Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 1 Contact Hypothesis in Context: Household Characteristics, Community Perception, and Racial/Ethnic Prejudice in the U.S. Amanda Watson Meredith Dye, Ph.D. and Brian P. Hinote, Ph.D. Abstract Although overt racism has diminished, there remain vast racial and ethnic disparities in the United States. Many households are isolated from these disparities simply by where they live. Contact theory hypothesizes that under certain conditions, individual contact with minorities can decrease prejudice. Using a nationally representative sample from the 2000 General Social Survey, this paper explores contact, residential segregation, and anti-minority prejudice in American households. We employ linear regression techniques to analyze the characteristics of White respondents in White households by perception of community composition, region, city size, education, and household type, to identify prejudice against minority groups. Results indicate strong regional effects, with higher levels of prejudice in the South when compared to other regions. Anti-Hispanic prejudice is higher in the Northeast than in the South. Contact theory is not supported, except to show that the effect of more contact is greater on anti-Black prejudice in the Northeast than in the South. Following prior research, education was associated with lower prejudice, and age exhibited a positive relationship with prejudice. We also discuss the general implications of our findings. Scientia et Humanitas 2 Spring 2011 INTRODUCTION The twenty-first century is a time of nearly unprecedented diversity when it comes to race and ethnicity in the United States (e.g., Larsen, 2004; McKinnon, 2003; Ramirez & de la Cruz, 2002; Reeves & Bennett, 2003). As Jeffrey Dixon (2006) succinctly notes, “if the problem of the 20th century was that of the color line, then the problem of the 21st century is that of increasingly numerous color lines.” As a result, the character and quality of interactions among various racial and ethnic groups are of particular interest to sociologists. Current demographic trends permit continued development in the area of social theory, but perhaps more importantly, there is also the opportunity to peer more deeply into important racial/ethnic issues facing us today, like immigration, prejudice, and many others. Social science research emphasizes two contrasting theoretical backgrounds in analyzing the relationship between racial/ethnic composition and prejudice (Ulrich et al., 2006). The first deals with social contact (see Allport, 1954), while the second focuses on threat effects (see Blalock, 1957). Previous findings offer evidence both in support and in critique of these frameworks (see Dixon, 2006), but more recent analyses (e.g., Crisp & Turner, 2009; Pettigrew & Tropp, 2008) focus current research efforts more and more upon the idea of social contact and its effects upon phenomena like prejudice and racism. This paper investigates racial- and ethnic-based prejudices in non-Hispanic White households, and examines the ways that contact with minorities affects prejudice in the United States. Specifically, this research looks to disentangle prejudice against Blacks, Hispanics, and Asians from the characteristics of residential segregation. While much research focuses on institutional discrimination, and prior research has shown that overt racism is in decline, the racial and ethnic disparities existing within the United States warrant further exploration ( Johnson & Jacobson, 2005). In addition, the recent election of the first African-American president in the U.S. has renewed public interest in race and has increased race and ethnic dialog in media outlets. Immediately after the election of President Barack Obama, a Gallup poll found that 67% of Americans thought that a solution to relations between Blacks and Whites would eventually be worked out. This was the highest percentage Gallup has measured on that particular question (Gallup, 2008). While that is certainly positive news, we cannot so quickly dismiss the past or disregard the current situations of minorities in the U.S. In the 1950s, “White flight” allowed Whites to flee urban neighborhoods, leaving behind Blacks who did not have the same opportunities or life chances, thus increasing segregation. Trends show that the Black middle-class has recently increased in size, and although many of its members remain in the central city, some have begun moving into the suburbs. However, this does not necessarily translate into equality. Research also indicates that Blacks, even if they are in the middle-class, still tend to live in poorer neighborhoods than their White counterparts (Logan, 2002). Hispanic populations are rapidly expanding as well, along with Hispanic presence within many communities. The Pew Hispanic Center (2009) reports that the Hispanic populations of the United States (native- and foreign-born) represent fifty percent of the increase in population from 2000 to 2007. Little research has investigated the effects of White non- Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 3 Hispanic and Hispanic prejudice or contact (Weaver, 2007). Asians, on the other hand, are considered a ‘model minority’ and experience higher levels of integration within American society (Gans, 1999). Is there a basis for anti-minority prejudice in White households, since there is such a high level of residential segregation in the United States? And is prejudice mediated by the perception of a higher level of minority presence by the respondent within his or her community? These are questions of particular interest in this research. This study evaluates the assumptions and propositions of contact theory by looking at selected characteristics of non-Hispanic White households in the United States (e.g., region, city size, family income, household type, and perception of community racial/ ethnic composition) to identify relationships between these characteristics and prejudice. Anti-Black, anti-Hispanic, and anti-Asian prejudice are analyzed separately for a more comprehensive study. This research seeks to uncover the extent to which anti-minority prejudice still exists among Whites, and if the perception of community racial/ethnic composition affects that prejudice in any observable way. The research conducted here is important because we have not yet clarified the causes and correlates of residential segregation in U.S. society. Doing so could feasibly inform future economic and social policies capable of addressing the causes persistent forms of residential segregation and prejudice. CONTACT AND RESIDENTIAL SEGREGATION In considering solutions to racial/ethnic prejudice, Allport (1954) hypothesizes that on an individual level, meaningful contact with another individual of a different race or ethnic group decreases prejudiced attitudes, further specifying that so-called ‘meaningful contact’ exhibits specific characteristics that must exist for the reduction in prejudice to occur. As Johnson and Jacobson (2005:388) explain, positive outcomes to social contact occur only (1) in cooperative events, (2) among participants of equal status, (3) between those who hold common goals, and (4) with those who have supportive authority. Allport’s contact hypothesis has since been widely tested and is now recognized as a theory (Pettigrew, 2008). However, is contact theory based on individuals without prejudice having more contact with minority individuals? Pettigrew (2008), in a review of the literature, claims that this is not the case. As for residential segregation, Allport (1954:263) notes that, “Where segregation is the custom, contacts are casual, or else firmly frozen into superordinate-subordinate relationships.” This is negative contact, and the vast inequalities