As the size of the Black middle class has grown in the post-civil-rights era, scholars have debated whether improved living standards and conditions of political equality have caused racial consciousness to be supplanted by class consciousness among African Americans. Wilson (1980) argued that the salience of race among African Americans would diminish with the decline of racial discrimination, on the assumption that as the life prospects of African Americans became more dependent on their economic status, their attitudes would be guided less by racial considerations and more by social class concerns.
Dawson (1994), however, found that identification with racial group interests was not weaker among higher status African Americans. On the contrary, income and education were positively correlated with the feeling among African Americans that they shared a common fate with other Blacks. According to Dawson and other scholars (Cose 1995; Hochschild 1993, 1995), racial consciousness persists because racial discrimination remains sufficiently pervasive that individuals believe their personal advancement is tied to improvements in the status of the entire group. Middle-class African Americans continue to focus on economic disparities between Blacks and Whites and to equate their self-interest with the interests of the group despite their personal achievements.
The socioeconomic barriers created by racial prejudice and discrimination have led scholars to question whether the most recent immigrants to the United States from Latin America and Asia will follow the classic pattern of assimilation exhibited by earlier generations of European immigrants in which ethnic identities faded as individuals were structurally assimilated in American society (e.g., Alba and Nee 1997; Portes and Rumbaut 1996; Rumbaut 1997). Socioeconomic mobility creates opportunities for “equal status contacts across ethnic lines in workplaces and neighborhoods” (Alba and Nee, 831), but minorities may in fact discover that their economic achievements do not erase racial and ethnic boundaries that limit their opportunities. In the housing market, for example, middle-class minorities often find they do not have the same access to the more desirable suburbs that are available to nonminorities with similar economic resources (Massey and Denton 1988, 1993). If Asian Americans and Latinos face more formidable barriers to social acceptance by the majority population than did previous European immigrants, will they react to discrimination and impediments to mobility in the same way as African Americans?
GROUP CONSCIOUSNESS AMONG AFRICAN AMERICANS, LATINOS, AND ASIAN AMERICANS
Most of what we know about the sources and effects of racial and ethnic group consciousness in U.S. politics derives from studies of African Americans (e.g., Chong and Rogers 2005; Dawson 1994; Gurin, Hatchett, and Jackson 1989). Only recently with the collection of representative survey data on Latinos and Asian Americans have researchers begun to explore whether racial and ethnic consciousness affects the political attitudes and behavior of other minority groups in the same degree as consciousness affects African Americans. This work largely follows the approach taken in research on African Americans by examining perceptions of shared interests among individuals and the potential for group consciousness to foster political participation (e.g., de la Garza et al. 1992; Jones-Correa and Leal 1996; Lien, Conway, and Wong 2004).
These studies have just begun to examine whether Latinos and Asian Americans will place the same priority as African Americans on racial and ethnic group interests over economic considerations in their political attitudes and policy preferences. To explain similarities and differences in the political attitudes of African Americans, Latinos, and Asian Americans, we need a theory flexible enough to accommodate variations among minority groups in their socioeconomic status and their evaluations of opportunities and conditions in American society. It is a reasonable first approximation to place these three groups into the general category of minorities and to contrast their structural position and subjective states to those of Whites. But within the common outlook of minority populations are significant differences that are likely to have a bearing on their political attitudes and behavior.
An important source of variation among African Americans, Latinos, and Asian Americans that may influence the salience of race and ethnicity among them is their contrasting levels of socioeconomic success. Compared to African Americans, Asian Americans have achieved greater levels of economic success, although the group averages calculated on the basis of this broad racial category masks significant variation among Asian Americans. In terms of education, income, and home ownership, Asian Americans appear to be rapidly integrating into American society and making the kind of progress that befits their disputed title as a “model minority” (Alba and Nee 2003; Ong 2000). Asian Americans also report experiencing discrimination significantly less frequently than do African Americans. They are more optimistic than other minorities about their prospects and believe they are doing well according to most conventional measures of success (Kaiser Family Foundation 1995).
