Hypersegregation

Identifying Hypersegregation

In an often-cited 1988 study, Douglas Massey and Nancy Denton compiled 20 existing segregation measures and identified five dimensions of residential segregation: evenness, exposure, concentration, centralization, and clustering. Evenness refers to the distribution of groups across neighborhoods according to their proportion of the population. Thus, if 10 percent of the population of a given city is Asian, then 10 percent of each neighborhood should be Asian for there to be no segregation. Exposure refers to the probability that members of a group will have any form of interaction with other groups. Concentration refers to how much space a minority group occupies; if a minority group lives primarily in a few neighborhoods, they are very concentrated. Centralization refers to how close to the center of an urban area a group lives. In many U.S. metropolitan areas, living in the suburbs (as opposed to the central city) is associated with a higher standard of living and better access to high-quality schools and amenities. Clustering refers to the extent to which a group lives in contiguous or adjoining neighborhoods.

Researchers typically use census data to measure segregation, mainly because such measures require geographically detailed population counts (such as in neighborhoods) that are collected in decennial censuses. More specifically, segregation measures usually take into account the distribution of various groups in particular neighborhoods in a metropolitan area relative to their numbers in the total metropolitan population. Neighborhoods are most often defined in terms of “census tracts,” which contain 1,500 to 8,000 people, though sometimes in smaller geographic units, such as blocks or block groups. The statistical indexes developed to measure each of these dimensions typically range from 0 to 1, with 0 indicating complete integration on a particular dimension and 1 indicating complete apartheid or segregation on a particular dimension. Massey and Denton (1989) defined hypersegregated metropolitan areas as those where indexes exceed .60 on at least four of the five dimensions of segregation.

To date, hypersegregation appears to be a phenomenon that applies almost exclusively to blacks who reside in certain U.S. metropolitan areas. Although blacks are also segregated from whites in other countries, the levels are not as high as they are in many U.S. metropolitan areas. Table 1 provides a list of U.S. metropolitan areas where blacks are hypersegregated from whites, by year. In 2000, Chicago, Cleveland, Detroit, Milwaukee, Newark, and Philadelphia had segregation scores above .60 on all five dimensions of segregation. Also worth noting are the metropolitan areas where African Americans were persistently hypersegregated for several decades: Baltimore, Buffalo, Chicago, Cleveland, Detroit, Gary, Los Angeles, New York, and St. Louis. Two metropolitan areas in 2000 also had Hispanic-white hypersegregation (Los Angeles and New York).

...