INTRODUCTION
In 2012, The Manhattan Institute’s Edward Glaeser and Jacob Vigdor released a report based on the 2010 Census, declaring that, with respect to housing, we were at “The end of the segregated century” (p. 1). This conclusion was based largely on two empirical observations: levels of residential segregation of Blacks had moved steadily downward, and all White neighborhoods had become “effectively extinct” (Glaeser and Vigdor, Reference Glaeser and Vigdor2012, p. ii). The reaction from researchers and advocates was swift and heated, with objections ranging from technical details of the analysis to its interpretation and implications. Indeed, using the same data, another pair of researchers, John Logan and Brian Stults (Reference Logan and Stults2011) drew a different conclusion, naming their report: “The Persistence of Segregation in the Metropolis.” In addition to interpretational differences, one of their primary objections to the Glaeser and Vigdor analysis was how segregation was calculated: although they used standard measures (the index of dissimilarity and isolation index), they did not calculate Black segregation from Whites, as is generally done, but instead they compared Black residential locations to all other racial/ethnic groups combined.
Among other concerns with this approach is that, if one wants to understand what has ‘replaced’ segregation, it is critical to know what the comparison group is. By examining Black segregation from all other groups combined, Glaeser andVigdor (Reference Glaeser and Vigdor2012) are, by default, identifying as ‘integrated’ a great variety of neighborhoods, those: where Blacks and Whites are living together; where Blacks and Hispanics share neighborhoods; where Asians and Blacks are living side by side; and where Whites, Blacks, Asians and Latinos are all sharing residential space. Richard Alba and Steven Romalewski (Reference Alba and Romalewski2012) in their response to Glaeser and Vigdor (Reference Glaeser and Vigdor2012) were explicit about the significance of this technical decision, which lead them to ask: Is all integration the same? As they write:
The residential integration of Black Americans is increasing, and that is a development to be welcomed. But we are nowhere near the end of segregation, or even of a century of segregation. New forms of residential segregation and disadvantage have developed in the wake of large-scale immigration. One especially prevalent form is that of poor, largely minority neighborhoods in which Hispanics and African Americans live side by side (Alba and Romalewski, Reference Alba and Romalewski2012).
This debate highlights that as our nation becomes more racially and ethnically diverse, there is a need to be more explicit about how we think about and operationalize the ‘opposite’ of segregation. Demographers have been working for the past decade at trying to capture and understand (and describe) the different ways in which neighborhoods are constituted with respect to their racial/ethnic composition. There are an almost dizzying array of different ways this has been done: in one study, ‘racially mixed’ is a neighborhood with fewer than 70% of a single group (e.g., Peterson and Krivo, Reference Peterson and Krivo2010); in another an integrated neighborhood is between 10–89% Black (e.g., Lee and Wood, Reference Lee and Wood1990); in others, it is some (varying) combination of minimum and maximum levels of different racial/ethnic groups (e.g., Ellen Reference Ellen2000; Friedman Reference Friedman2008; Krysan and Bader, Reference Krysan and Bader2009). The existence of a multitude of definitions is symptomatic of a general conceptual fuzziness of what is meant by integration in general, and integrated neighborhoods in particular (Sin and Krysan, Reference Sin and Krysan2015). Footnote 1
Our purpose in this paper is not to solve this problem but to draw attention to the issue and report the results of a study that highlight in a different way the importance of developing specificity in our work on integration and its related cousin, diversity. Our understanding of these ‘ideals’ needs to be grounded in real life experiences and perceptions of how they are played out rather than kept at the level of an abstraction or, at the other extreme, a question of technical definitions and indices. Specifically, our study illustrates how vastly different attitudes toward integration are, depending on the specific configuration of integration. We do this through analysis of a novel dataset in which Whites, Blacks, and Latinos in the Chicago area are asked to create their most and least desired neighborhood racial compositions; they are then asked to explain why they hold these preferences. By focusing on people who create diverse neighborhoods in each of these contexts, and examining how they talk about them, we shed light on the diversity of integration and highlight the complexity of attitudes toward it. Our aim is to move scholarly research in a direction that takes into consideration this complexity as it tries to determine the causes and consequences of neighborhood stratification or integration—in whatever form it takes.
In doing so, we seek to illustrate the diversity of integration attitudes and highlight the pitfalls of failing to imagine and specify more clearly what is meant by integration. Thus, our purpose is to explore and describe, rather than explain and test. While we do not test specific hypotheses, our analysis and interpretation are shaped by existing research on racial residential preferences and general theories of contemporary race relations and attitudes. To that end, after providing some general background on research on integrated neighborhoods, we review relevant literature that provides insights into two underlying questions: (1) Why might people like integrated neighborhoods? and (2) Why might people dislike integrated neighborhoods?
BACKGROUND
For pragmatic reasons if no other, researchers using Census and other large scale survey data must decide how to operationalize integration to do their calculations. Often these decisions are not made with an explicit conceptualization of integration; it is, instead, mostly defined as whatever segregation is not. A cursory inspection of quantitative research on the topic of residential integration reveals how our understanding of segregation and integration— both operationally and conceptually—has changed over time, implicitly if not explicitly. To take as one example, in 1970, just two years after the Fair Housing Act was passed, Norman Bradburn and colleagues (1970) in their landmark study of integrated neighborhoods called a community ‘open’ if it had two or more Black households (but fewer than 1%); moderately integrated if it was 1–10% Black; and substantially integrated if it was more than 10% Black. Two decades later, in an important article that went against the tide of research at the time, because it made the case that integrated neighborhoods could be stable, Barrett Lee and Peter B. Wood (Reference Lee and Wood1990, Reference Lee and Wood1991) defined an integrated neighborhood as one that was between 10–89% Black.
In the most recent decades, there has been an explosion of research on integrated neighborhoods, and a concomitant increase in the metrics used to define a neighborhood as integrated, in part due to the increasingly complex racial/ethnic landscape (for a synthesis of this research, see Sin and Krysan, Reference Sin and Krysan2015). For example, an integrated neighborhood is: in one case a place where there is no group greater than 70% (Peterson and Krivo, Reference Peterson and Krivo2010); in another case, it is one in which no group has a majority status (Lee et al., Reference Lee, Iceland and Sharp2012); and in another it is one that is between 30–70% Black (Quillian Reference Quillian2002). Researchers conducting ethnographies of integrated neighborhoods must choose their cases based on some criteria as well, and here the tendency of late is to study places that are multiethnic/racial—specifically, places where three and sometimes four different groups are present at more than token levels (e.g., Burke Reference Burke2012; Maly Reference Maly2005). The list could go on, but the overarching point is that there is little consistency in what constitutes integration (Sin and Krysan, Reference Sin and Krysan2015).
Notwithstanding the fact that there are multiple definitions as to what constitutes an integrated neighborhood, scholars have begun to generate valuable research that helps us make sense of contemporary neighborhoods: demographers providing us with detailed descriptions of neighborhoods and their trajectories; and ethnographers documenting how integration works, especially at the institutional or organizational levels in these communities. But less work has unpacked regular people’s attitudes toward different kinds of integrated neighborhoods. While there is a long tradition of research on attitudes toward living in neighborhoods with people of different races and ethnicities, the measures and more importantly the framing of the analysis has centered on how these attitudes contribute to segregation. The purpose of the present research is to shift the emphasis toward developing a richer picture of the related but distinct question of attitudes toward integration and integrated neighborhoods. This difference may be subtle—and we will see that many of the patterns are consistent with the long tradition of research on racial residential preferences—but by emphasizing regular people’s attitudes toward integration we reveal the pitfalls of failing in our scholarly efforts to imagine integration in all of its diverse forms.
One of the first survey questions measuring attitudes toward integrated neighborhoods was asked of Whites about Blacks in 1963: “If a Negro with the same education and income as you moved into your block would it make any difference to you?” Interested in doing a better job of understanding how people responded to different proportions of Black and White neighbors,Reynolds Farley and his colleagues (1978) introduced an innovative approach in the 1970s, providing cards that portrayed (with small houses that were shaded White or Black) different racial compositions and asking Whites and Blacks in Detroit how they would feel living in neighborhoods like those shown on the card. Time series analyses of Detroit residents from 1976 to 2004 show that Whites report being increasingly open to greater numbers of African American neighbors, though resistance to living in neighborhoods where Whites fall short of being the numerical majority remains substantial. For African Americans, the openness to living with substantial numbers of Whites (up to half) has always been strong, though there is some indication in more recent years that Blacks prefer somewhat fewer Whites in their neighborhood than in prior decades (Farley Reference Farley2011).
A major step forward occurred in the 1992 MCSUI-Los Angeles Study when Camille Charles (Reference Charles2006) built on the approach introduced by Farley and colleagues (1978) by making it more flexible and responsive to the multiethnic and multiracial city in which her study was conducted. Survey participants were asked to draw their ‘ideal neighborhood racial composition’—that is, to fill in the houses themselves with their ideal racial/ethnic mix (see Figure 1). Charles (Reference Charles2006) provides a rich and detailed picture of racial residential preferences using this multiethnic show-card technique.
Fig. 1. The Card Used to Measure Respondent’s Most and Least Desired Neighborhood Racial/Ethnic Composition.