that exist as a result of dilapidated and segregated minority neighborhoods begin to appear as characteristics of that particular race (Allport, 1954). However, while Whites may be highly segregated at their residences, it follows that contact will become more and more unavoidable as minority populations continue to expand. It is easy to point to the Civil Rights Movement or the Fair Housing Act as indications of dismantling segregation. However, these very public actions did not solve covert discrimination and racism (Massey & Denton, 1993) and residential segregation persists. Scientia et Humanitas 4 Spring 2011 Recent and previous studies (e.g., Clark & Blue, 2004; Freeman, 2008; Massey & Denton, 1993) emphasize economic status and interpersonal or institutional discrimination in explaining the persistent gap between White and minority neighborhoods. The economic explanation is perhaps the most straightforward: Blacks and Whites differ, on average, in both income and wealth, with Blacks falling behind some in the former and even more in the latter (Oliver & Shapiro, 1997). Whites, it follows, can afford higher quality housing, which is coincidentally located in different neighborhoods than housing that is financially available to Blacks. According to the spatial assimilation model, this condition should be temporary; as racial and ethnic minorities increase their levels of education, occupational prestige, and income, integration should follow (Charles, 2003). For example, Asians have been better able to assimilate into American society, presumably because of their higher education levels (Gans ,1999). The place stratification model, on the other hand, more fully considers the effects of institutional discrimination, emphasizing the “persistence of prejudice and discrimination – key aspects of inter-group relations – that act to constrain the residential mobility options of disadvantaged groups, including supraindividual, institutional-level forces” (Charles, 2003:3). In the past, this included ‘redlining’ throughout the United States (Wilson, 2008), or selecting out primarily Black neighborhoods from any mortgage, regardless of the socioeconomic status of the individual or family. Through government programs initiated after World War II, Whites were encouraged to move out into the suburbs and buy homes, leaving dilapidated urban areas behind (Massey & Denton, 1993). Institutional discrimination also affects housing searches, encouraging Blacks to settle or stay in predominantly Black neighborhoods even if they could afford to move into the White suburbs (Gans, 1991; Lake, 1981; Williams & Williams-Morris, 2000). Indeed, Logan (2002) analyzes 2000 decennial census data and uncovers a neighborhood gap between Black and Hispanic minorities and Whites. As Whites earn more money, they move to neighborhoods that correspond with their income. In 2000, the average White household with an income of $60,000 or more lived in a neighborhood where the median income was above $64,000. However, the average Black and Hispanic household with an income of $60,000 or more lived in neighborhoods where the median income was below $50,000 (Logan, 2002:4). This same trend follows in the South, with the average White household earning $60,000 or more living in a neighborhood where the median income was $52,000 and the average Black household living in a neighborhood where the median income was $41,918. Based on these data, Logan (2002:7) concludes that this gap is “not merely a reflection of income differences between the races. Comparable Whites and Blacks face a very different structure of opportunities about where to live, yielding considerable advantage to Whites.” This is not the result of Blacks’ preference to stay in their familiar poor and segregated neighborhoods. Like other racial and ethnic groups, African Americans aim to pair social and geographic mobility, moving to better neighborhoods as they become Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 5 financially able, but they are nonetheless more likely to stay in poorer areas (Patillo-McCoy, 1999:23). As Quillian (2002) summarizes, income differentials and institutional discrimination do not fully account for contemporary levels of segregation in the United States. This and other research (e.g., Clark, 1991; Krysan, 2002; Zubrinsky & Bobo, 1996) turns our attention to preferences, or the way that personal desires (particularly the desire to live near neighbors who are of the same or a different racial and ethnic background) shape choices of home and neighborhood. These preferences are measured in diverse ways, both directly – asking respondents about their willingness to buy in real or hypothetical neighborhoods of varying racial compositions – or indirectly, by modeling moves to and from neighborhoods with varied racial compositions, and inferring preferences as motives behind these moves. This latter line of inquiry documents actual patterns of White flight, linking them with neighborhood composition. For example, Quillian (2002) shows that as the number of Blacks in a census tract rises, Whites’ probability of moving to a different census tract increases, as does probability of moving to a different type of tract – one with fewer Black residents. “Whites move to neighborhoods that are Whiter than their origin and are by far most likely to move to the Whitest possible destinations. This is consistent with concerted efforts by Whites to avoid Black neighbors” (Quillian, 2002:212). In a similar study, Crowder (2000) finds that regardless of other individual- and neighborhood- level predictors of mobility, Whites living in minority neighborhoods are more likely to move out than those in Whiter areas. The availability of White destination tracts has a positive and significant effect on Whites’ probability of moving, as does the presence of recently constructed housing. Although this research suggests that Blacks prefer to live in racially-mixed and upwardly mobile neighborhoods, the racial composition of their current neighborhood influences the likelihood of moving and reflects movement toward racial homogeneity. In other words, Black residents are most likely to switch census tracts if they are in predominantly White areas, often moving to blacker neighborhoods (Quillian, 2002; Zubrinsky & Bobo, 1996). Since the findings on preference point away from racial homogeneity, there are likely other factors preventing racially diverse neighborhoods. While these studies focus on mobility patterns, they offer little insight into the perception of race itself. Race is not just a factor in the perception of a particular community, it is also a factor in Blacks’ and Whites’ awareness that a community exists. Asking about communities in several metropolitan areas (of varying social, economic, and racial composition), Krysan (2002) found that Blacks consistently classified each community as more desirable than did White respondents. However, when Black respondents had information on a possible negative racial climate of the area (for example, Whites discriminating or acting out against Black neighbors), the desirability changed. Black respondents were also less likely to know about neighborhoods that were further away from the core of the metropolitan area than Whites, suggesting that “housing searches for Blacks may be biased away from suburban- whiter- communities about which they have less knowledge; Whites’ housing searches may be biased away from racially mixed communities, in part, for the same reason” (Krysan, Scientia et Humanitas 6 Spring 2011 2002:526). Furthermore, Blacks may assume that predominantly White areas about which they have little other knowledge are more racist, leading them to prefer more mixed neighborhoods (Krysan, 2002:537). Whether or not areas have