GWEN IFILL: Let’s start with definitions. What do you mean when you say economic segregation?

PAUL TAYLOR: For the purpose of this analysis, we divided all households in the country into three groups, lower, middle and upper.

We define the lower group as having less than two-thirds of the median and the upper group has having more than double the median. In real dollars, for the nation, that works out to about $34,000. If you have less than that, you’re in the lower group — $104,000, if you have more than that, you’re in the upper group. In between, you’re in the middle.

GWEN IFILL: And how has that gap grown wider?

PAUL TAYLOR: What has happened is that the middle has shrunk.

There are fewer people in that middle segment now than there used to be. And that’s a large reason that we see more income segregation. The middle has hollowed out. We know that not just from this report. We know that from following census data over decades now.

So, the question is, because there are fewer people in the middle, has it affected where we live and with whom we live and what our neighborhoods look like? And the answer from this report is yes.

GWEN IFILL: So where does this happen? Where does this play out the most?

PAUL TAYLOR: It’s really fascinating, because while it has happened all across the country, there are pretty stark differences by metro areas and by regions.

The three cities that have seen by far the greatest growth in income segregation are all in Texas, the three big cities of Texas, Houston, Dallas, and San Antonio. They are now — among the nation’s 30 largest metro areas, they have the highest levels of income segregation.

GWEN IFILL: What is it about Texas?

PAUL TAYLOR: You know, I don’t know. But if you think about what has happened to those three cities in the last 30 years, they have grown a lot and they have grown a lot as a result of two different kinds of migration streams.

There has been a migration stream that it’s been attractive because they’re economically vibrant. So you have high-end workers. You have — maybe in Sunbelt areas, you have some well-to-do retirees. These folks might be attracted to living in more well-to-do neighborhoods.

You have a tremendous amount of low-end, low-skill migration, typically from south of the border, to build the new houses, to mow the lawns, to do the lower-end service jobs. So that has been a description of the changing demography of those cities. And it seems to play out in the changing residential patterns as well.

Residential segregation by income has increased during the past three decades across the United States and in 27 of the nation’s 30 largest major metropolitan areas, according to a new analysis of census tract and household income data by the Pew Research Center.

The analysis finds that 28% of lower-income households in 2010 were located in a majority lower-income census tract, up from 23% in 1980, and that 18% of upper- income households were located in a majority upper-income census tract, up from 9% in 1980.

These increases are related to the long-term rise in income inequality, which has led to a shrinkage in the share of neighborhoods across the United States that are predominantly middle class or mixed income — to 76% in 2010, down from 85% in 1980 — and a rise in the shares that are majority lower income (18% in 2010, up from 12% in 1980) and majority upper income (6% in 2010, up from 3% in 1980).

Despite the long-term rise in residential segregation by income, it remains less pervasive than residential segregation by race, even though black-white segregation has been falling for several decades.

The Pew Research analysis also finds significant differences among the nation’s 10 most populous metropolitan areas in the patterns and degree of residential segregation by income.

About six-in-ten Americans say they like the idea of living in politically, racially, religiously or economically mixed communities, while about a quarter take the opposite view: They would rather live in communities made up mostly of people like themselves. The rest say they have no strong opinion on the issue, according to a new nationwide Pew Research Center Social & Demographic Trends survey.

This preference for diverse communities is greater among Democrats, liberals, college graduates, blacks, and secular Americans than it is among the population as a whole. But virtually all major groups, at least to some degree, choose diversity over homogeneity when asked where they would like to live.

Despite these pro-diversity attitudes, however, American communities appear to have grown more politically and economically homogenous in recent decades, according to analyses of election returns and U.S. Census data.

Most notably, nearly half (48%) of all votes for president in 2008 were cast in counties that went either for Barack Obama or for John McCain by a margin of at least 20 percentage points. Back in 1976, only 27% of all voters lived in such “landslide counties,” according to figures compiled by Bill Bishop and Robert Cushing, authors of “The Big Sort,” a book which argues that Americans are clustering into politically like-minded enclaves.¹

It’s not clear whether this residential political polarization is a result of explicit decisions by people who move or a byproduct of other demographic trends. But either way, the pattern runs contrary to the attitudes expressed by a sizable majority of the public in the Pew Research Center telephone survey, which was conducted from Oct. 3 through Oct. 19 among a nationally representative sample of 2,260 adults. Survey respondents were asked if they would rather live in a community where most people share their own political views or one with a mix of different political views. Similar questions were posed about attitudes toward racial, religious and socioeconomic diversity, and about the public’s receptivity to living among immigrants. The survey’s key findings:

Politics

Some 63% of Americans say they would rather live in a community where people have a mix of political views; while 24% say they would prefer to live mainly among people who share their political views. The groups most inclined to say they prefer to live in politically like-minded communities include Republicans generally (35%), conservative Republicans (43%) and Hispanics (35%). The groups most inclined to say they prefer politically diverse communities include blacks (69%), independents (74%) and people earning more than $100,000 a year (72%).