Charles (Reference Charles2006) focused mostly on identifying which factors predict a desire to have greater numbers of one’s own racial/ethnic group in the neighborhood. Less attention was given to using this technique to understand if and why people were interested in neighborhoods that were integrated. Based on basic descriptive statistics from Charles’ measurement tool, it appears that Americans are very interested in living in integrated neighborhoods. Specifically, when asked in the 2000 General Social Survey (GSS) to construct their ideal neighborhood racial composition using this technique, Whites’ ideal neighborhood was (on average) 57% White, 17% African American, 13% Latino, and 13% Asian American. African Americans created a neighborhood where the modal—but not majority—group was African American (43%), and the second largest group was Whites (30%), with about an even split of the balance between Latinos (14%) and Asians (13%). For their part, Latinos created a neighborhood with about equally-sized Latino (33%) and White (32%) populations, and about equal percentages of African Americans (19%) and Asian Americans (16%) (Charles Reference Charles2003). But studies have not yet unpacked these results to get at questions such as: 1) Beyond these averages, what kinds of neighborhoods do people actually draw as ideal?; 2) And, importantly, what do they like about them?; 3) And when asked, what kinds of neighborhoods do these same people not want to live in? And why?
Why Might People Like Integrated Neighborhoods?
In our review of what we know about attitudes toward integration in general, we start with the observation that diversity has become a new norm in the United States. For example, Joyce Bell and Douglas Hartman (Reference Bell and Hartmann2007) observe:
Everyone in America—school administrators and business leaders, political activists, marketing gurus, and Supreme Court Justices—seems to be using the language of diversity these days…Extolling the virtues of differences, celebrating diversity as a value in itself, and describing diversity as the new cornerstone of American democratic idealism (p. 895).
Summarizing national telephone survey data they find: “nearly half of Americans believe that diversity is ‘mostly a strength’ for the country…and less than five percent see diversity as an unqualified weakness” (Bell and Hartmann, Reference Bell and Hartmann2007, p. 895). A review of the history of diversity (and its close cousin multiculturalism) is beyond the scope of this paper, but it is clear that observers across the political spectrum agree that diversity has become a value to be embraced with virtues to be extolled in American popular culture (e.g., Glazer Reference Glazer1997). In the context of residential integration, John Iceland (Reference Iceland2009) observes that diversity in neighborhoods and cities has become quite fashionable:
To a cosmopolitan person, the increasing diversity of many American metropolitan areas may be a source of stimulation. It can afford the opportunity to eat a variety of foods, observe different customs, and share in others’ celebrations, such as Cinco de Mayo or Chinese New Year (p. 2).
Iceland’s observation is consistent with others who suggest that people are enthusiastic about diversity because it is something interesting and different to be consumed and enjoyed (Bonilla Silva et al., Reference Bonilla-Silva, Forman, Lewis and Embrick2003; Gallagher Reference Gallagher2003). Therefore, one might expect that people who express a desire to live in an integrated neighborhood would explain these preferences in terms of a desire for new and interesting experiences and exposure to people who are not like themselves.
Second, although analysts who focus on diversity rhetoric frequently lament that this new language of diversity allows people to talk about race without talking about racism, racial subordination, and racial stratification (Bell and Hartmann, Reference Bell and Hartmann2007; Berrey Reference Berrey2005; Bonilla-Silva et al., Reference Bonilla-Silva, Forman, Lewis and Embrick2003; Gallagher Reference Gallagher2003), it is conceivable that when asked why they desire integrated neighborhoods people will explain their positive attitude toward integration because of the prospects it holds for improving race relations and addressing the racial inequities that have been perpetrated for decades.
A third reason people may desire integrated neighborhoods may be the perception, particularly among racial/ethnic minorities, that integrated neighborhoods will be of higher quality than segregated minority communities. Footnote 2 That is, compared to segregated minority neighborhoods, an integrated neighborhood may be perceived as offering things like higher property values, lower crime rates, better schools, and superior city services. This is because of the inequality between predominantly White and predominantly minority (in particular predominantly African American) neighborhoods in the United States (Peterson and Krivo, Reference Peterson and Krivo2010).
Why Might People Dislike Integrated Neighborhoods?
To foreshadow our results, we also expected to find a great deal of distaste for integrated neighborhoods—even among those who have created ideal neighborhoods that are integrated. What are the reasons people might dislike integrated neighborhoods? For one, integrated neighborhoods might be viewed as undesirable because of the racial conflict they engender.
Through a series of interpersonal and institutional actions, Whites have systematically attempted to keep Blacks in particular out of their communities (see, e.g., Loewen Reference Loewen2005; Massey and Denton, Reference Massey and Denton1993; Meyer Reference Meyer2000), routinely using violence to accomplish that. When unsuccessful, a pattern of ‘White flight’ was common. Thus, the reality of neighborhood-based racial conflict throughout our nation’s history is undeniable. This foundation creates the possibility that individuals— particularly, but not exclusively, among the racial/ethnic minorities who were primarily the victims of this violence—will perceive integrated neighborhoods as undesirable because of the expectation that they will be rife with interracial conflict.
Residents may also view integrated neighborhoods as undesirable because they are filled with people the individual does not want to be around. Consistent with research on racial residential preferences, people may dislike living in neighborhoods with people other than their own racial/ethnic group because of negative attitudes or prejudices toward people of different races/ethnicities (Bobo and Zubrinsky, Reference Bobo and Zubrinsky1996; Charles Reference Charles2006; Farley et al., Reference Farley, Steeh, Krysan, Reeves and Jackson1994; Lewis et al., Reference Lewis, Emerson and Klineberg2011). For example, existing research has demonstrated that among Whites, those who hold negative stereotypes about racial/ethnic minorities want more members of their own group as neighbors (Charles Reference Charles2006; Farley et al., Reference Farley, Steeh, Krysan, Reeves and Jackson1994; Krysan et al., Reference Krysan, Couper, Farley and Forman2009). Similarly, research has shown that African Americans who believe that Whites will mistreat and discriminate against them hold preferences for more own group residents (or, conversely, for fewer Whites) than those who do not (Charles Reference Charles2006; Krysan and Farley, Reference Krysan and Farley2002). There have been fewer studies of the predictors of Latino’s residential preferences, but Charles (Reference Charles2006) found that Los Angeles Latinos who had a high level of attachment to their in-group and who perceived greater social distance from other groups were less interested in integration with Blacks and Asians. Finally, a dislike for integration may reflect a desire by all to avoid neighborhoods where the other residents do not share their cultural perspectives and orientations (Clark Reference Clark1986, Reference Clark1988, Reference Clark1989).
Although some objections to integrated neighborhoods may be an outgrowth of stereotypes and ‘traditional’ prejudice against specific groups, a more sociologically informed theory (Blumer Reference Blumer1958) directs attention to prejudice as a sense of group position. Reluctance to live in a particular kind of integrated neighborhood may reflect the sense that such a neighborhood disrupts the proper positioning of one’s group. Although originally a theory focused on the dominant (White) group, Lawrence Bobo and Vincent Hutchings (Reference Bobo and Hutchings1996) apply it in a multiethnic context. Specifically, they highlight how racial alienation is critical with respect to the levels of threat experienced by different racial/ethnic groups, with those who are dominant unlikely to feel threatened, while those in the least advantaged group feel the most threatened (Bobo and Hutchings, Reference Bobo and Hutchings1996). With respect to this analysis, objections to living in an integrated neighborhood may be framed in terms of the concern that integrated neighborhoods of a certain sort disrupt the sense of the proper position of groups. How this will play out in a multiracial/ethnic sample in which people are asked to consider living with a range of different racial/ethnic groups is informed by the general observation that there is a racial hierarchy in the United States.
For example, those higher on the racial hierarchy may be reluctant to live in a neighborhood with racial/ethnic groups lower on the hierarchy. The hierarchy in the United States puts Whites at the top, followed by Asians, Latinos, and African Americans clearly at the bottom. Prior research (Bobo and Zubrinsky, Reference Bobo and Zubrinsky1996; Charles Reference Charles2006) has shown this hierarchy in residential preferences both in terms of differences in preferences based on the individual’s position in the hierarchy, but also in terms of the groups who might be included in one’s neighborhood. For example, Charles (Reference Charles2006) reports that Whites’ ideal neighborhoods are much more likely to exclude African Americans (who are at the bottom of the hierarchy) than Asians (who are closer to the top of the hierarchy). Similarly, most African Americans, Latinos, and Asians are more interested in greater contact with Whites as neighbors than they are with any of the other racial/ethnic groups (apart from their own). Therefore, when considering desirable and undesirable integration, as we do in this analysis, evidence of the racial hierarchy may emerge in several different ways.
