Skip Main Navigation Skip To Navigation Content
 Housing Patterns

Chapter 7

Cross-Group Analyses

The previous four chapters discussed residential segregation within metropolitan areas separately for the four major racial and ethnic minority groups in the United States -- American Indians and Alaska Natives, Asians and Pacific Islanders, Blacks or African Americans, and Hispanics or Latinos. This chapter examines these groups together and makes comparisons across them in the extent and patterns of segregation.

Because the size of these populations vary, as does their geographic distribution, this chapter examines the five residential segregation indexes by focusing mainly on all metropolitan areas, rather than the "selected" areas that were highlighted in the previous chapters. As a basis of comparison, Tables 7-1 and 7-2 include descriptive statistics both for all metropolitan areas and for the selected metropolitan areas from chapters 3-6.(1)

It is clear from Table 7-1 that Blacks were the most residentially segregated of the four groups examined. They had the highest mean index score for all metropolitan areas for all five indexes for all three censuses. They also had the highest index score for selected metropolitan areas for all five indexes for all three censuses, with only one exception (spatial proximity for American Indians and Alaska Natives for 1990).

Hispanics were the second-most segregated group, with the second-highest index score for all metropolitan areas for all five indexes for all 3 years, with the same exception as that for Blacks. Similarly, they had the second-highest index score for selected metropolitan areas for all five indexes for all 3 years, this time with only two exceptions: spatial proximity for American Indians and Alaska Natives for 1980, and Blacks with the second highest spatial proximity in 1990.

Asians and Pacific Islanders were more residentially segregated than American Indians and Alaska Natives, as measured by four of the five indexes for all years for both all metropolitan areas and selected metropolitan areas. The one exception, again, was spatial proximity.

As noted in Chapter 4, in 2000, the residential segregation indexes for Asians were close to those for Asians and Pacific Islanders (not surprising as they make up the vast majority of the combined group), while the indexes for Native Hawaiians and Other Pacific Islanders (also referred to as Pacific Islanders) tended to be somewhat lower (though not universally so). When all metropolitan areas are considered, both Asians and Pacific Islanders were more segregated than American Indians and Alaska Natives for three of the five indexes -- dissimilarity, isolation, and delta -- while Pacific Islanders were less segregated than American Indians and Alaska Natives for the other two indexes -- absolute centralization and spatial proximity.

The distribution of index values is presented in a histogram for all metropolitan areas for 1980, 1990, and 2000 in Figures 7-1(a-e), 7-2(a-e), and 7-3(a-e), respectively. These generally confirm the findings described above. The dissimilarity index shows the same ordering from most to least segregated: Blacks, Hispanics, Asians and Pacific Islanders, American Indians and Alaska Natives. This conclusion also holds, though not as clearly because of their relatively narrower distributions, for the isolation index and the spatial proximity index.

Table 7-2 displays the percentage changes in the weighted averages for each index over each decade. No index shows a uniform pattern for all groups. Of the five dimensions examined, declines in segregation were most evident in centralization (absolute centralization), where all groups experienced declines over the 1980 to 2000 period when all metropolitan areas were considered (declines were not registered for American Indians and Alaska Natives in their 13 selected metropolitan areas). Three of the four groups experienced declines in concentration (delta) when all metropolitan areas are considered -- Asians and Pacific Islanders, who experienced no change, are the exception. Trends for the evenness (dissimilarity) and clustering (spatial proximity) dimensions were split, with two racial/ethnic groups experiencing increases and two experiencing declines. Finally, exposure (isolation) was the one dimension where increases predominated, with only African Americans experiencing declines. Because the isolation index is sensitive to the overall size of the minority group, it is unsurprising that this index showed the greatest increases as the population of all of the minority groups grew substantially over the 1980-2000 period.

Figure 7-4(a-e) shows the change in the indexes between 1980 and 2000 for all "selected" metropolitan areas graphically (use of all metropolitan areas for such a figure would not give the reader much information because the extent of the overlap obscures too many individual points). Not only do the figures confirm the findings from Table 7-1 (Blacks tend to be more highly segregated than other groups), they also show groups that experienced changes in the indexes from 1980 to 2000 (as shown in Table 7-2).