Rural residents are a bit less inclined than those who live in cities, suburbs or small towns to say they favor politically diverse communities.

Likewise, people who live in “landslide counties” are less favorably disposed to politically diverse communities than are other Americans.

For example, while just 35% of all Republicans say they would prefer living in a politically homogeneous community, 44% of Republicans who live in a county that voted 60% or more for the GOP presidential candidate in both 2008 and 2004 say they would prefer to live in a politically homogenous community. The pattern is similar on the Democratic side. Among all Democrats, just 24% say they would rather live in a politically homogeneous community. But among Democrats who live in a county that voted for the Democratic presidential candidate by at least 60% in 2004 and 2008, some 31% say they prefer living among people who share their political views.

These findings raise some intriguing chicken-and-egg questions. Do people who are less enamored of diversity gravitate to communities where they know they will be among neighbors who share their political views? Or, alternatively, does living in a politically homogeneous community diminish a person’s tolerance for diversity? Or might there be other explanations for these correlations?

The data are inconclusive on these questions. Comparing the attitudes of respondents who have moved into a landslide county in the past decade with those who have lived in such communities most or all their lives, one finds no statistically significant difference between the attitudes of each group toward community diversity.

The survey also shows that, broadly speaking, public attitudes about political diversity in a community line up closely with attitudes about racial, religious and socioeconomic diversity. A rundown of these other findings:

Race

Some 65% of Americans say they would rather live in a community composed of a mix of racial and ethnic groups, while 20% say they would rather live in a community made up mainly of people who are the same race as they are. Democrats say they prefer racial diversity over homogeneity by 73%-15%; Republicans say they prefer racial diversity over homogeneity by a smaller margin — 52%-32%. Blacks say they prefer racially and ethnically mixed communities to mainly black communities by a lopsided 83%-to-9% margin. Whites say they prefer racially and ethnically mixed communities over mainly white communities by a smaller, but still sizable margin of 60% to 23%. Hispanics fall in between — 69% prefer racial and ethnic diversity in their communities, while 21% would rather live in a place where most people are the same race and ethnicity as them.

There are also significant differences on this question by age, with the preference for homogeneity greater among older adults than young ones. And geographically, the Midwest stands out as the region where there is the least support for racially diverse community — though even among Midwesterners, diversity is preferred over homogeneity by a margin of more than two-to-one.

Religion

Some 59% of Americans say they would rather live in a community where there are many people with different religions, while 25% say they would prefer to live mainly among people who are the same religion as they are. Some 40% of white evangelicals, 42% of Hispanic Catholics, 38% of those who attend religious services at least weekly and 41% of conservative Republicans say they would rather live in communities mainly filled with people who share their religion. By contrast, just 14% of liberal Democrats, 20% of all Democrats, 14% of those who seldom or never attend religious services, 18% of college graduates and 21% of white non-Hispanic Catholics say they would prefer to live mainly among people who are the same religion as they are.

Socioeconomic status

Some 61% of Americans say they would rather live in a community which is home to a mix of people in the upper, middle and lower classes, while 32% say they would rather live in a place where most people are the same social and economic class as they are. Liberal Democrats prefer diversity on this front by a margin of 71%-24%; conservative Republicans by a narrower margin of 59%-34%. The preference for living in an economically diverse community rises with rising socioeconomic status.

Some 66% of college graduates say they would rather live in socially and economically mixed communities; compared with just 56% of those who have a high school degree or less. Similarly, 68% of those who earn $100,000 a year or more opt for diversity, compared with just 55% of those who earn less than $30,000 a year.

Immigrants

This question was not worded the same as the others: Respondents were asked if they preferred living in a community with a large or small immigrant population. Perhaps because of the wording difference (necessitated by the fact that just one-in-eight Americans are immigrants), this is the only question in the battery to which a minority of respondents gave the pro-diversity answer. By 56%-24%, Americans say they prefer living in a community with a small rather than large immigrant population. Groups in which the smallest shares of people say they would want to live among a lot of immigrants include those over age 65 (12%), those who live in rural areas (12%), and conservative Republicans (14%). Groups most receptive to living in communities with a lot of immigrants include liberal Democrats (44%), Hispanics (40%), city dwellers (33%), Westerners (33%) and college graduates (32%).