Latinos face markedly different contemporary social and economic conditions compared to Asian Americans and African Americans (de la Garza 2004; Hero 1992; Jones-Correa 1998). In median income, housing, and education, Latinos have a status that is similar to, if not worse than, African Americans. Latinos themselves believe their socioeconomic status compares unfavorably to that of Whites, but at the same time many Latinos came to this country because they believe the United States offers more economic opportunities than their mother countries.1
See the survey toplines from the Washington Post/Kaiser Family Foundation/Harvard University 1999 National Survey on Latinos in America, available at http://www.kff.org.
Such differences in the economic circumstances of African Americans, Latinos, and Asian Americans, and in their interpretations of group conditions and economic opportunities, should affect the centrality of racial and ethnic group interests in their evaluations of political issues. In particular, we should not assume the political attitudes of Asian Americans and Latinos will be as impervious to changes in economic status as those of African Americans given their disparate assessments of prospects for success and varying experiences with discrimination.
With these premises as our starting point, we examine in this paper the effects of economic status, opportunities, and perceptions of discrimination on the propensity of African Americans, Latinos, and Asian Americans to pursue racial and ethnic group interests. We use the variation among minorities on these dimensions to develop a theory of opportunities and group consciousness that explains how and when economic status will affect support for group interests.
From this theory, we derive two hypotheses relating support for group interests to the economic status and socioeconomic experiences and perceptions of individuals. We first examine variation among African Americans, Latinos, and Asian Americans in the extent to which economic status influences support for racial and ethnic group interests. We then test conditions when higher economic status is more likely to diminish support for group interests among members of all three groups. In our conclusion, we discuss the implications of these results for political assimilation and minority group politics.2
We maintain a distinction between perceptions of a group interest and support for policies that promote group interests. We refer to these perceptions as “racial group consciousness” (in the case of African Americans and Asian Americans) and “ethnic group consciousness” (in the case of Latinos), and to the policy preferences as “support for racial and ethnic group interests.” The conceptual distinction is blurred because the two concepts are closely related and significantly correlated empirically. Some researchers (e.g., Verba, Schlozman, and Brady, 1995) have ignored the distinction between identification with group interests and support for them by including indicators of both in a measure of racial consciousness. The theory we develop in this paper is meant to account for variation in both group consciousness and support for group interests, and in discussing the dynamics of that theory we will often refer interchangeably to both perceptions and support. Our empirical tests of the theory focus on support for group interests in public policy.
A THEORY OF OPPORTUNITIES AND GROUP CONSCIOUSNESS
The persistence of racial consciousness among African Americans in reaction to unequal opportunities highlights a general relationship between group consciousness and discrimination that has been replayed through the course of American history. The assimilation of a minority group into American society depends not only on the actions of group members but also on the reception accorded that group by the majority population. An impediment to the assimilation of immigrants in different eras has been the reluctance of the majority to allow new groups equal access to the mainstream social and political institutions of society (Higham 1992). Whether the majority opens its doors to minorities—in its neighborhoods, schools, corporations and businesses, private clubs, and political institutions—affects the propensity of minorities to relinquish their identification with a racial or ethnic subculture (Gordon 1964). Immigrants who enjoy more extensive social and economic opportunities are more likely to assimilate to the mainstream and reduce their attachment to racial or ethnic groups as a means to improve their own life chances and those of their children (Alba and Nee 2003; Dahl 1961).3
An important qualification of this dynamic is the idea of segmented assimilation proposed by Portes and Zhou (1993), which raises the possibility that upward mobility does not weaken group values among those immigrant groups that possess strong social networks based on race or ethnicity.
These social psychological dynamics conform readily to the general theory of social identity developed by Henri Tajfel and his associates (Hogg and Abrams 1988; Tajfel 1981; Tajfel and Turner 1979). According to this theory, dissatisfaction with one's social status can motivate efforts to improve one's position through either individual or collective means. The choice between these strategies depends on whether one believes society provides individuals with an equal opportunity to advance. Individuals who believe that social mobility is possible are less likely to identify with their current group because they feel the boundaries between groups are permeable and that hard work, education, and other investments will allow them to move to a higher status group. A belief in the possibility of social mobility therefore tends to be associated with individualist strategies that work within the status quo.