explicitly racist histories or local cultures, middle class Blacks often experience both overt and subtle discrimination in predominantly White neighborhoods (Feagin, 1991), and some may prefer more mixed areas for this reason. Overt discrimination may include such clear discriminatory acts as prejudiced comments, while subtle discrimination may include avoidance or exclusion from neighborhood activities. Given the research on continuing forms of discrimination and residential segregation, are Whites’ preferences rooted in racism, or based on the correlations between neighborhood racial composition and property values, crime, or school quality? Is the avoidance of Black neighbors simply a reflection of the desire to avoid the problems associated with poor neighborhoods of color? Emerson, Yancey, and Chai (2001) investigate these factors by presenting respondents with a hypothetical ideal house, and controlling for factors including school quality, crime, housing values, and racial composition. The first three variables were always strong predictors of the willingness to buy the home, but as far as racial composition of neighborhood (which was presented as proportionately White and one other racial group), only the presence of Black residents impacted the respondents’ willingness to buy. The likelihood to purchase declined at all levels of Black population, particularly for respondents who had minor children. Race, it seems, does shape housing preferences, in both the kinds of preferences held by racially differentiated groups and the types of neighborhoods that individuals prefer. This research suggests that Blacks may still experience discrimination among those who live in a segregated neighborhood. Within suburban neighborhoods, we see trends toward gated communities that separate these residents from surrounding areas. These communities cater to families and offer amenities geared toward households with children. The areas surrounding these communities have also refused public housing, further insulating themselves from urban problems and surrounding themselves with wealth and prestige ( Jackson, 2000). On the other hand, Carter et al. (2005) found that city residents have more tolerant views toward Blacks than non-city residents. When they measured views by region, they also found that Southerners were less likely to be tolerant of Blacks. This discrimination could be rooted in prejudice, but the current analysis aims to disentangle prejudice from the other previously mentioned causes of residential segregation. RACIAL AND ETHNIC PREJUDICE This study follows the definition of prejudice provided by Jeffrey Dixon (2006), and used in his prejudice index, which includes a combination of the traditional meaning and elements dealing with the emergence of laissez-faire racism. “Prejudice has traditionally been defined as ‘an antipathy based upon a faulty and inflexible generalization.’” It has also come to include the feelings that “a minority group is inferior, different, alien and threatening to one’s own racial/ethnic group” (Dixon, 2006:2180). However, research indicates that overt Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 7 racism (called Jim Crow racism, based on the post-Civil War South) has declined because it goes against American values (Bobo, Kleugal & Smith, 1997). Rather, a new racism, termed ‘laissez faire,’ has emerged. Bobo et al. (1997) distinguish this racism from other types in that it removes social responsibility from the continued economic disadvantages for Blacks, including residential segregation. “In short, a large number of White Americans have become comfortable with as much racial inequality and segregation as a putatively nondiscriminatory polity and free market economy can produce: hence the reproduction and, on some dimensions, worsening of racial inequalities” (Bobo et al., 1997:41). Laissez-faire racism lends itself well to the prejudice index, as it measures attitudes toward qualities Whites might believe Blacks or other minority groups to possess as the reasoning behind their lack of economic mobility (e.g., laziness). While residential segregation is primarily focused on the White/Black dichotomy, other research suggests that Whites flee neighborhoods when any minority group representation increases (Crowder, 2000:226). Gans (1999) presents the possibility of a new racial hierarchy, which adds a third category that does not fit into the Black/non-Black dichotomy that exists today. One of the first minority groups who seemingly have been able to assimilate into the non-Black dichotomy includes part of the Asian and Asian-American population. Gans (1999:267) terms this group the “model minority,” as they have been able to “eradicate many of the boundaries between it and Whites.” It remains to be seen what will occur with the Hispanic population, as some groups within the Hispanic ethnicity are able to ‘pass’ as non-Black, while others have darker skin color that could be labeled as Black. Hence, there is a need to include Asian and Hispanic prejudice in the current study. Prior research on prejudice suggests that while Whites may not have completely negative attitudes toward Blacks, it does not follow that they necessarily have positive attitudes (Federico, 2006:345). For example, a White person may not be completely prejudiced against a Black person, or hold entirely prejudiced attitudes towards Black people, but that does not mean that no prejudice exists. Therefore, questions that ask respondents to make a choice strictly between positive and negative prejudice may be missing a chance to delve deeper into the issue. Conversely, this research relies on a scale, which takes this point into consideration. Another previous study (Carter et al., 2005) indicates that two types of modules are needed to fully test for prejudice. The first should contain questions based on principles which literally assess a person’s principles or beliefs regarding race and equality. The second module should contain questions on implementation, which typically involve feelings on government interaction with minority groups. This includes questions regarding support for affirmative action or enforcement of equal opportunity housing laws. Our study contains questions regarding principle, since implementation questions were not asked on this particular topic in the 2000 General Social Survey. Scientia et Humanitas 8 Spring 2011 DATA AND METHODS The data for this study comes from the 2000 National Opinion Research Center’s General Social Survey (GSS). GSS samples are nationally representative of English- speaking persons eighteen years of age or over living in non-institutional households in the United States (see Davis & Smith, 1992). Respondents for the 2000 survey were chosen using probability sampling, and the survey was administered to 2,817 respondents. Only a portion of the questions used to create the prejudice index were asked of a subset of the full sample, so we removed all cases that were not asked each question, as well as those not asked about prejudice. We then selected White respondents from White households and adjusted our sample for missing cases. This left a total of 789 cases analyzed for anti-Black prejudice, 774 cases for anti-Hispanic prejudice, and 770 cases for anti-Asian prejudice. Hypotheses Previous research (e.g., Clark, 1991; Krysan, 2002; Zubrinsky & Bobo, 1996) into the White flight phenomenon explores residential preferences and current and/or previous neighborhood racial composition, but has failed to look at the characteristics of households and their relationship to racial and ethnic group prejudice. This analysis investigates the relationship between characteristics of non-Asian and non-Hispanic Whites in White households and perceptions of racial/ethnic community composition