A Demographic Approach to Detecting Racial Gerrymanders in U.S. Congressional Districts

The practice of racially gerrymandering political districts entails segregating one racial group from others into and out of individual districts. This study measures the racial segregation of black people in U.S. Congressional Districts using the Areal Unit Segregation Measure (abbreviated S +/− ), a statistic introduced to the literature recently. S +/− measures how much a district packs or cracks some racial group by determining whether a district contains an unusually greater (or unusually lesser) share of black people than the share of black people among a district’s residents’ nearest N neighbors (where values of N equal the number of people in a state’s Congressional Districts). Since S +/− is a new measure, the main goal of this paper is to interpret values of S +/− . To do this, S +/− values are generated for all U.S. Congressional Districts drawn after each census from 1990 to 2020 and for high-profile court cases in which judges overturned Congressional Districts that unduly segregated black voters into and out of districts. S +/− values of actual districts are also compared with samples of computer-generated maps to determine when a 2022 district’s S +/− is statistically unusual. Collectively, these results clearly indicate that when a district’s S +/− value is about 18 or more points from zero a district is unduly packed or cracked.