Please bear with us… We shall attempt to explain each measure first and then explain why we shall use two of these in the analyses and presentations that follow.
Dissimilarity. If a racial minority is segregated from the majority in a spatial sense, how many people would have to move to mix the two populations evenly? This number is expressed as a proportion, from 0 to 1. If the value is zero, nobody would have to move. If it is 1 all would have to move. Dissimilarity measures this "need to move" to achieve desegregation. If the values on the left axis are multiplied by 100, the percent that would have to move is given. In the case of Detroit, in this graphic, the value for dissimilarity is.874 — suggesting that 87.4% of the African Americans in Detroit would have to move to achieve complete desegregation.
Isolation. This is another spatial measure. It shows the relationship between minority and majority group members in an area. If the index is 1 (i.e., 100%), all the people are of the minority. A number like 0.823 (Detroit's "isolation" index) means that 82.3 percent of the minority population lives with 17.7% of the majority population. The higher the value, the fewer whites are present, in the data shown here — hence the more "isolated" the minority population is. The lowest value in this display is for Atlanta, 0.667, suggesting that in Atlanta one third of residents are of the majority population.
Both Dissimilarity and Isolation implicitly assume the value of a desegregated society rather than, say, a multicultural society.
Centralization. This index measures how close to the central business district (CBD) the minority population is located. A high value means a strong aggregation of the minority population in the city's center. In New York and in Chicago this value is low for blacks, no doubt indicating the high value of real estate in Lower Manhattan and on Chicago's lakefront (which is surrounded by enclaves with names like Greektown, Chinatown, and Ukrainian Village). It is high in Detroit where redevelopment of central areas has been sporadic and isolated. The measure is inapplicable to cities with no well-defined CBD.
Concentration. This index measures density relative to the white population. It measures how small an area a minority occupies relative to the majority population, all other things equal (normalized for population size). Relative concentration, the measure used here, means that if the average density for the minority is the same as for the majority, the index is zero. The higher the index, the more densely settled the minority is. If the measure drops below zero, it means that the majority is more densely settled, the minority more dispersed (often the case with Indians, as we shall see). The upper bound of this measure is 1 — maximum density. The lower bound, expressed in negative numbers, can be quite high (negative), with populations spread thinly over an area. The underlying assumption is that a one-acre lot is better than a two-room flat for a family of five. It's difficult to argue.
Clustering. This indicator is high if minority neighborhoods are adjoining. The more neighborhoods are clustered, without intervening white neighborhoods, the higher the index. We show "relative clustering," a measure based on the average distance between minority members compared with the average distance between majority members. High values indicate close clustering of minority enclaves. Negative values signal dispersion of settlements. Values plotted are on the right scale. High concentration combined with high clustering suggests dense, adjoining neighborhoods. If concentration is high but clustering low, there are many dense neighborhoods but separated by majority settlements.
In the following panels, we shall use "relative concentration" as our measure of choice, illuminated, where meaningful, by the isolation measure. Concentration has a certain objective quality. Minorities might wish to live together. How densely must they live to do so — that is the issue. It is also useful for showing situations where minorities are dispersed. We shall encounter such cases. Isolation is one of the traditional measures of segregation. It shows the degree to which races and ethnicities are in contact with the majority white population.
Source: U.S. Bureau of the Census. Housing and Household and Economic Statistics Division. Online. Available: census.gov/hhes.www.housing/resseg/def.htm. Presentation is from the work of Roderick J. Harrison and Daniel H. Weinberg. Racial and Ethnic Residential Segregation: 1990.
User Comments Add a comment…