with racial and ethnic group-based prejudices, and what role contact plays. The characteristics measured here include: 1) region, 2) household type, 3) city size, 4) education of the respondent and 5) family income of the respondent, and we specifically examine White respondents’ anti- Black, anti-Hispanic, and anti-Asian prejudice. We first hypothesize that White respondents in Southern White households have higher levels of anti-Black, anti-Hispanic, and anti-Asian prejudice than households in the Northeast, Midwest, and West. This is based on previous research (e.g., Crowder, 2000) indicating that White flight occurs more prominently in the South. It also follows evidence of historical racism in this region, and Carter et al.’s (2005) findings of more racist attitudes among southerners. H1: White respondents in White households in the South will have higher levels of anti-Black, anti-Hispanic, and anti-Asian prejudice, compared to households in other regions. Following the propositions of contact theory (Allport, 1954), we next hypothesize that lower levels of racial/ethnic presence within the community are associated with higher levels of racial/ethnic prejudice net of all household characteristics. H2: As perceived levels of racial/ethnic presence increase, anti-Black, anti- Hispanic, and anti-Asian prejudice will decrease. Due to the importance of the variable used to test contact theory and the preliminary results showing highly significant results by region, we also test for interaction effects. Since Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 9 contact theory is grounded in place, we examine interaction between variables representing place (i.e., region) and community composition. We hypothesize that prejudice is explained by the interaction of perception of community composition by race and the region where they live. H3: The effect of perceived levels of racial/ethnic presence on anti-Black, anti- Hispanic, and anti-Asian prejudice will vary by region. We next posit that households with children exhibit higher levels of prejudice than households without children. This is based on the characterization that families participate in White flight, and as Jackson (2000) explains, communities located in White suburbs are geared toward households with children. In addition, Emerson et al. (2001) report that Whites with children are more likely to decline purchase of a home within a neighborhood at any level of Black population. Therefore, it follows that individuals with children would be more likely report prejudice and to flee neighborhoods that have a large minority population. H4: White respondents with children in White households will have higher levels of anti-Black, anti-Hispanic, and anti-Asian prejudice, compared to households without children. Because White flight occurs toward the suburbs (e.g., Crowder, 2000), we next hypothesize that White respondents in suburban households exhibit greater prejudice than those in city or rural areas. Blacks have not relocated to the suburbs at the rate of Whites either due to racial factors or institutional discrimination (Logan, 2002). Since these neighborhoods tend toward more homogeneity, perhaps the White residents of suburbs exhibit higher levels of prejudice against minorities. H5: White respondents in suburban White households will have higher levels of anti-Black, anti-Hispanic, and anti-Asian prejudice, compared to households in urban areas. Finally, because white flight is typically a middle-class phenomenon, prejudice should decrease as education and family income increases. While White neighborhoods tend to have higher education and family income levels than minority neighborhoods (Logan, 2002), the more educated persons are, perhaps the less prejudiced they are as well. H6: As education and family income of White respondents in White households increase, anti-Black, anti-Hispanic, and anti-Asian prejudice will decrease. In summary, this study tests the idea that education, household type, city size, region, and perception of community composition among White respondents in White households influence the degree to which Whites hold anti-Black, anti-Hispanic, and anti-Asian attitudes. Scientia et Humanitas 10 Spring 2011 Dependent Variables Dependent variables are based on the prejudice index developed by Jeffrey Dixon (2006). Rather than ask a strict yes or no question on prejudice, the index is based on scaled questions, which permits a deeper look into the issue of prejudice. A scale allows the respondent to provide an answer within a range of responses, so there may be less of a stigma associated, rather than just choosing a response that says “yes, I am prejudiced toward….” The first index item asks respondents to rate on a scale of 1 to 7 whether or not they think a certain group is committed to strong families, with 1 being a strong commitment and 7 being no strong commitment. The next item asks whether they think a certain group is hard working (1) or lazy (7). The final item asks whether they think a certain group is unintelligent (1) or intelligent (7). The question on intelligence is reverse coded (negative to positive) following the other items used in the scale. The prejudice index is based on previous research indicating that Whites position their views of other groups in reference to how they view themselves (e.g., Blumer, 1958). Therefore, Dixon (2006) subtracted Whites’ placement of each minority group from the placement of their own group, then combined scores to create scales. Higher scores indicate the White respondent rated the other race more negatively than their own, indicating greater prejudice. A score of zero indicates that the White respondent rated the other race at the same level as his or her own, indicating no prejudice. Negative scores indicate that White respondents rated the other race more positively than their own, indicating they hold members of the other race or ethnic group in higher esteem than their own (see Figure 1). Figure 1 Independent Variables We also employ a series of variables assessing the racial/ethnic make-up of respondents’ local communities. Respondents were asked, “Just your best guess - what percentage of the people who live in your community is each group?” Since we examine anti-Black, anti- Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 11 Asian, and anti-Hispanic prejudice, we use variables that gauge minority presence within the respondent’s community. Respondents reported a percentage between 0-100, and we employ composition variables corresponding to the respective dependent variable (e.g., percent Black composition is used with anti-Black prejudice, etc.). Based on prior research (e.g., Allport, 1954; Carter et al., 2005; Crowder, 2000; Emerson et al., 2001; Jackson, 2000; Krysan, 2000; Logan, 2002), we operationalize the characteristics of White households as independent variables. We first include region, which coded as a series of dummy variables for various sections of the United States - West, Midwest, Northeast, and finally the South, which is used as a reference group. This schema is based on definitions provided by the U.S. Census Bureau (see http://www.census.gov/geo/www/us_regdiv.pdf ). Size of place is measured at the nominal level, and is dummy coded into categories for city, suburb, other city, and rural households, with suburb used as the reference group. Household type is measured at the ordinal level. We dummy coded variables for married with children households, not married with children households, and households without children, and use married with children households as the reference group. Accordingly, the reference group in each model consists of married Whites who live in Suburban households in the South with children. We employ family income and