On the other hand if socioeconomic boundaries are less permeable and opportunities for mobility are blocked, individuals will become more conscious of their group memberships, especially if those group affiliations are responsible for restricting their mobility. One avenue for individuals to improve their personal status is by working with fellow group members to advance the group's status, because their individual interests are tied to the fortunes of the group as a whole.
This basic dynamic between mobility and group identification lies at the center of our theory of opportunities. We expect the effect of economic status on support for group interests will depend significantly on the degree to which individuals perceive opportunities for social mobility. Whether middle-class minorities experience equal opportunity or encounter prejudice and discrimination in their daily lives will affect the extent to which they remain conscious of their minority status and support policies that address racial and ethnic inequality. Variation across minority groups in the effect of economic status on racial and ethnic consciousness and support for group interests can therefore be explained by systematic differences in the perceptions and experiences of group members rather than by contrasting psychological processes.
The theory of opportunities suggests two hypotheses about the relationship between economic status and group consciousness among African Americans, Latinos, and Asian Americans. Our first hypothesis is that African Americans' support for group interests, relative to that of Latinos and Asian Americans, will be the least responsive to changes in economic circumstances because African Americans face more daunting economic and social barriers than other racial and ethnic minorities. When African Americans achieve higher economic status, they continue to experience discrimination and to evaluate their life prospects in racial terms.
Our second hypothesis tests whether the same factors that explain intergroup differences can also explain intragroup variation. As Gay (2004) recently demonstrated, greater satisfaction among African Americans with the quality of their residential neighborhoods tends to weaken their belief that they share a common fate with other African Americans. We examine more generally whether the effect of economic status on support for group interests is conditioned by individual experiences of status among all minorities. If prejudice and discrimination reinforce the tendency of minorities to think in group terms, then the effect of improvements in economic status on group consciousness will depend on the socioeconomic experiences of individuals. Individuals who experience social and economic integration alongside their higher economic status should attribute less importance to race or ethnicity as a determinant of life circumstances. Economic security in such cases should reduce support for group interests.
On the other hand, individuals who achieve nominal material success but continue to feel excluded on racial or ethnic grounds from the economic and social institutions that normally accompany higher status should be more likely to retain their group consciousness. They will be more inclined to view politics in racial or ethnic terms and to support policies designed to remove structural and institutional barriers to equality. This specification of the circumstances in which economic status will moderate group consciousness should account for differences across racial and ethnic groups as well as within them.
There are of course other factors that influence development of group consciousness beyond the socioeconomic elements examined here, but they fall outside the scope of the data analyzed in this paper. A comprehensive model also would take account of various agents of socialization, patterns of social interaction, and the racial composition and quality of neighborhoods—among other life experiences and contextual factors—to explain similarities and differences in consciousness across racial and ethnic groups (e.g., Demo and Hughes 1990; Gay 2004; Welch et al. 2001). We develop a central component of this comprehensive model with our theory of opportunities, by offering general propositions about the effect of economic status and socioeconomic experiences on racial and ethnic consciousness that are applicable across minority groups.
DATA
Our analysis uses data gathered in a national survey conducted jointly by the Washington Post, the Kaiser Family Foundation, and Harvard University (Post/Kaiser/Harvard) in 2001. This survey allows us to study how minority groups regard their own position in society, and how their perceptions of social conditions combine with their material life circumstances and experiences to influence their support for racial and ethnic group interests. Unlike typical national surveys, this survey included an over-sample of 696 minority respondents (230 African Americans, 237 Latinos, and 229 Asian Americans) in addition to a regular national random sample of 1,013 respondents. This resulted in a nationally representative sample of 1,709 randomly selected respondents aged 18 and older consisting of 779 whites, 323 African Americans, 315 Latinos, and 254 Asian Americans. Sampling bias was reduced by conducting interviews with Latinos in either English or Spanish, and with Asian Americans in English, Korean, Cantonese, Mandarin, Vietnamese, or Japanese depending on the respondent's preference. The descriptive statistics and model estimations reported here are weighted to reflect the actual racial and ethnic distribution in the nation.4
The margin of sampling error for the whole sample is ±3%; for non-Hispanic Whites—±4%; for African Americans—±6%; for Latinos—±7%; and for Asian Americans—±9%.