It is again clear that Blacks or African Americans had the biggest declines in dissimilarity (Figure 7-4a) and delta (Figure 7-4c) and almost all the declines in isolation (Figure 7-4b), since those points lie below the 45-degree line. The other groups are clustered around the 45-degree line for dissimilarity and delta, indicating little change over the 20-year period, with perhaps a preponderance of increases for Hispanics for dissimilarity. American Indians and Alaska Natives experienced increases in isolation for low levels of that index and decreases for higher levels; the other two groups -- Asians and Pacific Islanders, as well as Hispanics -- experienced increases in segregation using the isolation index (and Hispanics had the four highest isolation scores for 2000).

The absolute centralization index (Figure 7-4d) does not show much change from 1980 to 2000, with most of the points scattered fairly closely around the 45-degree line. Spatial proximity (Figure 7-4e) also clusters around the 45-degree line, indicating little change. Blacks experienced a decrease in this index between 1980 and 2000, while Hispanics experienced an increase in this index.

Only one metropolitan area, Los Angeles-Long Beach, had sufficient minority group population in 1980 to qualify as a selected metropolitan area in chapters 3 through 6, which analyzed the groups individually. For that reason, we thought it worthwhile to present those data in Table 7-3 and discuss their consistency with the general findings stated above. Blacks and Hispanics were the two most segregated groups in Los Angeles, with Blacks more segregated than Hispanics according to most measures in most years. Hispanics were, however, more isolated than Blacks in 1990 and 2000, and were tied in their absolute centralization index in 2000.

In all 3 years, Asians and Pacific Islanders were more segregated in Los Angeles than American Indians and Alaska Natives (tied for one measure, dissimilarity, in 2000) but less segregated than Hispanics or Blacks. In 2000, Asians were also more segregated in Los Angeles than American Indians and Alaska Natives (and tied for dissimilarity), but less segregated than Hispanics or Blacks. Pacific Islanders had a higher dissimilarity index in Los Angeles than either Asians or American Indians and Alaska Natives, were tied with Asians for delta (both higher than American Indians and Alaska Natives), but were the least segregated group using the other three segregation measures (and substantially less isolated, a function of their relatively small size). So, in general, the patterns in Los Angeles-Long Beach seem to mirror the patterns in the Nation as a whole.

Table 7-4 presents the percentage changes in residential segregation over the 1980-2000 period for Los Angeles-Long Beach. It should be noted that while there was a decline of 6 percent in the number of American Indians and Alaska Natives between 1980 and 1990, there was a tremendous growth between 1990 and 2000 (a tripling of the population). Much of this growth is attributable to multiple-race identification, as only 76,988 people identified as American Indian and Alaska Native only, whereas 138,696 identified either as American Indian and Alaska Native alone or in combination with at least one other race.

For the other three minority groups, there were the expected changes -- decreases in dissimilarity and spatial proximity for Blacks and increases in isolation for Hispanics and for Asians and Pacific Islanders. Delta and absolute centralization, however, showed declines in segregation for all groups (with the exception of no change for American Indians and Alaska Natives).

Figure 7-5(a-d) presents a map of the population distribution of all four groups, contrasted with the settlement pattern of non-Hispanic Whites.

To recap, a comparison of segregation patterns of the different groups indicated that African Americans were the most segregated (vis-a-vis the reference group, non-Hispanic Whites). Hispanics or Latinos were generally the next most highly segregated, followed by Asians and Pacific Islanders, and then American Indians and Alaska Natives across a majority of the measures. However, African Americans experienced declines, albeit modest ones, in segregation across all dimensions, while other groups showed either mixed patterns or small increases over the 1980-2000 period.


1. As discussed in chapter 2, the set of indexes averaged across all metropolitan data suffers from the weakness of including metropolitan areas where the minority group is so small that segregation estimates may appear peculiar or anomalous. This weakness is at least partly compensated for by computing averages based on weights in all tables, where the weights are the number of the minority group in question present in the metropolitan area.

Chapter 7:
Go to Appendix A: Measuring Segregation Using Alternative Methods of Identifying Racial/Ethnic Group Members
Go to "Racial and Ethnic Residential Segregation in the United States: 1980-2000" in HTML format

Contact the HHES Information area at 301-763-3242 or visit for further information on Housing Patterns Data.

Source: U.S. Census Bureau, Housing and Household Economic Statistics Division
Last Revised: September 30, 2011