education measures to operationalize social class. Education is measured in years of education, and family income is measured at the ordinal level. Respondents were asked to group their total family income before taxes from the previous year. In the GSS, family income is grouped according to midpoints, making it an ordinal-level variable (see Ligon, 1988 for further discussion), but due to significant skewness and kurtosis, we logged the income variable to yield a more manageable distribution for analysis (see Tabachnick & Fidell, 2007). We also include control variables consistent with other studies of racial attitudes, including age (in years) and sex (e.g., Dixon, 2006; Krysan, 2000). Preliminary analysis revealed several missing cases for family income, and we replaced these cases with the sample mean income of $38,331.93. Accordingly, we created a flag variable for the missing cases to use as a control in regression equations. Sensitivity Analysis We analyze these data using cluster analysis, conducted in STATA, and linear regression, conducted in SPSS. Since households are nested within region, we also employ cluster analysis to detect differences in prejudice across region that might also explain differences across households. This is to ensure any significant differences at the household level do not actually reflect regional differences. This test detects bias in standard errors within the linear regression model, and allows the researcher to correct the standard errors if clustering does indeed exist (Long & Freese, 2003). The statistical models for the cluster analysis are as follows: Scientia et Humanitas 12 Spring 2011 Anti-Minority Prejudicej = A + B1Northeastj + B2Midwestj + B3Westj + B4Community Compositionj + B5Household Not Married with Childrenj + B6Household No Childrenj + B7Cityj + B8Ruralj + B9Other Urbanj + B10Educationj + B11Family Incomej + B12Income Flagj + B13Agej + B14Femalej + uj where uj = Σ ei*xi jcluster Regression Analysis Cluster analyses did not suggest significant standard error biases, and we thus proceeded with our regression models. We conduct models independently for each outcome variable, with regression equations shown below (anti-minority prejudice is substituted for each separate dependent variable): Model 1: Anti-Minority Prejudicej = A + B1Northeastj + B2Midwestj + B3Westj + B4Community Compositionj + B5Household Not Married with Childrenj + B6Household No Childrenj + B7Cityj + B8Ruralj + B9Other Urbanj + B10Educationj + B11Family Incomej + uj Model 2: Anti-Minority Prejudicej = A + B1Northeastj + B2Midwestj + B3Westj + B4Community Compositionj + B5Household Not Married with Childrenj + B6Household No Childrenj + B7Cityj + B8Ruralj + B9Other Urbanj + B10Educationj + B11Family Incomej + B12Income Flagj + B13Agej + B14Femalej + uj Model 3: Anti-Minority Prejudicej = A + B1Northeastj + B2Midwestj + B3Westj + B4Community Compositionj + B5Community Composition*Northeastj + B6Community Composition*Midwestj + B7Community Composition*Westj + B8Household Not Married with Childrenj + B9Household No Childrenj + B10Cityj + B11Ruralj + B12Other Urbanj + B13Educationj + B14Family Incomej + B15Income Flagj + B16Agej + B17Femalej + uj Model one tests all hypotheses by separately testing for the difference between means for anti-minority prejudice among the household characteristics of region, perception of community minority composition, household type, city size, and education. Again, we analyze each dependent variable independently of the others (anti-Black prejudice, anti-Hispanic prejudice, and anti-Asian prejudice) and community minority composition corresponds to Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 13 the prejudice being tested (e.g., perception of percent Blacks in the community with anti- Black prejudice). Model two examines control variables, and estimates the effect of anti- minority prejudice among the household characteristics of region, perception of community minority composition, household type, city size, and education while controlling for sex, age, and family income. This helps ascertain whether the effect of household characteristics and perception of community composition remain significant on anti-minority prejudice while simultaneously holding individual-level variables constant. Finally, model three tests interaction effects. Preliminary analysis suggested a strong regional effect over the other independent variables, so we included regional interaction effects with the perception of community racial composition to look more closely at the relationship between the two variables tied to place in relation to contact theory. Model three estimates the impact of household characteristics and perception of community composition on anti-minority prejudice and introduces an interaction term to determine if the effect of perception of community minority composition on prejudice differs by region. This is important because the aspect of contact theory that this study explores is grounded in place, and the measure of location in this study is region. Contact theory posits that interaction with minorities decreases prejudice, and the region in which a White household is located may have an impact on the perception of community composition. RESULTS Table 1 presents means and standard deviations for each variable in the analysis. Tables 2 through 4 present regression results and are discussed below. We discuss results from the third and completed model, unless otherwise noted. Scientia et Humanitas 14 Spring 2011 Table 1. Means and Standard Deviations of Prejudice Indices, Perception of Community Composition, and Household Characteristics (n = 798)   Non-Hispanic and Non-Asian White Respondents   Mean Std. Dev. Minimum Maximum Anti-Black Prejudice 2.233 3.449 -6 18 Anti-Hispanic Prejudice 0.940 3.162 -8 13 Anti-Asian Prejudice -0.486 3.073 -10 16 Region   South 0.310 0.463 0 1 Northeast 0.182 0.386 0 1 Midwest 0.267 0.443 0 1 West 0.242 0.428 0 1 Community Composition   Estimated Percent Black 14.280 15.553 0 99 Estimated Percent Hispanic 13.850 16.791 0 97 Estimated Percent Asian 6.760 9.708 0 85 Household Type   Married with Children 0.267 0.443 0 1 Not Married with Children 0.098 0.297 0 1 No Children 0.635 0.481 0 1 City Size   City 0.173 0.378 0 1 Suburb 0.269 0.444   Rural 0.112 0.315 0 1 Other Urban 0.446 0.497   Family Income (Logged) 10.173 0.964 5.81 11.86 Education 13.750 2.646 3 20 Age 44.510 16.481 18 89 Female 0.524 0.500 0 1 Income Flag 0.073 0.259 0 1 Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 15 Anti-Black Prejudice Regression results indicate that region, the product of community composition and region, education, age and sex are significant. While controlling for individual-level characteristics, anti-Black prejudice in the households located in the Midwest (B = -0.145) is significantly lower when compared to anti-Black prejudice in households located in the South (p ≤ .05). Introducing the interaction terms gives a more detailedpicture of the effect of region on anti-Black prejudice. The percent Black composition/Northeast interaction term is significant at the .01 level (B = 0.162). The positive coefficient for the product term indicates that the difference in levels of prejudice between regions is greater when respondents report a higher perceived percentage of Blacks in their communities. This means that the effect of perceived percentage of Blacks living in the community area has a more pronounced effect on prejudice towards Blacks in the Northeast rather than the South. Thus, as respondents’ perceptions of the percentage of Blacks living in their community increases, levels of anti-black prejudice decrease, but the effect is only significant for the Northeast region. Results also indicated that as the level of education increases for the White respondent, anti-Black prejudice decreased by 0.184 (p ≤ .001), controlling for the other household factors, perception of Black community composition, and individual characteristics. In addition, the individual-level control variables were significant. Age had a positive relationship with anti-Black prejudice. As age increases, anti-Black prejudice increases (B = 0.226, p ≤ .001). Females have lower levels of anti-Black prejudice than Males (B = -0.107, p ≤ .01). Model 1, without controlling for household factors and individual characteristics, had an R2 of 0.053, meaning that 5.3% of the variance in anti-Black prejudice is explained by the variance in the household variables. In further models that included control variables and interaction terms, the R2 increased to 11.1% and then to 12.2% (see Table 2). Scientia et Humanitas 16 Spring 2011 Table 2. Standardized Coefficients from OLS Regression of Anti-Black Prejudice for White Respondents in White Households (n = 789)   Model 1   Model 2   Model 3   Region   South --- --- ---   Northeast 0.000 -0.013 -0.119 * Midwest -0. 