Finally, negative attitudes toward integrated neighborhoods may be based on the assumption that such neighborhoods will have a number of undesirable characteristics. For example, ethnographic accounts of communities grappling with changes in their racial composition (e.g., Maly Reference Maly2005) report that community leaders are well aware of the perception held by some that integrated neighborhoods—or, perhaps more accurately, integrating neighborhoods—are unstable and have (already or will soon develop) a range of undesirable characteristics, including being crime-ridden, having low and/or falling property values, poor upkeep, and inferior city services. This perspective is also clear in Ingrid Ellen’s (Reference Ellen2000) neighborhood stereotyping hypothesis, which is based on the expectation that a neighborhood with any minority presence may be perceived as racially unstable. Thus the argument goes, if such a neighborhood becomes increasingly minority, it will slip in terms of its quality of life. Footnote 3
Implicit in the above descriptions about why one might or might not like integrated neighborhoods is that both the kind of integrated neighborhood that might be desirable or undesirable, as well as the reasons underlying these perceptions, will vary by a person’s race/ethnicity and, as will become evident, the race/ethnicity of the people living in the imagined neighborhood. Racial residential preferences research has long documented racial/ethnic differences in such attitudes. For example, although the results from the GSS reviewed earlier reveal openness to substantial integration across all groups, they also show that Whites, Blacks, and Latinos differ in how many co-ethnics they prefer in their ideal neighborhood. Whites were strongest in this preference (favoring on average a 57% own-group preference, in comparison to 43% among African Americans and 33% among Latinos). Also illustrative from these GSS data (Charles Reference Charles2003) is that 20% of Whites created an all-White neighborhood as ideal; and 25% included no African Americans.
Latinos and African Americans were far less likely to create all in-group neighborhoods (1% and 7%, respectively). In short, while all groups on average appear to embrace integrated neighborhoods, African Americans, and to a lesser extent Latinos, appear to have stronger preferences for integration than do Whites.
In this paper, we unpack White, Black, and Latino attitudes toward integrated neighborhoods. We do this with an unfolding and increasingly fine-grained analysis of two survey questions asking respondents to identify their most and least desired neighborhood racial composition. We first present descriptive statistics showing the average racial/ethnic composition of the most and least desired neighborhoods for the three groups. Similar to Charles (Reference Charles2003; Reference Charles2006), we find substantial enthusiasm for sharing neighborhoods with people of different races and ethnicities when Chicago area Whites, Blacks and Latinos are asked to identify their ideal neighborhood racial/ethnic composition. But because we also asked them to identify their least desired neighborhood racial/ethnic composition, we also show in this first analysis that all groups also find integrated neighborhoods to be undesirable. We next use cluster analyses that enable us to single out the respondents who drew integrated neighborhoods as their most desired and we continue our analysis with only this sub-group. Focusing on this group, we examine more carefully, again using cluster analyses, what people who desired integrated neighborhoods created as undesirable neighborhoods. We shed light on the reasons for the neighborhoods’ undesirability by analyzing an open-ended question that asked them to explain their answers. We conclude with a qualitative analysis in which we illustrate the fragility and ambivalence toward integration that comes across when viewing side-by-side the reasons individual respondents give for why they like certain kinds of integration and dislike others.
DATA AND METHODS
Our analysis uses the 2004–2005 Chicago Area Study (CAS), a face-to-face multi-stage area probability sample of adults twenty-one years and older living in households in Cook County, Illinois. Cook County (which includes the city of Chicago and a number of suburban areas), was first stratified by racial/ethnic composition, based on tract-level data from the 2000 Census, and over-samples were drawn of African Americans, Latinos, and those living in racially mixed neighborhoods. A total of 789 interviews were completed by professional interviewers at the Survey Research Laboratory at the University of Illinois at Chicago, with a 45% overall response rate. Footnote 4 Approximately one hour interviews were conducted in the respondent’s home (or a location convenient and of their choosing) in either English or Spanish, from August 2004 through August 2005. One respondent was randomly selected by the interviewer from all eligible household residents (anyone over age twenty-one). Footnote 5 The survey covered a range of topics related to racial residential segregation, including racial attitudes, housing search behaviors and experiences, other measures of residential preferences, and an array of background characteristics. All analyses use a weight that incorporates a selection weight and an adjustment for nonresponse (which was the inverse of the response rate in each primary sampling unit), which attempts to compensate, to the extent our data permit, for the possibility of nonresponse biases. Our study is limited to White, Black, and Latino residents because there were an insufficient number of respondents from other racial/ethnic groups. Respondents self-identified their racial/ethnic background and were allowed to select more than one response, though very few respondents did. Although exact race-matching was not possible, most survey respondents (82%) were interviewed by someone of their own race/ethnicity. Footnote 6
A comment about our study location is in order. That is, a limitation of this study is that it examines a single city—one that has a long history of high levels of Black-White segregation. Our study must be qualified therefore with the observation that race/ethnicity, particularly in the context of residential issues, is likely to have more salience for residents of the Chicago metropolitan area than in some other areas of the country. Nevertheless, there are a number of large rust-belt cities that share demographic, social, and racial histories similar to Chicago, so it is also not likely to be unique in its patterns. Moreover, as we will show, our initial analyses of the average ideal neighborhood racial composition (the only ones for which we can make comparisons) very closely mirror those of national survey data (General Social Survey).
The survey included a replication and slight modification of the technique developed by Charles (Reference Charles2006) for measuring racial residential preferences. Footnote 7 Respondents were given a card like that shown in Figure 1, which contained fifteen houses (including their own, in the middle), and were instructed:
Now I’d like you to imagine an ideal neighborhood that had the ethnic and racial mix you personally would feel most comfortable in. Using the letters W for White, B for Black, H for Hispanic, AS for Asian American and AB for Arab American, please put a letter in each of these houses to represent your ideal neighborhood where you would most like to live. Please be sure to fill in all the houses.
To measure the least desired neighborhood racial composition, respondents were given a second card, and told:
Now I would like you to do the same sort of thing, except this time I would like you to construct a neighborhood that has the racial and ethnic mix that you would feel least comfortable in. Use the same letters to identify the race/ethnicity of each house.
Our first analysis examines the average percentage of Whites, Blacks, Asian Americans, Latinos, and Arab Americans living in respondents’ most and least desired neighborhood. Then, in order to summarize—and create a typology of—the neighborhoods that were drawn by respondents as most and least desired, we conducted a two-step cluster analysis (Norušis Reference Norušis2010) separately for each racial/ethnic group. The variables used to generate the clusters were percent White, percent African American, percent Arab American, percent Asian American, and percent Hispanic in the respondent’s most and least desired neighborhood (separately). We then describe the average racial composition of the most/least desired neighborhoods within each cluster. We used the two-step clustering command in SPSS, allowing the SPSS algorithm to determine the number of clusters, which it does based on the change in BIC and ratio of distance measures (Norušis Reference Norušis2010). We used these clusters to identify respondents who drew ‘integrated’ ideal neighborhoods and the subsequent analyses of the least desired neighborhoods included only those respondents whose ‘most desired’ neighborhood was classified into one of the racially integrated clusters. Footnote 8
In order to reveal the complexity and nuance of attitudes toward integration we also conduct two kinds of analyses (one more quantitative and the other more qualitative) of open- ended follow-up questions asking people to explain what they like (or dislike) about the neighborhoods they drew. Specifically, after filling in the ideal neighborhood card, the respondent was asked: “Looking at the neighborhood you’ve created, could you please tell me what makes this an ideal racial and ethnic mix to you?” Then, after drawing their least desired neighborhood they were asked: “Looking at the neighborhood you just drew, can you tell me why you would feel least comfortable in this kind of neighborhood?”
The responses to both open-ended questions were collected in up to two ways: 1) interviewers typed the responses into the laptop computer used to conduct the interview; and 2) a digital recorder was activated which provided exact responses (with the exception of instances of poor recording quality and respondents not consenting to being recorded). The transcriptions of the recordings and the interviewer recorded responses were both used in coding.
A complex coding scheme was initially conceived deductively, and included the following core themes from racial residential preferences research: 1) perceptions (stereotypes) of racial/ethnic groups; 2) feelings toward the in-group; 3) feelings toward the out-group; 4) perceptions of inter-group relations (conflict or cooperation); 5) relative numbers of each group; 6) perceptions of cultural similarities or differences; and 7) perceptions of neighborhood characteristics. After an initial attempt to code the data with these themes, we revised our scheme (and thus it was also an inductive process) to include explicit statements about a preference for diversity. After multiple revisions and tests, we finalized the coding scheme and two graduate research assistants independently coded all responses, made comparisons, and reconciled any disagreements (see Appendix A for the detailed coding scheme). Footnote 9
For our more quantitative analysis of the open-ended questions, we grouped individuals based on which cluster their neighborhood drawing was assigned to, and then calculated the distribution of themes within the clusters; we use selected quotations to illustrate the key themes. For our more qualitative analysis of the open-ended questions, we analyzed how individual respondents who drew specific combinations of most and least desired integrated neighborhoods explained what they liked and disliked about the neighborhoods. This analysis illustrates how the same individual struggles with issues of neighborhood integration, showing how it is that a person can both like and dislike integration—and what it is that shifts the balance from desirable to undesirable.
RESULTS
We begin with the overall picture of the neighborhoods that the total sample of Whites, Blacks, and Latinos created as their most and least desired (Table 1)—specifically, the average racial/ethnic composition of the average White, Black, and Latino respondent. With respect to the most desirable neighborhood composition, our results are similar to the national GSS sample (Charles Reference Charles2003): 1) all three racial/ethnic groups draw, on average, diverse ideal neighborhoods; 2) all three groups, on average, want a substantial in-group presence, though Whites alone create a majority own-group neighborhood; and 3) the exact mix of different racial/ethnic groups varies for Whites, Blacks, and Latinos.
