Segregation, Residential

Residential segregation most often refers to the separation of people into different neighborhoods based on their race and/or ethnicity, though neighborhood-level segregation can be computed using other characteristics of people, households, or neighborhoods, for example, social class, single-parent families, or home ownership. The association between the term residential segregation and race/ethnicity reflects the fact that segregation by race/ethnicity is far higher than segregation by other characteristics. Segregation is a social problem because where people live determines [Page 828]other aspects of their lives: what schools their children attend, access to transportation, availability of municipal services, job opportunities, and so on.

Since residential segregation is based on neighborhoods, it is important to understand how neighborhood is defined. Neighborhoods are usually approximated by census tracts, non-overlapping areas that cover an entire city or metropolitan area and contain approximately 4,000 persons. With the exception of subdivisions to accommodate population growth, tracts have the advantage of changing relatively little over time, making comparisons across time easier. Researchers who argue tracts are too large to capture what most people mean by a neighborhood instead use block groups, areas of about 1,500 people, as proxies for a neighborhood, though these get redefined at each census. The larger the unit used to define the neighborhood, the lower segregation will be, because the probability of finding two different people, say a black and a white, increases as the total number of people examined increases.

In its simplest form, residential segregation is measured by the Index of Dissimilarity. This compares the population of each neighborhood to the overall racial composition of the city or metropolitan area in which it is located. Neighborhoods whose population more closely matches the overall racial composition contribute less to the index, those that do not contribute more. The index ranges from 0 to 100, where zero indicates no segregation and 100 indicates total segregation, with no groups sharing any neighborhoods at all. Values below 30 are considered low, those between 30 and 60 moderate, and above 60 high. The value of the index can be interpreted as the percentage of either group that would need to change their neighborhood so that each neighborhood would reflect the race/ethnic distribution of the city as a whole. The advantages of this index are that it is easy to compute and understand, takes the city's overall racial composition into account and does not impose an artificial definition of integration, is unaffected by the relative sizes of the groups, and has a long history of use. The main disadvantage of the index is that it compares only two groups at a time, a feature that is becoming more problematic as U.S. society becomes increasingly diverse.

Using the Index of Dissimilarity as the measure allows description of patterns of segregation for different groups in the United States in recent decades. First, African Americans are almost always the most segregated group, followed by Hispanics, and then Asians. In 2000, on average, African American segregation from non-Hispanic whites was 65.2, Hispanic segregation was 51.6, and Asian segregation 42.2. Second, the segregation of African Americans has been declining in recent decades, but that of Hispanics and Asians has been increasing, in part because of continued immigration of those groups; new immigrants tend to live near others of their group, thus replacing those who have moved out. On average, while black-white segregation declined 4.6 points between 1980 and 1990, Hispanic-white segregation increased 4.9 points, and Asian-white increased 2.1 points. Third, segregation tends to be highest in older, formerly industrial cities, in the Northeast and Midwest, with large black populations. For example, in 2000 the segregation of African Americans from whites in Detroit was 84.7, compared to segregation of 45.7 for Hispanics and 45.9 for Asians. Conversely, it is lower in newer cities in the South and West whose populations are growing. For example, in 2000 the segregation of African Americans in Phoenix-Mesa was 43.7, compared to 52.5 for Hispanics and 28.1 for Asians. Metropolitan areas that attract more immigrants also tend to have somewhat lower black segregation than those that do not. Fourth, the largest declines in black segregation are in smaller metropolitan areas with small black populations. Between 1980 and 2000, African American segregation declined by nearly 20 points in Dallas and Phoenix-Mesa, while it declined by less than 5 in Detroit and Newark, and increased by 0.2 points in New York City. Thus, in the larger metropolitan areas where more than half of metropolitan blacks live, segregation declines have been modest.

...