TESTING THE THEORY OF OPPORTUNITIES
The Post/Kaiser/Harvard survey contains extensive data on the material quality of life of racial and ethnic minorities, and their assessments of opportunities for advancement in American society. In addition, the survey contains self-report data on the amount of prejudice and discrimination experienced by respondents in everyday life. These data allow us to test how variations in economic status and socioeconomic experiences and perceptions affect support for group interests among African Americans, Latinos, and Asian Americans.
We estimate two statistical models corresponding to our two hypotheses. Our first model examines the relative influence of economic status, perceptions of life opportunities, and direct experiences with discrimination on support for racial and ethnic group interests:
The second statistical model examines the effect of economic status conditional on perceptions of opportunities and experiences of discrimination:
We test in the second model, using interaction terms, whether improvements in economic circumstances are more likely to weaken support for group interests among African Americans, Latinos, and Asian Americans who experience equal treatment in society than among individuals who continue to face discrimination. The key comparison for our purposes is between those individuals who have had positive experiences (i.e., those who perceive equal opportunity and encounter little discrimination) and those who have had negative experiences (i.e., those who perceive unequal opportunities and encounter frequent discrimination). Higher economic status should significantly reduce support for group interests among those in the former category, but have a negligible effect among those in the latter. We will pay less attention to respondents in the hybrid mixed category (i.e., those who either perceive equal opportunity or encounter little discrimination) because our theory does not offer a definite prediction about how these individuals will respond to their conflicting perceptions and experiences.
OPERATIONALIZATIONS5 The exact wording of all items is provided in the Appendix along with scale reliabilities where applicable.
The exact wording of all items is provided in the Appendix along with scale reliabilities where applicable.
Racial and Ethnic Group Interests
In previous research, racial and ethnic consciousness have been most commonly measured with items probing feelings of shared interests and closeness with other individuals of the same race or ethnicity. The Post/Kaiser/Harvard survey does not contain these traditional consciousness items, but it contains several questions assessing whether racial or ethnic group interests are central to how one thinks about politics. We use three of these indicators of support for group interests on political issues as our dependent variables. Our first measure is a question asking respondents whether more attention or less attention should be paid to racial issues in this country (or if just the right amount of attention is currently being paid to race). Our two other measures of group interests are based on preferences toward affirmative action and government policies to ensure equality between minorities and Whites in health care, education, jobs, and the administration of the law. The affirmative action question asks whether special efforts should be made by colleges and businesses to recruit qualified minorities. The four questions about government's role in ensuring equality, which we aggregated into a scale, present a tradeoff between group interests and individual economic costs by asking respondents if they support government action in each policy domain even if it leads to higher taxes.
Each of these measures probes the extent to which one thinks about public affairs along racial or ethnic lines and gives priority to group benefits when evaluating public policy. In recognition of the difference between these items and conventional measures of racial consciousness, we will generally refer to our dependent variables as measures of support for racial and ethnic group interests.6
The three indicators of group interest are significantly correlated with one another and together form a scale with a reliability coefficient of .57. In the analysis to follow, we disaggregate the scale and examine how individuals respond to its separate components, because there is some variation across the measures, but we also report our findings when we use the scale as our dependent variable in the same models.
This assumption is substantiated in Dawson's (1994) analysis of the predictors of support among African Americans for government racial policies similar to those used in our analysis, including affirmative action, aid to Blacks, programs to improve the quality of education for Blacks, and government efforts to improve the economic position of Blacks. Dawson found that perceptions of a “linked fate” with other African Americans “played the greatest role in predicting support for government racial policies. … On the whole, the expectation that the perception of linkage between individual and group interest would dominate in policy areas where race is clearly salient is supported by this analysis” (194).7
We returned to the 1984–88 NBES Panel Study and built a scale using the four racial policy attitude items described in the text to calculate the correlation between this scale and the measure of linked fate. The Pearson's r between the two measures is .21 and is statistically significant at the .001 level. Because the policy items in the NBES are similar in substance to the Post/Kaiser/Harvard items, we interpret this result as further evidence that the racial policy attitudes reflect racial consciousness.