Table 1. Racial/Ethnic Composition of Most and Least Desired Neighborhood, by Respondent Race/Ethnicity (all respondents). 2004–2005 Chicago Area Study (s.e. in parentheses).
It is striking how different these ideal neighborhood compositions are from the reality of respondent’s neighborhoods. Footnote 10 Cook County Whites created a 53% White neighborhood as their ideal, while they live in neighborhoods that are on average 80% White. African Americans created an ideal neighborhood with 39% African American residents, but they reside in neighborhoods that are on average 74% Black. Finally, Latinos come closest to a match between their ideal and actual neighborhood: the ideal neighborhood is 35% Latino, while their actual is on average 56% Latino.
The columns on the right of Table 1 reveal that in our full sample, on average, all three racial/ethnic groups also create integrated neighborhoods as their least desired racial/ethnic composition. Footnote 11 For Whites, the main difference between the average most and least desirable neighborhoods is the reduction in the percentage of Whites. Interestingly, it is not accompanied by an across-the-board increase in the size of all other racial/ethnic groups. Rather, there is a substantial increase in the Black and Arab American populations; a slight increase in the Latino population; and a modest reduction in the percentage of Asian American neighbors. Thus, although there was little sign of a racial hierarchy in the ‘most desirable’ neighborhood racial/ethnic composition for Whites, it is clearly evident in the results for the least desired neighborhood. The pattern is also clear for Latinos: the most striking difference between the most and least desired neighborhood is that both the White and Latino population shrinks while the Black population grows dramatically. Latinos follow a pattern suggested by the existence of a racial hierarchy, as there is a greater interest in residing in neighborhoods with those at the top (Whites) of the hierarchy than at the bottom (Blacks).
The most puzzling results are for African Americans, where the average racial/ethnic composition changes relatively little between the most and least desired neighborhoods. There is a change in the size of the Black population, but it is more muted than was the case for Latino and White respondents. The cluster analysis and the open-ended survey data, to which we now turn, will allow us to investigate this puzzling pattern as well as the more general question of understanding of the diversity of attitudes toward integrated neighborhoods.
What Kinds of Neighborhoods Do People Find Desirable?
Because our purpose is to isolate those who drew integrated ideal neighborhoods, and to be able to analyze our results in a way that focuses on integrated neighborhoods, our approach cannot use ‘average’ racial compositions or simple ‘percentage own group’ but instead we need a way to summarize and analyze the neighborhoods people actually drew, in all their complexity and variation. The results of the cluster analyses, presented in Figure 2, do this. Three patterns stand out. First, the clusters for Whites and African Americans each include one cluster/neighborhood that is nearly entirely co-ethnic (White cluster 3; Black cluster 3), though almost twice as many Whites as Blacks are classified into it (17% vs. 9%). By contrast, no “all-Latino” cluster was extracted for Latino respondents. Footnote 12 The second pattern that stands out is that the ideal neighborhoods that most respondents drew are integrated. Footnote 13 The most popular cluster for Whites is a neighborhood in which Whites are just over the majority (on average, Whites are 58% of the neighborhood), and then has about equal percentages of Blacks, Asians, and Latinos—and essentially no Arab Americans. The two most popular clusters (Clusters 1 and 2) for African Americans are integrated, though in different ways. All of the Latino respondents are classified into two different kinds of integrated neighborhoods as their ideal.
Fig. 2. Cluster Analysis Results for Most Desired Neighborhood Racial/Ethnic Composition, by Respondent’s Race/Ethnicity. (Numbers in parentheses are percentage of respondents classified into each cluster.) 2004–2005 Chicago Area Study.
Finally, the third pattern worth noting is that Whites, Blacks, and Latinos all have as one of their more commonly-drawn neighborhoods one in which their own group has a numerical advantage (or in the case of Latinos, parity) over the rest of the groups, and the rest of the neighborhood is about evenly divided across the remaining racial/ethnic groups, although the size of the Arab American population is always the smallest. We refer to these clusters (labeled in Figure 2 as White Cluster 2; Black Cluster 1; and Latino Cluster 1) as ‘five group’ integrated neighborhoods. 57% of African Americans, 39% of Whites, and 60% of Latinos are classified into this integrated cluster as their ideal neighborhood racial/ethnic composition.
Looking across these results, the story appears to be that all three racial/ethnic groups desire to share residential space with ample numbers of their own racial/ethnic group—though the size of the in-group is larger for Whites than for Blacks or Latinos—but there is also interest in having non-trivial percentages of neighbors from other racial/ethnic groups. In other words, in their imagined “ideal” neighborhood, all racial/ethnic groups in quite high percentages embrace integration with all kinds of racial/ethnic groups. Because our interest is in attitudes toward integrated neighborhoods, the rest of the analysis is limited to those who created integrated ideal neighborhoods (this includes 83% of Whites; 91% of Blacks; and all of the Latino respondents). We now turn to a brief discussion of what people who draw integrated neighborhoods say it is that they like about them.
Why Do People Say They Like These Integrated Neighborhoods?
When we ask people why they like the integrated neighborhoods they drew, a desire for ‘diversity’ swamps all other reasons. 75% of Whites, 70% of Blacks, and 58% of Latinos who were classified in the cluster that represents the ‘five group’ integrated neighborhood said they felt the integrated neighborhood they drew was desirable because it was diverse. Footnote 14 Respondents explained the benefits of diversity in two key ways: 1) the exposure to different people both for the respondent and for their children, and 2) the perception that diverse neighborhoods will help improve racial/ethnic relations. The ‘exposure’ theme is the most popular among Whites (40%), though Blacks and Latinos also mentioned it (31% and 19%, respectively). A few examples illustrate the theme:
Just a variety of different people. [Why is this desirable?] Um I think it’s a healthier atmosphere, good for kids. [Why?] I think it’s good for children to be exposed to all types of people. [Anything else?] No. (White respondent: 36% White, 14% Black, 21% Hispanic, 29% Asian American, 0% Arab American) Footnote 15
Because you have children getting to know different cultures and not being afraid of different races. (African American respondent: 36% White, 43% Black, 7% Hispanic, 0% Asian American, 14% Arab American)
Although the exposure theme was mentioned by many African Americans, the somewhat more popular explanation for Blacks (38% mentioned it) was that diverse neighborhoods could help improve race relations and create racial harmony: Footnote 16
Cause there’s some of every race living there. I got everybody. [Why is that desirable?] Oh. That you could. That you could be in the neighborhood and you have the children and they’re at the park together and getting along. Letting you know people can get along, regardless of what they say. Or how it used to be. (29% White, 21% Black, 14% Hispanic, 14% Asian American, 21% Arab American)
Cause this represents a true melting pot [What is it about a melting pot that is good?]. Well they have more variety and you learn more when you have more people around. I don’t want to be in an all White or all Black neighborhood. Can develop racial tolerance for one another when you live next to one another. (21% White, 21% Black, 21% Hispanic, 21% Asian, 14% Arab American)
Thus, consistent with the observation that integrated neighborhoods can be sites for interracial cooperation (Iceland Reference Iceland2009), and in line with the contact theory of inter-group relations (Allport Reference Allport1954), African Americans who created very diverse integrated neighborhoods generally did so because of an optimism that such inter-group contact would ease the tensions and conflict that characterizes much of U.S. race relations.
To this point, the analysis reveals little evidence of a racial hierarchy. Chicago residents justify their preference for diversity by focusing on the ideals of integration, the value of exposure to differences, and the interest in contact. But there is little evidence that who those groups are makes much difference to how integration is interpreted. This story changes dramatically once we ask the opposite question of our respondents: What kind of neighborhoods do you find undesirable?
For the next stage of our analysis, where we put together the good and bad sides of integration, we discuss the results for each racial/ethnic group in turn, beginning with the cluster analysis of least desired neighborhoods (among those whose ideal neighborhood is integrated). After describing the reasons given for why the neighborhoods were undesirable, we then put side-by-side the explanations for why integration is good and bad in order to reveal the complexity and nuance of these attitudes.
White Results
For Whites Who Prefer Integrated Neighborhoods, What Kind of Neighborhoods Are Undesirable?
The first step is to examine what constitutes an undesirable neighborhood racial/ethnic composition among Whites whose most desired neighborhoods were integrated. One possibility is that those who create integrated desirable neighborhoods will create segregated neighborhoods as least desirable. What we find instead is that undesirable neighborhoods are also integrated neighborhoods. They are just integrated differently.
Specifically, the results of the two-step cluster analyses for Whites’ undesirable neighborhoods (shown in Figure 3a) reveal two clusters. About one-quarter of Whites created a neighborhood dominated by Arab Americans (on average, the neighborhood/cluster included 57% Arab Americans); by our standards from the desirable neighborhood analysis (that a neighborhood with fewer than 60% of one group is integrated) this is an integrated neighborhood. But it is clearly dominated by a single racial/ethnic minority group and some might view it as segregated. The majority of Whites (76%) created as their least desired neighborhood one with substantial percentages of African Americans, Latinos, and Arab Americans, and very few Whites or Asian Americans. Thus, the ‘most popular’ least desired neighborhood can clearly be classified as integrated. But unlike the results for the desirable neighborhood, the racial hierarchy appears quite clear in the undesirable neighborhood question since African Americans and Latinos—but not Asian Americans—figure prominently in them. Moreover, we have some insights into the position of Arab Americans, which is uncertain in the existing literature. Given that Cluster 1 has a relatively large Arab American population, and that one-quarter of Whites specifically created a majority Arab American neighborhood (Cluster 2) as their least desired, Arab Americans appear to be vying with African Americans for the bottom of the racial hierarchy.
