An examination of the available statistics to determine the degree to which socioeconomic factors, rather than racial prejudice, determine observed nonwhite residential segregation in the Chicago and Detroit metropolitan areas. Multiple regression techniques are used to test a socioeconomic model, and intensive examination of the results reveals that only a fraction of observed segregation can be "explained" by socioeconomic factors. Additional tests provide evidence of prejudice in the white community--either actual or anticipated--as the major constraining force in the "unexplained" segregation. The implication is that efforts to improve nonwhite socioeconomic conditions must be continued and that specific antisegregation programs have a role to play. 239 pp. Bibliog.