The marginal distributions of opinion on our three indicators of group interests confirm that African Americans place more emphasis on racial considerations than do Latinos and Asian Americans. Sixty-four percent of African Americans believe that too little attention is paid to race, compared to 41% of Latinos and 33% of Asian Americans. Similarly, 77% of African Americans endorse extra efforts by employers and colleges to recruit qualified minorities versus 63% of Asian Americans and 62% of Latinos.
African Americans are also more likely to believe the federal government is responsible for ensuring equality in the areas of employment, education, health care, and law. Seventy two percent of African Americans believe the government is responsible for ensuring that minorities have jobs of equal quality as Whites. Identical proportions (89%) of African-Americans believe the government should: make sure that schools are of equal quality; guarantee that minorities and Whites receive equal health care services; and ensure that minorities and Whites are treated equally by the courts and police. Majorities of Latinos and Asian Americans also feel the federal government is responsible for enforcing equality in these areas, but there is greater division of opinion within these two groups. The average level of support for government action across the four domains is 77% among Latinos and 71% among Asian Americans, compared to 85% among African Americans.
Economic Status
Economic status is a composite measure that combines annual family income and ownership of stocks and bonds with subjective assessments of financial well-being and reports of past financial problems with rent and mortgage payments, medical care, and saving money for the future.8
The Cronbach's alpha measure of reliability for this scale is .72. The scale reliability declines to .68 if we omit the item measuring respondents' subjective perceptions of their personal economic situations. Subjective and objective indicators of well-being prove to be closely related. The simple correlation between subjective perceptions and the economic status scale constructed without this item is .51.
According to the 2000 U.S. Census, the median incomes of African American, Latino, and Asian families were $33,255, $34,394, and $59,394, respectively. By comparison, the median income of White (non-Hispanic) families was $53,356.
When asked to describe their personal financial situations, 64% of Asian Americans, and 53% of Latinos report that their financial state is either excellent or good, but a majority (51%) of African Americans say their financial situation is either poor or “not so good.” African Americans and Latinos are also more likely than Asian Americans to report they have experienced difficulty paying mortgages or rents, getting medical care, or saving money for future needs.10
In the analysis to follow, we sometimes refer to the effect of “improvements” or “changes” in economic status, even though we are analyzing cross-sectional data. This phrasing is in keeping with the conventional interpretation of a regression coefficient as the change in the dependent variable produced by a unit change in the independent variable. Panel data would be needed to assess the effect of changes in economic status in the same individuals.
Perception of Opportunities
Individuals were asked to evaluate their life opportunities relative to those enjoyed by Whites. African Americans offer the most negative evaluations of their own group's chances in society. Seventy-four percent of African Americans feel they have less opportunity than Whites, compared to 45% of Latinos and 34% of Asian Americans. Only 24% of African Americans think they have as many opportunities as Whites. Asian Americans are more positive about their own group's opportunities than are Latinos. Sixty percent of Asian Americans feel they have either the same or better opportunities than Whites, whereas 53% of Latinos believe they have at least as much opportunity as Whites.
Personal Experience of Discrimination
Respondents were asked a series of seven questions about whether they had experienced racial or ethnic prejudice and discrimination in the workplace and in their daily lives. Answers to these questions were summed and converted to a scale measure. As in the case of economic status and perceived opportunities, there are systematic differences among minority groups, with African Americans suffering the indignities of racial discrimination to a significantly greater extent than other minorities. Latinos and Asian Americans also report being subjected to intolerance, but they are not exposed to prejudice as extensively and routinely as African Americans.
One out of every two African Americans report personal experiences with discrimination during the past 10 years, compared to four in 10 Latinos and Asian Americans. When asked about specific forms of discrimination, more than three out of every 10 African Americans report having been unfairly stopped by the police. One in five Latinos and one in 10 Asian Americans also report they have been subjected to racial profiling by the police. African Americans feel the slights and insults of prejudice and discrimination on a regular basis. A majority of African Americans report they have at least occasionally been treated with disrespect, received poor service, or encountered people who acted fearfully toward them because of their race. By contrast, a majority of Latinos and Asian Americans say they have never experienced discrimination in these ways.