Fig. 3a. Racial/Ethnic Composition of Least Desired Neighborhood (by Cluster), among Whites Who Drew Diverse Desirable Neighborhoods. 2004–2005 Chicago Area Study.
What Do Whites Say Is Wrong with the Undesirably Integrated Neighborhoods?
We continue to follow this group of Chicago area Whites who created desirable integrated neighborhoods, but then went on to draw undesirable neighborhoods that were clearly integrated (e.g., this next analysis excludes the Whites whose undesirable neighborhood was classified in Cluster 2, the majority Arab American neighborhood). In Table 2 (column 1), we report the themes mentioned by those Whites whose undesirable neighborhood was classified in the integrated “Black/Latino/Arab American” neighborhood (Cluster 1). Footnote 17 The theme mentioned most frequently is neighborhood quality (30%):
Because of crime and drugs and stealing, robberies, shootings, and so on and so forth. [What is it about the neighborhood that makes it this way?] Well going by what I have seen from neighborhoods close to here, the Black people have destroyed their neighborhood with crime and drugs. (21% White, 29% Black, 29% Hispanic, 21% Arab American)
Table 2. Themes Mentioned by Respondents Explaining Why They Disliked the Neighborhood, By Undesirable Neighborhood Cluster into which They Were Classified, by Respondent Race/Ethnicity. 2004–2005 Chicago Area Study. (% mentioning theme; themes not mutually exclusive.)
A second example combines concerns about the neighborhood and its social class characteristics with cultural concerns, another common theme:
If this were the neighborhood makeup, you could almost be certain that it would be lower income, higher crime, worse school. Not because of inherent characteristics of these people but people of these backgrounds have a lower socio-economic status and opportunity for education. Also I am just not used to….it would be a culture shock. Just like moving into any new culture. It is difficult in general to see people that don’t look like them. (43% Black, 57% Hispanic)
The two themes that capture different levels of concern about how the out-group will treat the respondent also constitute a common objection to this kind of integrated neighborhood. About even numbers of White respondents point to either general ‘discomfort’ being around the out-group or the outright hostility they think they would experience from the out-group. The two versions are illustrated in the following examples:
It is more of one race not as mixed. I would feel like an outsider. More singled out. (57% Black; 29% Hispanic; 14% Arab American).
I would be the only White person. I would be intimidated and picked on. (57% Black, 7% Hispanic; 37% Arab American)
We next dig even deeper and examine how individual respondents shift in the racial composition of their most and least desired neighborhoods and how they explain the differences between the two. First, in a quantitative analysis that calculates, for individual respondents, the increases and decreases in the size of each racial/ethnic group in their two neighborhoods, there are many similarities to what was reported in Table 1 at the group level (see Appendix B). The difference between desirable and undesirable integration for individual Whites is that there is an increase in the size of the African American and Arab American populations; and a concomitant loss of White residents. Even more can be learned about what underlies the shift from most to least desirable by comparing, side-by-side, how a given respondents explained their most and least desired integrated neighborhoods. Footnote 18 We begin with a White respondent whose ideal neighborhood was classified in the ‘five group’ neighborhood cluster. This respondent explained their ideal choice in this way:
Multicultural. [Why is that ideal?] Why is it ideal? [Yes.] It’s nice to have a nice mix of people together. [Can you be more specific?] You can learn a lot from their cultures and values and incorporate that into your family and community. (43% White, 14% Black, 14% Hispanic, 21% Asian American, 7% Arab American)
The respondent went on to draw a least desirable neighborhood that was also integrated, and in the explanation we see that the desire for multiculturalism expressed above has waned:
Um. It would be supported by different values that I wouldn’t necessarily be able to support their different value systems. [Why do you say that?] It goes back to culture. There are differences in culture. [Can you be more specific about the differences?] Not really. (14% White, 43% African American, 14% Hispanic, 7% Asian American, 21% Arab American)
These responses, placed side-by-side, reveal limitations to a commitment to integration and diversity, i.e., this person values exposure to different cultures, but also finds different cultures to be problematic. It appears that it is the cultures of African Americans and Arab Americans that are especially troubling, since these are the groups whose size has increased in the undesirable neighborhood. This is consistent with the overall patterns in our differences analysis (Appendix B).
The next example highlights the discomfort some Whites feel in certain kinds of integrated neighborhoods—often articulated as concerns about being the minority. First, the respondent explains why they would value the ideal integrated neighborhood they created:
Um. I just want my kids to be around everybody. [Why?]. Um. I just want. I think they’ll be more open-minded. [Anything else?]. No. (21% White, 32% Black, 29% Hispanic, 14% Asian American, 14% Arab American)
This respondent went on to create an integrated undesirable neighborhood and had this to say about why it was undesirable:
Um. Because I would not like to be the minority. [Why]. I don’t know. I just wouldn’t [How would you feel?] I wouldn’t be comfortable. I don’t know if I’d be safe or not. But wouldn’t be comfortable. [Anything else?] No. (14% White, 57% Black, 29% Hispanic, 0% Asian American, 0% Arab American)
Notice that this respondent was in the minority in both neighborhoods; but in the second (undesirable) neighborhood, there were more African Americans and no Asian Americans or Arab Americans. In this case, it was apparently this particular mixture that became troubling and raised the salience of being ‘the minority;’ a concern that was absent in the first also very diverse neighborhood. This respondent’s desire for integration is circumscribed, hinging on both the relative size of the White population and the particular makeup of the rest of the neighborhood. The shift in the composition of the neighborhood is consistent with a desire on the part of Whites to avoid living with too many neighbors who are at the bottom of the racial/ethnic hierarchy.
African American Results
For African Americans Who Prefer Integrated Neighborhoods, What Neighborhoods Are Undesirable?
The cluster analysis for African Americans’ least desired neighborhood (Figure 3b) resulted in a larger number of clusters than was the case for Whites. In addition, four of the five clusters have a majority of a single racial/ethnic group, and three have in excess of 60% of a single group. Three clusters each had just 11% of African American respondents who were classified into it (Cluster 3 (predominantly White); Cluster 4 (predominantly Latino); and Cluster 5 (predominantly Arab American). Another one-third of African Americans were classified into the cluster best described as the “all Black” neighborhood (Cluster 2). Thus, African Americans come closest to the expectation that among those who desire an integrated neighborhood, the least desirable neighborhoods will be ones that are not integrated: 65% of Blacks created a neighborhood dominated by one or another of the racial/ethnic groups. Footnote 19 This leaves 35% of African Americans who are classified into an integrated least desirable neighborhood; and this cluster is best described as a ‘five group’ integrated neighborhood (Cluster 1). Interestingly, there is little difference in the racial/ethnic composition of this undesirably integrated neighborhood and the racial/ethnic composition of the most popular of the desirably integrated neighborhoods. This pattern suggests that different forms of ‘demographic’ integration—that on their face seem similar—can be apparently understood very differently. What is at the root of these differences becomes clearer as we turn to the words of Black Chicagoans themselves who explain what is wrong with the integrated neighborhoods that were demographically very similar to the ones they found to be desirable.
Fig. 3b. Racial/Ethnic Composition of Least Desired Neighborhood (by Cluster), among African Americans Who Drew Diverse Desirable Neighborhoods. 2004–2005 Chicago Area Study.
What Do African Americans Say Is Wrong with the Undesirably Integrated Neighborhoods?
Before reviewing the objections to the five-group integrated neighborhoods, because there were so many Black respondents who created ‘segregated’ neighborhoods as their least desired, we provide a brief summary of the reasons for objecting to the various predominately single-race neighborhoods (e.g., predominantly White, Latino, Arab American, or Black). The reasons vary somewhat depending on which group was in the majority. Concerns about hostility from the out-group dominate objections to the predominantly White neighborhood (60%). Objections to the predominantly Latino and predominantly Arab American neighborhoods included a range of negative perceptions about the dominant out-group and concerns about cultural differences. Footnote 20 For the one-third of African Americans who drew a predominantly Black neighborhood as the least desirable, the reasons were more varied, although the most common (mentioned by about one-quarter) were concerns about the quality of the neighborhood.
The results for the one-third of African Americans who created integrated desirable and undesirable neighborhoods are reported in Table 2. The most frequent are mentions of how the respondent feels they will be treated by the out-group (resulting in feelings of discomfort or experiencing hostility from the out-group). Despite the fact that the least desired integrated neighborhood includes people from a variety of racial/ethnic groups, the emphasis for many is on the size of the White population. For example:
I don’t want to live around all White people because they would exclude me from everything. (43% White; 29% Black, 0% Hispanic, 14% Asian American; 14% Arab American)
Another respondent echoed these concerns, but also mentioned another common theme focusing on cultural differences:
One, I’m the only Black person on the block. A lot of times a lot of my neighbors would not relate to me because we are of different backgrounds. I know the communications would not be as open. In a neighborhood like this if something did happen, you would probably be blamed for it. They may think property values may go down because I am there. Others might not want to move there because I’m there. [Anything else?] No. (64% White, 0% Black, 0% Hispanic; 21% Asian American; 14% Arab American)
Finally, one respondent highlighted the problem with this much diversity succinctly, with again an emphasis on ‘cultural’ differences:
I don’t think that many different races could get along. There are too many different beliefs. People don’t trust people that much. (29% White; 21% Black; 7% Hispanic; 14% Asian American; 29% Arab American).