075 -0.078 * -0.145 ** West -0.066 -0.057 -0.077   Community Composition, Percent Black -0.010 0.043 -0.058   Interaction Terms   Percent Black * South --- --- ---   Percent Black * Northeast --- --- 0.162 ** Percent Black * Midwest --- --- 0.086   Percent Black * West --- --- -0.009   Household Type   Married with Children --- --- ---   Not Married with Children -0.020 0.010 0.000   No Children 0.024 -0.039 -0.042   City Size   Suburb --- --- ---   City -0.081 * -0.053 -0.048   Rural -0.028 -0.018 -0.014   Other Urban -0.112 ** -0.082 -0.067   Education -0.191 *** -0.185 *** -0.184 *** Family Income (logged) --- 0.040 0.046   Income Flag --- 0.038 0.042   Age --- 0.224 *** 0.226 *** Female --- -0.112 ** -0.107 ** Constant 6.369 *** 3.027 3.138 * R2 0.053   0.111   0.122   *p ≤ .05, **p ≤ .01, ***p < .001 Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 17 Anti-Hispanic Prejudice The regression results in Table 3 indicate that region, education, and age are significant. While controlling for other individual-level factors, anti-Hispanic prejudice for White respondents in White households located in the Northeast (B = 0.085) is significantly higher when compared to anti-Hispanic prejudice in the South (p ≤ .05, Model 2). As for the West, there are significantly lower levels of anti-Hispanic prejudice for White respondents in White households than in the South even with the individual-level characteristics (p ≤ .05). Whites living in White households in the West are 0.100 lower on the anti-Hispanic prejudice scale, controlling for other factors. Education was statistically significant, even with the control variables. For every year increase in education, anti-Hispanic prejudice decreased by 0.166 (p ≤ .001). Age, one of the individual-level control variables, was also highly significant. As age increases, anti- Hispanic prejudice increases (B = 0.228, p ≤ .001). Model 3 did not produce significant interaction effects for anti-Hispanic prejudice. For Model 1, without controlling for household factors and individual characteristics, had an R2 of 0.070, meaning that 7.0% of the variance in anti-Hispanic prejudice is explained by the household variables. With the inclusion of control variables in Model 2, the R2 increased to 11.8% (see Table 3). Scientia et Humanitas 18 Spring 2011 Table 3. Standardized Coefficients from OLS Regression of Anti-Hispanic Prejudice for White Respondents in White Households (n = 774)   Model 1   Model 2   Model 3 Region   South --- ---   --- Northeast 0.096 * 0.085 * 0.082 Midwest -0.040 -0.050   -0.058 West -0.087 * -0.100 * -0.040 Community Composition, Percent Hispanic -0.019 0.004   0.035 Interaction Terms   Percent Hispanic * South --- ---   --- Percent Hispanic * Northeast --- ---   0.009 Percent Hispanic * Midwest --- ---   0.023 Percent Hispanic * West --- ---   -0.093 Household Type   Married with Children --- ---   --- Not Married with Children -0.017 -0.002   0.001 No Children 0.041 -0.022   -0.025 City Size   Suburb --- ---   --- City -0.077 -0.055   -0.057 Rural 0.008 -0.001   -0.001 Other Urban -0.100 * -0.082   -0.076 Education -0.175 *** -0.162 *** -0.166 *** Family Income (logged) --- 0.004   0.006 Income Flag --- 0.033   0.034 Age --- 0.228 *** 0.230 *** Female --- -0.029   -0.033 Constant 4.173 *** 2.114   2.013 R2 0.070   0.118   0.122 *p ≤ .05, **p ≤ .01, ***p < .001         Anti-Asian Prejudice Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 19 The regression results in Table 4 indicate that region, size of place, education, and age are significant. While controlling for other individual-level factors, anti-Asian prejudice for White respondents in White households located in the West (B = -0.112) is significantly lower when compared to anti-Asian prejudice in the South (p ≤ .01). In other words, for each White respondent located in a White household in the West as opposed to the South, anti-Asian prejudice decreased by 0.112, controlling for other factors. In addition, White respondents living in White households located in rural areas had significantly higher levels of anti-Asian prejudice compared to those located in suburban areas of the largest SMSA’s, net of individual-level control variables (B = 0.099, p ≤ .05). Education was also found to be significant. For every year increase in education, anti- Asian prejudice decreased by 0.195 (p ≤ .001). In addition, one of the individual level control variables was also significant. Age had a positive relationship with anti-Asian prejudice. As age increases, anti-Asian prejudice increases (B = 0.087, p ≤ .05). Again, Model 3 did not produce significant interaction effects for anti-Asian prejudice. For Model 1, without controlling for household factors and individual characteristics, had an R2 of 0.095, meaning that 9.5% of the variance in anti-Asian prejudice is explained by the variance in the household variables. In the second model, which included control variables, the R2 increased to 10.4% (see Table 4). Scientia et Humanitas 20 Spring 2011 Table 4. Standardized Coefficients from OLS Regression of Anti-Asian Prejudice for White Households (n = 770)   Model 1   Model 2   Model 3 Region   South --- ---   --- Northeast 0.071 0.067 0.088 Midwest -0.037 -0.044   -0.062 West -0.120 ** -0.126 ** -0.112 * Community Composition, Percent Asian -0.026 0.034   0.055 Interaction Terms   Percent Asian * South --- ---   --- Percent Asian * Northeast --- ---   -0.044 Percent Asian * Midwest --- ---   0.035 Percent Asian * West --- ---   -0.031 Household Type   Married with Children --- ---   --- Not Married with Children 0.025 0.020   0.023 No Children 0.037 0.000   0.000 City Size   Suburb --- ---   --- City -0.037 -0.038   -0.040 Rural 0.101 0.096  * 0.099 * Other Urban 0.052 * 0.048   0.047 Education -0.214 *** -0.192 *** -0.195 *** Family Income (logged) --- -0.054   -0.048 Income Flag --- 0.029   0.028 Age --- 0.087 * 0.091 * Female --- -0.031   -0.033 Constant 2.647 *** 3.598 **  3.407 * R2 0.095   0.104   0.107 *p ≤ .05, **p ≤ .01, ***p < .001         Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 21 DISCUSSION Previous research has not directly examined the link between household characteristics, contact theory, and levels of anti-minority prejudice for Whites. We do know that residential segregation exists and that this could be a result of preference, institutional access, or perhaps prejudiced views toward minorities. Segregation allows Whites to isolate themselves from minority groups. Is this isolation associated with racial/ethnic prejudice? In addition, as minority populations continue to grow, it follows that it will become more difficult to limit contact. This study used data from the 2000 General Social Survey to ascertain