Group Differences

Looking at the responses by various groups to this full battery of questions, one finds that the largest gaps in attitudes are by partisanship and ideology. Some 58% of Democrats, compared with just 35% of Republicans, give the diversity-oriented response to at least four of the five questions. Combining ideology with partisanship produces an even larger gap; 65% of liberal Democrats, but just 32% of conservative Republicans, give the pro-diversity responses to four of the five questions. Self-described independents are much closer to Democrats than to Republicans in their views; 57% of independents give the pro-diversity response to at least four of the five questions.

Educational attainment is another characteristic that divides respondents to these questions. Some 62% of college graduates, compared with just 46% of those with no more than a high school degree, give the pro-diversity response to at least four of the five questions.

There are smaller group differences by age, race and church attendance. Some 59% of blacks, compared with 50% of whites and 46% of Hispanics give the pro-diversity response to at least four of the five questions. Some 55% of those who seldom or never attend religious services, compared with 44% of those who attend weekly or more, give the pro-diversity responses to at least four of the five questions. And 53% of those ages 18 to 64, compared with 41% of those ages 65 and older, give the pro-diversity response to at least four of the five questions.

The Facts on the Ground

All of these survey findings raise an obvious question: Is the public’s generally strong preference for diverse communities to be taken at face value, or might it be based in part on respondents choosing the answers they deem to be socially desirable?

While it’s impossible to know for sure, it is worth noting that residential segregation has long been a fact of life in America, and that, at least by certain measures, some varieties of segregation appear to be on the rise.

The most pervasive forms of residential segregation, by far, are racial and ethnic. According to the U.S. Census Bureau, while blacks comprise just 12% of the U.S. population, about half of all blacks in 2000 lived in majority-black neighborhoods.² Hispanics also tend to be clustered into segregated enclaves, though not quite to the same degree as are blacks. Latinos made up 12.5% of the U.S. population and 43% lived in majority-Latino neighborhoods as of 2000. (The table below ranks the nation’s most populous multi-racial and multi-ethnic metropolitan areas by their levels of black/non-black and Hispanic/non-Hispanic residential segregation.)

Trends in residential segregation have been mixed over the past several decades. Black/white segregation has declined significantly since 1960, when fully 70% of blacks lived in majority black neighborhoods.³ But immigrant segregation as well as Hispanic and Asian segregation has increased in recent decades,⁴ as has overall economic segregation. From 1970 to 2000, there was a 32% increase in the residential separation of high-income Americans (those in the top income quintile) from all other Americans, according to one analysis of Census data.⁵ Even with this increasing spatial isolation of the well-to-do, however, blacks are still nearly three times as segregated from whites as are affluent Americans from those who are less well off.

What about trends in political segregation? The very durability of the buzzwords “Red America” and “Blue America” is a testament to a fairly entrenched political map in which most of the nation’s 50 states have voted reliably for one party or the other (at least in presidential contests) for many decades. And, as Bishop and Cushing have demonstrated, this political polarization by geography increasingly now extends down to the county level as well.

However, a note of caution is in order: counties and communities are not one and the same. Large counties can contain scores of communities, many with differing demographic and political profiles. Thus, the increasingly polarized voting patterns at the county level, while consistent with a trend toward self-selection into politically like-minded communities, is not proof that such a trend exists.

Moreover, not even Bishop says he believes that people choose where to live in order to be among neighbors who vote as they do. Rather, he theorizes that they choose to live among people who share their backgrounds, tastes and lifestyles — and that these likenesses, increasingly, correlate with like-minded political views.⁶

There’s some evidence in support of this “big sort” theory from recent election returns and Census data – but very little from the new Pew Research survey. It may just be that, when the subject is community diversity, Americans talk one way but behave another.

Find survey methodology, topline questionnaire and acknowledgements at socialtrends.org.

Notes

¹ The share of the vote cast in “landslide counties” grew from 27% in 1976 to 38% in 1992 to 45% in 2000 to 48% in 2004 and again in 2008, according to Bishop. But this 48% figure for votes cast in landslide counties in 2008, while high by modern standards, is not unusual by historic norms. For example, in the landslide elections of 1972, 1964 and 1936, more than half of all votes were cast in a “landslide county.” This metric of political polarization is a product of two very different components – the overall competitiveness of a given presidential campaign (which can vary widely from one election to the next) and the political dissimilarities between counties (where trends are more stable and change is more gradual). For more, see this article by Bill Bishop: http://www.slate.com/blogs/blogs/bigsort/default.aspx Also, see Klinkner, Philip A. and Ann Hapanowicz, “Red and Blue Déjà Vu: Measuring Political Polarization in the 2004 Election.” 2005. The Forum. Volume 3, Issue 2.