Overall, either discomfort or hostility from the out-group was mentioned by more than one-half of Black respondents whose least desired neighborhood was integrated. Cultural concerns were also apparent (23%), but not as frequent.
The size of the White population as a critical feature of what makes desirable and undesirable integration for African Americans—and why—becomes even more clear and vivid when we compare, side-by-side, what a single respondent says about good and bad integration: both in the numbers and the words. First, the quantitative analysis (Appendix B) comparing changes in racial composition shows that the biggest change is in the size of the White population—increasing from 24% to 34%—and the corresponding loss of Asian American residents (of ten percentage points).
The reasons become clear when we turn to the open-ended responses. The first example is an African American whose most and least desired neighborhoods were both classified in the ‘five group’ integrated neighborhood. The explanation for the ideal integrated neighborhood highlights the most common theme mentioned by African Americans—that integration can foster racial harmony:
We can get along with each other. It would be interracial. Communication. Have block club meetings about keeping down the crime in the neighborhood. (21% White, 43% African American, 14% Hispanic, 21% Asian American, 0% Arab American)
This respondent went on to draw a least desired neighborhood that included all five racial/ethnic groups, and explained that it was undesirable because:
They have more Whites. They would be prejudice. They don’t want you in the neighborhood. Some might not like Black people. (36% White, 14% African American, 14% Hispanic, 14% Asian, 21% Arab American)
Even though the size of other groups also changed (the Arab American population went from none to 21%), it was the increase in the White population that was salient to this respondent. For this and other Black respondents, it appears less important which other racial/ethnic groups are in the undesirable neighborhood; it is the size of the White population that was most salient in this neighborhood that went from being desirably to undesirably integrated. This example highlights how a very slight demographic change can have very salient social consequences.
A second example illustrates a similar dynamic. First, in describing why the diverse neighborhood was ideal, this African American respondent explains:
It has some of everybody in the picture so there is a mixture because kids need to be exposed to different cultures and it balances out better. (36% White, 36% Black, 7% Hispanic, 14% Asian American, 7% Arab American)
Integration becomes undesirable when the mix has more Whites, slightly more Arab Americans, and the African American population dwindles. But the explanation focuses entirely on the size of the White population:
It’s only three Blacks and the other are different race. And when it’s like that you don’t get much voice in the neighborhood. [Why is that?] Mostly Whites are over-ruling. They feel superior and they are vicious when they want their way with things and the Blacks can be persuaded to whatever is going on. (43% White, 14% Black, 14% Hispanic, 14% Asian American, 14% Arab American)
Despite the fact that other groups shift in size between this respondent’s most and least desirably integrated neighborhoods, the social significance of these shifts is very much constructed in terms of the experience of being too much in the minority when there is a relatively large White population.
Latino Results
For Latinos Who Prefer Integrated Neighborhoods, What Kind of Neighborhoods Are Undesirable?
The analysis of Latinos’ least desired neighborhood cards generated a four-cluster solution (see Figure 3c). Two of the clusters have substantial African American populations: 19% of Latinos are classified into Cluster 4, which has on average 92% African Americans; another 24% were classified into Cluster 3 where Blacks were, on average, one-half of the residents. The remaining two (slightly more popular) clusters are different combinations that include sizeable percentages of several different racial/ethnic groups—in other words, clusters that were more integrated. Specifically, about one-third of Latinos create a neighborhood that is “White/Latino/Black” (Cluster 1); and just over one-quarter create an integrated neighborhood with substantial Arab American, African American, and Asian American residents (Cluster 2). Latinos are unusual in that one of the undesirable clusters has a sizeable Asian American population.
Fig. 3c. Racial/Ethnic Composition of Least Desired Neighborhood (by Cluster), among Latinos Who Drew Diverse Desirable Neighborhoods. 2004–2005 Chicago Area Study.
Overall, Latinos are intermediate between Whites and Blacks in the extent to which those who desire integration created ‘non-integrated’ neighborhoods. Footnote 21 43% of Latinos whose ideal neighborhood was integrated created majority single-group undesirable neighborhoods, compared to 65% of African Americans and just 23% of Whites.
What Do Latinos Say Is Wrong with the Undesirably Integrated Neighborhoods?
Again, given the fairly high percentage of Latinos who drew majority single-group least desired neighborhoods, we first briefly summarize how Latinos explained their dislike of these neighborhoods. Latino objections to heavily African American neighborhoods differ based on whether it was an overwhelmingly (average of 92%) Black neighborhood (Cluster 4) or whether it was a predominately Black neighborhood with substantial Arab American residents (Cluster 3). In the case of the former (Cluster 4), Latinos explained themselves by drawing on negative perceptions and stereotypes about Blacks themselves; while the Latinos who drew the African American/Arab American neighborhood (Cluster 3) focused on a combination of neighborhood quality (48%) and negative stereotypes about Arab Americans (25%) and African Americans (36%).
The reasons people gave for why they didn’t like the two more integrated neighborhoods (Clusters 1 and 2) depended on the specific racial/ethnic mix (see Table 2). A number of different themes were mentioned (negative perceptions of Arab Americans, Asian Americans, and Blacks, and concerns about hostility from the out-groups), but those who drew the Arab American/Asian American/African American neighborhood were most likely to refer to concerns about cultural dissimilarities. For example:
I guess the opposite of what I just said. It is not people that are in my social circle, and I do not know enough about those cultures. There is always hesitancy in what you don’t know. [Anything else?]. No. (0% White; 29% Black; 14% Hispanic; 29% Asian American; 29% Arab American).
There is no single theme that dominates the explanations given by those Latinos who drew an integrated White/Black/Latino neighborhood as their least desirable. Neighborhood quality, negative perceptions of Blacks, and cultural concerns were among those that were most frequently mentioned, with around 20–30% mentioning each. A few examples illustrate these themes, beginning with concerns about the neighborhood quality:
Because I’d feel surrounded with more danger. I’d feel more insecure. [Anything else?] Nothing else. (21% White; 36% Black; 36% Hispanic; 7% Asian American; 0% Arab American)
And another focused more on the cultural differences:
Because of the difference. Because of the culture clash. Sometimes they don’t understand my customs. Or I don’t understand theirs. The way of thinking is different. (43% White; 43% Black; 14% Hispanic; 0% Asian; 0% Arab American)
Turning to the side-by-side comparisons (see Appendix B), the quantitative patterns for Latinos whose ideal neighborhood is integrated are clear: moving from the most to least desired integration means a decline in White, Latino and Asian American residents and a sharp increase in the percentage of African Americans (the level differs depending on which neighborhood was most desired). The words of selected respondents reveal how it is the size of the African American population in particular that influences the assessment. Specifically, Latinos whose ideal neighborhood was classified into the ‘five group’ cluster were most likely to create a White/Latino/Black least desired neighborhood. When explaining their dislike, the focus was on cultural differences and neighborhood quality. The contrast is clear. In the first quotation, we see what this Latino respondent liked about their ideal integrated neighborhood:
There’s different cultures. There wouldn’t be any racism. I think that everyone would get along better. Life would be more positive. [Anything else?]. No. (29% White, 21% Black, 43% Hispanic, 7% Asian, 0% Arab American)
This respondent went on to explain why a differently integrated neighborhood was undesirable:
I wouldn’t feel security for myself or my family. Because of this people, because one race is predominant here and it doesn’t give me trust. I wouldn’t feel that the kids are safe. The value of the houses wouldn’t be so high because in neighborhoods like that values are low, I think. (29% White, 50% Black, 21% Hispanic, 0% Asian American; 0% Arab American).
For this respondent, their undesirable neighborhood is marked by a large increase in Black residents relative to their desirable neighborhood. And along with this comes negative perceptions of its quality.
A second example shows respondents concerns about culture that frequently emerge. First, when explaining the virtues of an integrated neighborhood, this respondent said:
We’d be like almost perfect. If there’s solely Hispanic or African American… It would be good to live mixed up. In each country, the African Americans live isolated. The different races living apart and I think that is wrong because we should live together. We should know how to live together. [Why is a mix desirable?] We would all get along as neighbors. It would come out something good—for example, more trust. [Anything else?] No, that’s all. (29% White, 14% Black, 43% Hispanic, 7% Asian American, 7% Arab American)
But when asked to make a least desirable neighborhood, this same respondent reduced the number of Whites and Latinos, and increased slightly the size of the Black, Asian American, and Arab American population and explained:
Maybe because of not knowing their customs. For example, Arab Americans, I don’t know their ways of life. And also the Asian. I don’t know their customs. But I do know how White and Hispanic people live. Living amongst those whom I don’t know would make me uncomfortable. [Anything else?] No, I believe that’s the only one. (14% White, 21% Black, 36% Hispanic, 14% Asian American, 14% Arab American)
This example reveals a comparatively fragile commitment to the virtues of integrated neighborhoods. On the one hand, there is a strong interest in learning about other cultures and a belief that it is important to not be segregated; but in the next breath, the same respondent expresses reluctance about living in an integrated neighborhood because of cultural differences and discomfort.