Equations (1) and (2) also include statistical controls for education and ideology. We control for education—a component of socioeconomic status—to isolate the effects of economic status that are captured by our measure. We control for the respondent's political ideology to test whether the effects of economic status hold across all ideological groups.11
We also estimated models in which we included additional controls for age and gender on the assumption that younger respondents and men may be more likely to experience discrimination. For Asian Americans, neither age nor gender is significantly correlated with reports of discrimination; but for African Americans and Latinos both variables are moderately correlated in the expected direction (Pearson's r's range between .13 and .18). However, in the multivariate models predicting African Americans' and Latinos' racial attitudes, age and gender have a weak and inconsistent effect. Among African Americans, women are more likely than are men to support government action to address racial inequality. Among Latinos, younger individuals and women are more likely to call for greater attention to race. Otherwise, age and gender are statistically insignificant factors. With respect to the two main hypotheses tested in this paper, controlling for age and gender does not affect our findings and conclusions about the interaction between economic status and perceptions of opportunities and discrimination. Given the inconsistent and usually insignificant effects of age and gender, we favored the more parsimonious model reported here that excludes these variables. The results for the alternative model estimations can be obtained from the authors upon request.
In this survey, 22% of African Americans identify themselves as liberal, whereas 16% and 29% of Latinos and Asian Americans, respectively, say they are liberal. This marginal distribution differs from a 1995 Washington Post survey that was a precursor to the 2001 Post/Kaiser/Harvard survey. In the 1995 survey, the percentages of African Americans, Latinos, and Asian Americans who claimed liberal identification were 38%, 29%, and 25%, respectively. We do not have an explanation for this shift during the 6 years between surveys, although a potential source of variation may be the method used to over-sample minority respondents.
In estimating equations (1) and (2), we assume a nonreciprocal causal structure. This is consistent with Gay's (2004) analysis of the effect of perceived neighborhood quality on racial consciousness among African Americans, but different from Dawson's (1994) analysis of the effect of perceived group status on African Americans' racial consciousness. Gay found that perceptions of neighborhood quality influenced feelings of linked fate, but not the reverse. Dawson, however, found a statistically significant reciprocal relationship between perceptions of racial group status and feelings of linked fate, although the effect of perceived status on feelings of linked fate was considerably stronger than the reverse effect.
We tested the possibility, using a form of Hausman specification test, of a reciprocal relationship between support for racial and ethnic group interests and perceptions of economic opportunities and reported discrimination (Hausman 1978; see Gujarati 1995 and Pindyck and Rubinfeld 1998 for variations on the test). This test of simultaneity was conducted using our three measures of group interests and was repeated separately for each minority group. In each case, we failed to reject, using an F test at either the .05 or .10 levels of statistical significance, the null hypothesis that the coefficients of the fitted values for perceived opportunities and discrimination were equal to zero. Given these results, we treated our measures of perceptions of opportunities and reported discrimination as exogenous variables.13
The original Hausman specification test evaluates the null hypothesis that there is no significant difference between estimates obtained from OLS and 2SLS. If the null hypothesis is rejected, then 2SLS estimation is necessary to obtain consistent and efficient estimators. If we fail to reject the null hypothesis, then OLS estimation can be used. Spencer and Berk (1981) and Davidson and MacKinnon (1993) derived a simpler version of this test, which is based on estimation of an augmented regression model that includes the original variables plus fitted values for any suspected endogenous variables. The mathematical derivation and proof of the compatibility of the Hausman test and this version of the test can be found in Spencer and Berk, and Davidson and MacKinnon (237–42). If the coefficients for the fitted values in the augmented model are not statistically significant, the original model can be estimated with OLS. The mechanics of the estimation can be found in Gujarati (1995, 669–73) and Pindyck and Rubinfeld (1998, 353–55). We obtained predicted values for perceptions of opportunity by regressing individuals' perceptions of their own group's relative opportunities on their perceptions of the opportunities of other minority groups and the following list of variables: economic status, ideology, residence in an integrated neighborhood, gender, age, number of phones at home, economic prospects for children, perceived discrimination against multiracial couples, attitude toward marriage with African Americans, and number of children in the house. Similarly, we regressed individuals' reported experiences with discrimination on their perception of discrimination faced by members of other minority groups and the same list of exogenous variables used to obtain the predicted values for perceived opportunities.