² In these analyses, a “neighborhood” is a census tract — a geographic unit which generally has about 4,000 inhabitants and, when first delineated, is designed to be relatively homogeneous with respect to population characteristics, economic conditions and political boundaries.

³ Glaeser, Edward L. and Jacob L. Vigdor. 2001. Racial Segregation in the 2000 Census: Promising News. Washington, DC: The Brookings Institution.

⁴ Logan, John R., and Brian J. Stults and Reynolds Farley. 2004. “Segregation of Minorities in the Metropolis: Two Decades of Change,” Demography, vol. 41 (1), pages 1-22.

⁵ Fischer, Claude S., Gretchen Stockmayer, Jon Stiles, and Michael Hout. 2004. “Distinguishing the Geographic Levels and Social Dimensions of U.S. Metropolitan Segregation, 1960-2000,” Demography, vol. 41 (1), pages 37-59.

⁶ Interview with author, Nov. 14, 2008.

The 5-4 decision by the U.S. Supreme Court in June to strike down school desegregation plans in Seattle and Louisville has focused public attention on the degree of racial and ethnic integration in the nation’s 93,845 public schools. A new analysis of public school enrollment data by the Pew Hispanic Center finds that in the dozen years from 1993-94 to 2005-06, white students became less isolated from minority students while, at the same time, black and Hispanic students became slightly more isolated from white students.

These two seemingly contradictory trends stem mainly from the same powerful demographic shift that took place during this period: an increase of more than 55% in the Hispanic slice of the public school population. Latinos in 2005-06 accounted for 19.8% of all public school students, up from 12.7% in 1993-94.1 During this same period, the black share of public school enrollment rose slightly — to 17.2%, from 16.5% — while the white share fell sharply, to 57.1% from 66.1%.

In part because whites now comprise a smaller share of students in the public schools, white students are now more likely to be exposed to minority students. In 1993-94, fully one-third (34%) of all white students attended a nearly all-white school (this report defines a school as “nearly all-white” if fewer than 5% of the students are non-white). By 2005-06, just one in five white students (21%) was attending a nearly all-white school. The number of nearly all-white public schools fell by 35%, from 25,603 in 1993-94 to 16,769 in 2005-06.

But even as the decrease in the white share of the public school population has led to a greater exposure of white students to minority students, it has also led to a diminished exposure of black and Hispanic students to white students. Roughly three-in-ten Hispanic (29%) and black (31%) students attended schools in 2005-06 that were nearly all-minority (by this report’s definition, a “nearly all-minority” school is one in which fewer than 5% of the students are white), and these percentages were both somewhat higher than they had been in 1993-94, when they stood at 25% for Hispanic students and 28% for black students. The number of nearly all-minority public schools almost doubled in this time period, increasing from 5,498 in 1993-94 to 10,135 in 2005-06.

The growing exposure of white students to minority students is largely due to the growth in Hispanic enrollments, which in turn is related to the sharp increase in the Hispanic share of the nation’s population. In 1993-94, 74% of white students attended schools in which fewer than 5% of the students were Hispanic. By 2005-06, just 58% of whites attended schools with that very low Hispanic share of enrollment.

Not only are black and Hispanic students similarly isolated from white students, they also tend to be isolated from one another. In 2005-06, 56% of Hispanic students attended a majority-Latino public school (a school in which at least half of the students are Hispanic). These majority-Latino public schools educated just 7% of the nation’s black students. Similarly, the nation’s majority-black schools, which educate nearly 50% of black students, educate just 4% of the nation’s Hispanic students.

To be sure, levels of racial and ethnic segregation and integration in the public schools are affected by factors other than the demographic changes in the school population at large. In particular, they are affected by the geographic dispersion of racial and ethnic groups; by local residential housing patterns; and by desegregation policies at the school district level. This report does not look at those factors, nor does it attempt to determine which factor has had the greatest impact on the changing patterns of integration and segregation in the public schools since 1993-94. Rather, it simply tracks the changes over a 12-year period in the levels of racial and ethnic isolation and exposure in public schools among black, white, Hispanic and Asian students.