CONCLUSION
There is no mistaking that integrated neighborhoods are popular among Chicago area residents. Equally unmistakable is that integrated neighborhoods are maligned, even among Chicago area residents who said they prefer integrated neighborhoods. The devil is very much in the details. Those details, according to our analysis, focus on: 1) How much integration there is in a neighborhood and therefore the extent to which one’s own racial/ethnic group is present; and 2) What groups are included, which excluded, and their location in the racial hierarchy. These findings speak to and reflect dynamics of race relations that highlight group position and contribute insights into the ways in which individuals experience and understand being in the ‘minority’ status.
Specifically, we find that Whites, Blacks, and Latinos all prefer a numerical advantage for their own group; but whether integration is desirable or not hinges on both the relative size of their own group as well as which groups make up the rest of the neighborhood and where their own and the other groups are in the racial hierarchy. The first of these—the general ‘numbers’ issue—has received attention in research on residential segregation, insofar as it relates to the ‘tipping point’ argument (Schelling Reference Schelling1971). And there is a tradition of research that highlights the importance of the relative size of groups for inter-group relations dating back to Hubert Blalock’s (Reference Blalock1967) observations about how size of the Black population influenced White racial attitudes. But by and large, studies of residential segregation and racial residential preferences have given relatively little attention to how individuals understand and experience numerical minority status—particularly in a way that allows comparisons across different racial/ethnic groups. Our study of attitudes toward neighborhood integration provides such an analysis, and what we find lends texture and nuance to general theories of the relevance of group size and the notion of prejudice as a sense of group position (Blumer Reference Blumer1958) as also extended by Bobo and Hutchings (Reference Bobo and Hutchings1996) to the multiethnic context. In essence, numbers matter for all three groups; but there are differences with respect to how and why they matter. And those differences are inextricably linked to the racial hierarchy in the United States.
Our data reveal that the prospect of being a minority is unappealing to most. As one White respondent put it, “I would be the minority. There’s something unpleasant about being the minority.” But our analysis suggests that what makes being a minority “unpleasant” looks different depending both on where one sits in the racial hierarchy and where the others in the neighborhood sit.
For Whites, the difference between acceptably and unacceptably integrated neighborhoods appears to be, on the one hand, the size of the White population. The percentage of Whites that a White person puts in their desirable neighborhood drops substantially when they create their undesirable neighborhood. But for many Whites, both the desirable and undesirable neighborhoods are integrated. When asked to explain what it is about certain integrated neighborhoods that make them undesirable, many refer to the lack of other Whites. But the Whites in a respondent’s ideal integrated neighborhood are not replaced ‘at random’ in the least desired integrated neighborhood. Rather, it is the African American and Arab American presence that increases the most. The Latino and Asian American populations either stay the same or decline modestly. Who replaces Whites to make a ‘good’ integrated neighborhood become a ‘bad’ integrated neighborhood, reveals the racial hierarchy. Thus, it is more complicated than simply being “a minority” in the neighborhood. It is being a minority in a particular kind of neighborhood; namely, a neighborhood with too many from the bottom of the hierarchy.
The concept of racial alienation figures prominently in making this interpretation (Bobo and Hutchings, Reference Bobo and Hutchings1996). By definition, given their advantaged status, Whites feel less threat and alienation than members of lower-status groups. In our study, this is suggested by the fact that Whites’ objection to being a numerical minority was explained with rather vague descriptions that were consistent with the group position sense of prejudice: not being in the numerical majority violated their sense of the proper order of things and it was “uncomfortable” and “lonely and isolating.” Whites’ explanations (though not likely their actions to avoid such a circumstance) had a different and more muted tenor than those offered by the groups not at the top of the hierarchy. Although it is the case that Latinos’ and African Americans’ patterns were also distinctive from each other as we will see.
Latinos occupy a middle position between Whites and Blacks on the racial hierarchy both in terms of who wants to share residence with them and the groups with whom Latinos want to share neighborhoods. First, Latinos’ orientation toward being a numerical minority focused on the experience being culturally unsettling—as we will see, the tone of their objections was not as vivid as that of African Americans. Perhaps this is attributable to their ‘in the middle’ status. The role of the racial hierarchy in shaping their preferences becomes very clear when looking at who is included and excluded in their most and least desired neighborhoods. There is a dramatic change in the size of the African American population between the most and least desired integrated neighborhoods: depending on which particular integrated neighborhood was most desirable, the size of the Black population in the undesirably integrated neighborhoods of Latino respondents was either twenty-six percentage points or a full forty percentage points larger than their desirably integrated neighborhoods (See Appendix B). The increase was accompanied by declines in the White and Latino population. For Latinos, who occupy a middle position in the hierarchy, sharing neighborhoods with Whites may be perceived as upward mobility, while sharing with African Americans may be perceived as downward mobility (Charles Reference Charles2006).
For African Americans, the picture is more complicated, perhaps because of their location at the bottom of the racial hierarchy. On average, as with other groups, African Americans prefer a numerical in-group advantage. But the quantitative analysis offered somewhat of a puzzle in that the most and least desirably integrated neighborhoods on average showed only very slight differences in their racial makeup. Fortunately, coupled with the qualitative insights, a clearer story emerged. The qualitative data reveal that for African Americans, being a numerical minority is not simply perceived as being uncomfortable. Rather, the descriptions provided by African Americans about why they are reluctant to live in certain kinds of integrated neighborhoods are far more vivid and included images of the dangers and disadvantages associated with such a racial composition. The focus for African Americans—as revealed by both the quantitative and especially the qualitative patterns—is not a generic concern about being in the minority, but is specifically related to the size of the White population.
Taken together, all of the patterns for African Americans reveal that Whites are at once both their most and least favored neighbors. On the one hand, Whites are the second largest group in Blacks’ most desired integrated neighborhood. This may reflect the value African Americans’ place in having some Whites—the group at the top of the racial hierarchy—in their neighborhood because they signal a better neighborhood quality because of better services, promise potentially upward mobility, and offer an opportunity to improve race relations. On the other hand, as the qualitative data highlight, a differently integrated neighborhood where there are “too many” Whites tips the balance and opens up the increased possibility of a racially hostile or discriminatory environment. The modest differences in the makeup of the most and least desired neighborhoods may reflect the ambivalence and complexity of sitting at the bottom of the racial hierarchy—being at once the most disadvantaged and most vulnerable to racial discrimination but also recognizing the value of integration to advance both racial justice and personal mobility.
The results of this analysis which draw attention to how the relative numbers of own group and other group, as well as the specific groups and their location on the racial hierarchy, combine to reinforce a central message of this analysis: there is no such thing as ‘one diversity’ or ‘an’ integrated neighborhood. The abstract concept of demographic integration, therefore, conceals a real-world variety of multiethnic and multiracial neighborhoods that are differently valued and perceived depending on what you ask and who you ask. An implication of this is that those who are interested in understanding ‘integrated neighborhoods’ cannot rely on the simple determinations of what constitutes integration that may have worked in the past. The operational complexity and sophistication (new ways to measure and index neighborhoods) has not been met with equally sophisticated conceptualization.
Our results also speak to more general questions of attitudes toward diversity and integration. Taken as a whole, we see that although diversity as an abstract value is widely embraced, the appeal of integration and diversity is not in its ability to transform the system of racial inequality that persists in the United States (e.g., Berrey Reference Berrey2005; Downey Reference Downey1999). In the kind of diversity captured by the ideal neighborhood results, and the explanations provided especially by Whites for why it is desirable, there are similarities and reflections of what Bell and Hartmann (Reference Bell and Hartmann2007) referred to as “Happy Talk.” As they and others have noted, by embracing the seemingly inclusive language of diversity, it becomes possible to gloss over profound and persisting racial inequalities and injustices (e.g., Andersen Reference Andersen, Kenyatta and Totowa1999).
Also as Bell and Hartmann (Reference Bell and Hartmann2007) found, we uncovered ambiguities and tensions surrounding the concept of diversity. In this analysis of attitudes toward integrated neighborhoods, this became evident when we put together the responses to the most and least desired neighborhood. Both the makeup of the integrated neighborhoods that are least desirable and the explanations respondents gave about why they considered them undesirable, reveal that individuals hold and are willing to admit to deeply racialized perceptions about neighbors and neighborhoods.
Our analysis therefore brings into sharp relief the fact that ambivalence toward diversity is not simply with respect to whether diversity, per se, is good or bad (one of Bell and Hartmann’s (Reference Bell and Hartmann2007) key findings). Rather, the ambivalence about diversity and integration that is captured in this analysis has to do with the makeup of the neighborhoods. Neighborhoods integrated with certain groups are acceptable and desirable (and this differs depending on the group); but neighborhoods integrated with other groups are unacceptable and undesirable.