any relationship between residential segregation as measured by certain household characteristics of whites, contact with minorities, and racial prejudice. Based on the analysis, the hypothesis that anti-minority prejudice would be greater among households located in the South was supported, except in one scenario. There were significant mean differences, and the mean of anti-Black prejudice for White respondents in White households was lower in the Midwest when compared to the South. This follows previous research on residential segregation, which found that White flight occurs more predominately in the South (Crowder, 2000). In addition, the mean of anti-Hispanic and anti-Asian prejudice in White households was significantly lower in the West when compared to the South. This finding is also consistent with previous research that found higher racist attitudes among Southerners (Carter et al., 2005). For this study, it means that there is some relationship that exists between the regional distribution of White households and racial/ethnic prejudice. There was a significant interaction effect with perception of Black community composition in the Northeast when compared to the South. An increase in perception of Black community composition for White respondents located in the Northeast was associated with a decrease in anti-Black prejudice, when compared to households located in the South. This indicates that contact with Blacks matters more in the prediction of lower prejudiced views among Whites in the Northeast than in the South. Allport (1954) specifies the aforementioned conditions that must exist for contact to truly reduce prejudice such as individuals of different race or ethnic groups having equal status or working together toward common goals. These conditions may be more prevalent in the North. Since White Flight occurs more predominately in the South (Crowder, 2000), it may be the case that Whites are more isolated in this compared to other U.S. region. An increase in perception of Black community composition may not make a significant enough impact in already isolated areas to affect prejudiced attitudes. Or, the conditions that must exist for contact to reduce prejudice may never occur because Whites may flee with an initial increase of Black residents. The one case that did not follow the first hypothesis was for anti-Hispanic prejudice in White households in the Northeast, which was significantly higher compared to the South. This could be due to a number of factors. As Weaver (2007) points out, Hispanics tend to be grouped geographically by origin. In this particular relationship, Puerto Ricans are likely to live in the Northeast while Mexicans are likely to live in the South (Weaver, Scientia et Humanitas 22 Spring 2011 2007:269). Another factor could be contact theory. While there was no significant interaction effect with perception of Hispanic community composition by region, and that variable represented contact, we cannot ignore the difference in Hispanic population by these two regions. From 2000 to 2006, the Hispanic population in the Northeast increased by 15.1%. In comparison, the Hispanic population in the South increased by 31.2% (U.S. Census Bureau, 2006). With such high Hispanic population growth in the South, it may be the case that more contact with this population has allowed prejudiced attitudes to decline. The hypothesis that White households in the Suburbs would have higher levels of anti- minority prejudice was not supported. White respondents in White households in rural areas had significantly higher levels of anti-Asian prejudice than households in suburban areas. This could be a result of less contact with Asians in rural areas. Although this is an assumption, this finding does show that contact operates in a specific way. Education levels had a significant effect on anti-minority prejudice. Across anti-Black, anti-Hispanic, and anti-Asian prejudice, an increase in years of education attained was associated with a highly significant decrease in prejudice. This follows previous research into racial prejudice (Logan, 2002), and is also a key component in the spatial assimilation model. Residential segregation should dissipate, this model claims, as levels of education increase across all races (Charles, 2003). This finding is a positive step towards reducing racial/ethnic prejudice. However, residential segregation still exists and minorities do not have the same access to resources as Whites. As for our contact hypothesis, there were no significant findings to suggest that higher levels of perception of minority community composition had an effect on racial/ethnic prejudice. This is aside from the interaction term previously discussed. However, as the other significant variables show, contact may have some indirect effects on the racial make-up of the region and city size that these households are located. Indeed, numerous studies have indeed shown the importance of contact (Pettigrew, 2008; Weaver, 2007). CONCLUSION Despite the important findings of this research, more research into the causes and mechanisms of residential segregation is still needed. Region, education, and age had effects across all minority prejudice. As indicators of residential segregation, household type and suburban location were not significant predictors of racial/ethnic prejudice in the current study. However, we know that this is where White flight is continuing to occur. Research on residential segregation focuses on the city/suburb dichotomy, as Whites “flew” from cities into suburbs. More research should delve into the prejudiced views of rural residents, as they are also isolated. Further research must disentangle these factors, and incorporate questions on implementation of racially prejudiced policies as well as principle, or feelings of prejudice (Dixon, 2006). While implementation questions can establish a possible underlying prejudice against minority groups, this research can still be looked at as a step in exploring the relation of household characteristics and community perception to prejudiced attitudes of Whites. Furthermore, the contact variable used does not satisfy Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 23 all the requirements of the contact theory. Ideally the contact variables would represent Allport’s conditions of contact that reduces prejudice (1954). In addition, to completely link racial prejudice to household characteristics and community perception, research would need to consider locational data of respondents. The racial make-up of the areas where respondents live would provide an alternative measure and more nuanced test of contact theory. This study emphasizes the importance of region, education, and age on levels of racial/ ethnic prejudice. Higher levels of education were associated with lower levels of prejudice, while increase in age was associated with higher levels of prejudice. This can be a tool to further reduce prejudice. Also, while many studies focus on the prejudiced and racist attitudes of the South, this research shows a more complicated picture. There were differences among households in the Midwest and the South on anti-Black prejudice, and among the West and the South on anti-Hispanic and anti-Asian prejudice. However, households in the Northeast had higher levels of anti-Hispanic prejudice than households in the South. This finding alone is important as research on the Hispanic population evolves as the population rapidly expands. Further research should look into Hispanic differences by origin, and if there are any trends of White prejudice toward different groups. With such high Hispanic population growth in the South, perhaps more contact with this population has allowed prejudiced attitudes to decline. This claim needs to be investigated