index of dissimilarity is the most widely used method for measuring racial segregation. When applied to Indianapolis, this index has returned results showing the city to be among the most segregated in the country. The resulting measure, however, suffers from two shortcomings. First, the index of dissimilarity is sensitive to the census-defined geographic unit chosen for the analysis; thus, this index returns different (though proportionate) results depending on whether the population data are aggregated to larger or smaller enumeration units. Second, the index of dissimilarity cannot account for the influence of spatial proximity; adjacent census blocks interact regardless of administrative boundaries. In place of the index of dissimilarity, we apply the segregation index in order to treat the phenomena as a surface that is simultaneously smooth and continuous. In this article, we calculate the segregation index for Indianapolis from 1990 to 2010 using the kernel density estimation method. The results of the analysis are presented in three pairs of decennial maps. These maps add to the understanding of residential segregation by resolving in a statistically reliable manner the phenomenon’s geographic component. Our visualization of segregation confirms its presence in distinct clusters, its growth over time, and a strong bias of this growth to be contiguous. In a manner akin to examinations of residential segregation’s impact on education attainment and health outcomes, careful description of segregation’s spatial aspect leads to a more nuanced understanding of phenomenon’s pervasiveness across social life.