On the one hand, people value the diversity and richness associated with living in neighborhoods with multiple cultures; but on the other hand, negative perceptions and experiences, racial discrimination, and inequality dictate undesirable integration. If we were to limit our analysis to attitudes toward the most desired neighborhood—which is all that has been available to date—we would conclude that there was a strong commitment to the principle of diversity and great interest in ‘integrated’ neighborhoods. But when contrasted with the findings from the least desirable neighborhood question we reveal a more complex picture of integration shaped by racialized perceptions—a picture that reflects the racial hierarchy in the United States, both in terms of where one sits and perceptions of the group with which one might share a neighborhood. Integration is far more complex and problematic than it is often portrayed. The results of this study seek to add grounding and nuance to this prominent feature of contemporary race relations and show how diversity and integration are structured and experienced in real social life—at least insofar as can be gleaned from expressions of attitudes.
Regardless of where you stand on the question of whether segregation is ‘dead’ or not, it is unmistakable that our nation has become more racially and ethnically diverse than in decades past. Although some progress has been made on unpacking how differently configured integrated neighborhoods look in terms of their social and economic characteristics, we have not kept pace in terms of understanding how integrated neighborhoods are perceived. We have not considered what replaces reductions in segregation, and asked ourselves if the resulting ‘integration’ is progress that helps to undo the inequalities that have been tightly wrapped up in the segregated neighborhoods that have defined our nation’s cities and communities for decades. Or if instead it is helping to perpetuate them.
It is no easy task to unpack the diversity of integration, but it is a necessary step, so that we can understand the social consequences of whatever is replacing—or might replace—segregation. And to understand the trajectories of these neighborhoods, it will be important to understand how differently configured ‘integrated’ neighborhoods are perceived and evaluated. To do this, we need to dig deeper into how people experience and understand these ideals of integration when applied to a real-life context like neighborhoods. Our purpose is to encourage and contribute to this conversation. Future research could head in at least two directions. First, social scientists need to begin to seriously tackle this question and develop models and theories that reflect the complexity of integration highlighted here; numbers matter, but so do the specific groups and the mixture therein. Demographers are developing increasingly sophisticated tools and indices intended to capture the realities of neighborhoods. But it is clear that these models need to be informed by theories and social psychological understandings of how these real neighborhoods are imagined and experienced by their residents. For their part, social psychologists also need to incorporate the way in which different mixes of different groups impact perceptions—with a greater understanding of how the racial hierarchy is constructed and what serves to build it up and break it down.
ACKNOWLEDGMENT
The authors gratefully acknowledge the support of the National Science Foundation (SES-0317740), the Ford Foundation, and the University of Illinois at Chicago, which funded the data collection reported in this article. We also appreciate Marco Roc’s research assistance in coding the data. The first author presented earlier versions of this research at Northwestern University, the University of Minnesota, Cornell University, and the University of Illinois at Champaign-Urbana, and would like to thank their audiences, as well the anonymous reviewers of the Du Bois Review for their very useful suggestions and comments. The authors are also very grateful to Doug Hartmann, Matthew Hall, Amanda E. Lewis, and Michael D. M. Bader for reading earlier versions and providing invaluable feedback and guidance.
APPENDIX A
This appendix describes the range of themes that were used in coding the open-ended questions (separately for most and least desired neighborhoods). This is provided for general information, but results for each of the detailed codes are not presented due to space constraints. The paper includes results for those themes that were more frequently mentioned by respondents.
Summary of Detailed Coding Scheme for “Ideal” Neighborhood
Positive Feelings Around In-Group
This is used when the respondent mentions that the reason they like the chosen neighborhood is because of how they would feel in a neighborhood with their own group. This includes mentions of being more comfortable, supported, accepted, empowered, or safer around their own group.
Concerns about Culture
This category includes mentions of cultural differences with various out-groups, cultural similarities with their own group, and a general familiarity with the culture in the neighborhood. Mentions of culture often focus on values, food, religion, language, and beliefs.
Negative Stereotypes and Affect toward Other Racial/Ethnic Groups
This series of codes included negative affect toward, as well as negative stereotypes (cognitions) of, other racial/ethnic groups. This does not include stereotypes about the neighborhoods, but only about members of the groups. For example, “Whites are snobby” is included here; but “a White neighborhood would be snooty” is coded under neighborhood quality (below). This theme was coded so as to permit the identification of which racial/ethnic group was the target of the stereotype/affect (Whites, Blacks, Asians, Hispanics or Arab Americans).
Neighborhood Quality
This category is reserved for remarks about the neighborhood itself; if the respondent’s answer reflects stereotypes about members of various racial groups (as opposed to the neighborhoods they live in) they are coded above, under “negative stereotypes and affect toward racial/ethnic groups” and not here. The range of qualities mentioned includes city services, schools, property values, crime levels, friendliness of the neighbors, the care of the property, and social class characteristics.
Numbers
This theme is used when a respondent indicates that something about the “numbers” in the neighborhood makes it desirable. This code is not used if the respondent was probed and provided an explanation for why the racial distribution (numbers) was ideal.
Desire for Diversity: Fun and Interesting
This theme includes people who say that they want a diverse neighborhood because it would be “fun and interesting” and also includes mentions that it would be “culturally interesting” or because it is “nice” or “good” or “beautiful” or “the spice of life.”
Desire for Diversity: Comfortable with Diversity
This includes responses that mention that either the respondent or their kids will feel comfortable with diversity and living in neighborhoods with different kinds of people. This includes respondents who say that they “can get along with anybody,” or “I like to be around different people,” or “I value diversity.”
Desire for Diversity: Exposure to Other Groups for Self or Children
This includes respondents who say they want diversity so they can get to know each other, or get to know other groups better. This includes mentions that they want diversity or the presence of other groups to learn about them, appreciate their cultures, and so groups can learn from each other. This includes any references to diversity being “enriching” and “because everyone would have different opinions, ideas, and thoughts.”
Desire for Diversity: Creates Racial Harmony
This includes mentions that the respondent wants the neighborhood they drew because it creates racial harmony and helps us “all get along,” and improve race relations. The focus here is on how groups get along with each other and the idea that a diverse neighborhood will reduce prejudice, discrimination, or racism among groups.
Other Reasons for Wanting Mixture/Diversity Not Elsewhere Classified.
This includes any other reasons the respondent gives for liking or preferring diversity that are not already coded above. Some examples include, “in a diverse neighborhood everybody is treated equally,” “everybody has their own voice,” “it’s the way it’s supposed to be,” and “It’s America.” Also included are responses that the neighborhood accurately reflects racial compositions, such as “it reflects the city.”
Summary of Detailed Coding for “Least Desired” Neighborhood
Negative Stereotypes and Affect toward Out-Group
This series of codes is reserved for those responses that include affect toward an out-group as well as stereotypes (cognitions) toward them. This does not include stereotypes of the neighborhoods, but only stereotypes or affect toward the group members themselves. This theme was coded so as to permit the identification of which racial/ethnic group was the target of the stereotype (Whites, Blacks, Asians, Hispanics, or Arab Americans).
Neighborhood Quality
This category is reserved for remarks about the neighborhood itself; if the respondent’s answer reflects stereotypes about members of various racial groups (as opposed to the neighborhood they live in) then they are coded above, under “negative stereotypes” and not included here. The range of qualities mentioned includes city services, schools, property values, crime levels, friendliness of the neighbors, the care of the property, and social class characteristics.
Feelings toward In-Group
This is used when the respondent mentions that the reason they like the chosen neighborhood is because of how they would feel in a neighborhood with their own group. This was further divided into those who mentioned having positive feelings toward their in-group and those who had negative feelings toward their in-group.
Feelings toward Out-Group: Feelings of Discomfort from Out-Group
This theme includes respondents who say that they are not friendly with the out-group or are not comfortable with them. It also includes mentions of isolation, lack of acceptance, or feeling unwanted.
Feelings toward Out-Group: Feelings of Hostility from Out-Group
This includes respondents who specifically connect any feelings of isolation or lack of acceptance to the dynamics of racism or discrimination. It also includes mentions that the respondent feels unsafe (as opposed to merely ‘uncomfortable’) around the out-group, including extreme responses by the out-group, directed to them, including ‘hate,’ ‘hostility,’ or ‘violence’.
Racial Conflict in Neighborhood
This includes respondents who indicate that such a neighborhood would have too much racial conflict or that the people in the neighborhood would not get along because of different cultures or attitudes.
Numbers
This theme is used when a respondent indicates that something about the “numbers” in the neighborhood makes it undesirable, including the mention that the individual doesn’t want to be ‘the only one’ of their group or because there aren’t any of their own group in the neighborhood. This code is not used if the respondent was probed and provided an explanation for why the racial distribution (numbers) was undesirable.
Desire for Diversity
This is a summary code comprised of the same detailed categories for the diversity response that were identified in the coding scheme for the desirable neighborhood. Sub-themes are not presented because of the limited use of this code by respondents who were explaining their least desired neighborhood.
Concerns about Culture
This theme includes respondents who make reference to cultural differences with the out-group and a desire to avoid such differences. Any mention of language, religion, or values is considered cultural. This includes mentions that the respondent says they would not be familiar with the culture or the people in the neighborhood.
Appendix B. Change in Racial/Ethnic Neighborhood Composition between Most and Least Desired Neighborhood, by Most Desired Neighborhood Cluster and Race/Ethnicity of Respondent. 2004–2005 Chicago Area Study.