further, but appears to be supported by this analysis. Scientia et Humanitas 24 Spring 2011 REFERENCES Allport, Gordon W. (1954). The Nature of Prejudice. Cambridge, MA: Addison-Wesley Publishing. Blalock, H.M. (1957). Percent Non-White and Discrimination in the South. American Sociological Review, 22, 677-82. Bobo, Lawrence, James R. Kluegal, and Ryan A. Smith. (1997). Laissez-Faire Racism: The Crystallization of a Kinder, Gentler, AntiBlack Ideology. Pp. 15-42 in Racial Attitudes in the 1990s: Continuity and Change, edited by Steven A. Tuch and Jack K. Martin. Westport, CT: Prager Publishers. Blumer, Herbert. (1958). Prejudice as a Sense of Group Position. Pacific Sociological Review, 1, 3-7. Carter, J. Scott, Lala Carr Steelman, Lynn M. Mulkey, and Casey Borch. (2005). When the Rubber Meets the Road: Effects of Urban and Regional Residence on Principle and Implementation Measures of Racial Tolerance. Social Science Research, 34, 408-425. Charles, Camille Zubrinsky. (2003). The Dynamics of Racial Residential Segregation.” Annual Review of Sociology, 29, 167-207. Clark, William A.V. (1991). Residential Preferences and Neighborhood Racial Segregation: A Test of the Schelling Segregation Model. Demography, 28(1), 1-19. Clark, William A.V. and Sarah Blue. (2004). Race, Class, and Segregation Patterns in U.S. Immigrant Gateway Cities. Urban Affairs Review, 32(6), 825-843. Crisp, Richard J., & Rhiannon N. Turner. (2009). Can Imagined Interactions Produce Positive Perceptions?: Reducing Prejudice Through Simulated Social Contact. American Psychologist, 64, 231-40. Crowder, Kyle. (2000). The Racial Context of White Mobility: An Individual-Level Assessment of the White Flight Hypothesis. Social Science Research, 29, 223-257. Davis, James A. and Tom W. Smith. (1992). The NORC General Social Survey: A User’s Guide. Newbury Park, CA: Sage Publications. Dixon, Jeffrey C. (2006). The Ties that Bind and Those that Don’t: Toward Reconciling Group Threat and Contact Theories of Prejudice. Social Forces, 84(4), 2179-2204. Emerson, Michael O., George Yancey, and Karen J. Chai. (2001). Does Race Matter in Residential Segregation? Exploring thePreferences of White Americans. American Sociological Review 66, 922- 935. Feagin, Joe R. (1991). The Continuing Significance of Race: Antiblack Discrimination in Public Places. Pp. 239-245 in The Inequality Reader, edited by David B. Grusky and Szonja Szelenyi. Westview Press. Federico, Christopher M. (2006). Ideology and the Affective Structure of Whites’ Racial Perceptions. Public Opinion Quarterly, 70(3), 327-353. Freeman, Lance. (2008). Is Class Becoming a More Important Determinant of Neighborhood Attainment for African-Americans? Urban Affairs Review 44(1), 3-26. Gallup. (2008). Americans See Obama Election as Race Relations Milestone. Retrieved July 10, 2009, http://www.gallup.com/poll/111817/americans-see-obama-election-race-relations-milestone.aspx. Gans, Herbert J. (1991). People, Plans, and Policies: Essays on Poverty, Racism, and Other National Urban Problems. New York, NY: Columbia U Press. Racial/Ethnic Prejudice in the U.S. Middle Tennessee State University 25 Gans, Herbert J. (1999). The Possibility of a New Racial Hierarchy in the Twenty-First Century United States. Pp. 266-275 in The Inequality Reader, edited by David B. Grusky and Szonja Szelenyi. Westview Press. Jackson, Kenneth. (2000). Gentleman’s Agreement: Discrimination in Metropolitan America. Pp. 185-217 in Reflections on Regionalism, edited by Bruce Katz. The Brookings Institution. Retrieved October 5, 2007, http://www.brookings.edu/es/urban/reflections.htm. Johnson, Bryan R. and Cardell K. Jacobson. (2005). Contact in Context: An Examination of Social Settings on Whites’ Attitudes toward Interracial Marriage. Social Psychology Quarterly 68(4), 387-399. Krysan, Maria. (2002). Community Undesirability in Black and White: Examining Racial Residential Preferences through Community Perceptions. Social Problems 49(4), 521-543. Lake, Robert. (1981). The New Suburbanites: Race and Housing in the Suburbs. Center for Urban Policy Research, Rutgers University. Larsen, Luke J. (2004). The Foreign-Born Population in the United States: 2003. Current Population Reports, P20-551, U.S. Census Bureau, Washington, D.C. Ligon, Ethan. (1989). “Rationale and Construction of Poverty Measures in the General Social Survey.” General Social Survey Methodological Report 64, Chicago: NORC (23 pages). Logan, John R. (2002). Separate and Unequal: The Neighborhood Gap for Blacks and Hispanics in Metropolitan America. Lewis Mumford Center for Comparative Urban and Regional Research, University at Albany (20 pages). Retrieved October 5, 2007, http://mumford.albany.edu/census/ SepUneq/SUReport/SURepPage1.htm. Long, J. Scott and Jeremy Freese. (2003). Regression Models for Categorical Dependent Variables Using Stata. College Station, TX: Stata Press. Massey, Douglas S. and Nancy A. Denton. (1993). American Apartheid: Segregation and the Making of the Underclass. Cambridge,MA: Harvard University Press. McKinnon, Jesse. (2003). The Black Population in the United States: March 2002. U.S. Census Bureau, Current Population Reports, Series P20-541. Washington, D.C. Oliver, Melvin L. and Thomas M. Shapiro. (1997). Black Wealth/White Wealth: A New Perspective on Racial Inequality. Pp. 258-265 in The Inequality Reader, edited by David B. Grusky and Szonja Szelenyi. Westview Press. Patillo-McCoy, Mary. (1999). Black Picket Fences: Privilege and Peril among the Black Middle Class. Chicago: University of Chicago Press. Pettigrew, Thomas F. (2008). Review: Future Directions for Intergroup Contact Theory and Research. International Journal of Intercultural Relations, 32(3), 187-199. Pettigrew, Thomas F., & Linda R. Tropp. (2008). How Does Intergroup Contact Reduce Prejudice? Meta- analytic Test of Three Mediators. European Journal of Social Psychology 38, 922-34. Pew Hispanic Center. (2009). Statistical Portrait of Hispanics in the United States. Retrieved October 22, 2009, http://pewhispanic.org/factsheets/factsheet.php?FactsheetID=46. Scientia et Humanitas 26 Spring 2011 Quillian, Lincoln. (2002). Why is Black-White Residential Segregation so Persistent? Evidence on Three Theories from Migration Data. Social Science Research, 31, 197-229. Ramirez, Roberto R., & G. Patricia de la Cruz. (2002). The Hispanic Population in the United States: March 2002, Current Population Reports, P20-545, U.S. Census Bureau, Washington, D.C. Reeves, Terrance, & Claudette Bennett. (2003). The Asian and Pacific Islander Population in the United States, March 2002. Current Population Reports, P20-540, U.S. Census Bureau, Washington, D.C. Smith, Tom W. (1978). Size of Place Codes on the 1972-1977 General Social Surveys. General Social Survey Methodological ReportNo. 4, Chicago: NORC (30 pages). Tabachnick, Barbara G. and Linda S. Fidell. (2007). Using Multivariate Statistics 5th edition. Boston, MA: Pearson Education Inc. U.S. Census Bureau. (2006). Hispanics in the United States. Ethnicity and Ancestry Branch, Population Division (39 pages). Retrieved October 2, 2009, http://www.census.gov/population/www/socdemo/ hispanic/files/Internet_Hispanic_in_US_2006.pdf. Weaver, Charles N. (2007). The Effects of Contact on the Prejudice Between Hispanics and Non-Hispanic Whites in the United States. Hispanic Journal of Behavioral Sciences, 29(2), 254-274. Williams, David R. and Ruth Williams-Morris. (2000). Racism and Mental Health: The African American Experience. Ethnicity and Health, 5(3-4), 243-268. Wilson, William Julius. (2008). The Political and Economic Forces Shaping Concentrated Poverty. Political Science Quarterly, 123(4), 555-571. Zubrinsky, Camille L. and Lawrence Bobo. (1996). Prismatic Metropolis: Race and Residential Segregation in the City of Angels.Social Science Research, 25, 335-374