THE EFFECT OF ECONOMIC STATUS ON SUPPORT FOR GROUP INTERESTS
The Post/Kaiser/Harvard survey reveals a consistent pattern of Asian Americans' enjoying more economic security than Latinos and African Americans as well as having a more positive outlook on the opportunities and status of their group in American society. African Americans experience the most discrimination and have the bleakest assessments of their opportunities relative to those of Whites. Given the sharp contrast in how African Americans, Latinos, and Asian Americans perceive their opportunities and standing in society, we expect a corresponding variation among these groups in the degree to which economic status will influence support for racial and ethnic group interests.
Our first statistical model therefore tests the hypothesis that economic status will have the least effect on the racial attitudes of African Americans compared to other minorities. We estimated equation (1) separately for each of the three measures of group interest, and for each of the three minority groups.
Attention to Race
The results in Table 1 generally support our hypothesis that economic status has a weaker impact on the attitudes of African Americans than on the attitudes of Latinos and Asian Americans. Economic status has a statistically insignificant effect on the emphasis that African Americans place on race. Instead, their attitudes are highly dependent on their perceptions of group opportunities; individuals who believe they have fewer opportunities in life than Whites want to increase the salience of race in public affairs. Political ideology, which has a significant impact among both Latinos and Asian Americans, has no effect on the degree to which African Americans call for attention to racial issues.
In contrast, economic status exerts a significant influence on the degree to which Asian Americans focus on race. Asian Americans who have a higher economic status are significantly more likely to downplay racial issues. Individuals who have a favorable assessment of Asian Americans' group opportunities also place less emphasis on race. But Asian Americans who have frequently experienced discrimination call for greater attention to be given to race.
Higher economic status also significantly diminishes the salience of race among Latinos, but the effect is somewhat weaker among Latinos than it is for Asian Americans. Latinos with a positive assessment of group opportunities assign a lower priority to racial issues. Higher education reduces attention to race only among Latinos. But conservative ideology in both Latinos and Asian Americans directs attention away from racial issues.
Affirmative Action
There is also no significant variation by economic status among African Americans in support for recruiting qualified minorities in education and employment. However, African Americans who have encountered discrimination or have a more negative assessment of the opportunities available to them are substantially more likely to support affirmative action.
Similarly, among Latinos, those who have faced discrimination are more inclined to support affirmative action, as are those who believe that opportunities are unequal. But contrary to our expectations, higher economic status elevates support for affirmative action. We cannot account for this result, but suspect it is an anomaly because it is inconsistent with other national survey data we analyzed on Latino attitudes.14
In the absence of comparable studies that examine the sources of Latinos' support for affirmative action, we analyzed the influence of economic status using the 1989–90 Latino National Political Survey. The simple correlation between income and “support for quotas in employment and college education” is −.15 (p < .01). In addition, a multivariate test of the effect of income, controlling for age, education, gender, ideology, and experiences with discrimination in daily life, yields a similar result. Those with higher income are significantly less likely (p < .01) to support quotas.
Economic status has the expected negative effect among Asian Americans; affluent Asian Americans are significantly more opposed to affirmative action than Asian Americans of lower economic status. But Asian Americans are no different from Latinos and African Americans in being more likely to support affirmative action if they perceive unequal opportunities. Asian Americans who identify themselves as political liberals place greater emphasis on recruiting minorities, whereas political ideology has no significant effect on African Americans and Latinos. A puzzle in Table 1 is that better educated African Americans and Asian Americans are more supportive of affirmative action but better educated Latinos are significantly less supportive. The negative relationship between education and support for affirmative action among Latinos is confirmed by other data we analyzed, but we do not have an explanation for the varied effect of education across racial and ethnic groups.15
The correlation between education and support for racial quotas in the 1989–90 Latino National Political Survey is −.14 (p < .01); the negative relationship remains statistically significant (p < .01) in a multivariate analysis controlling for age, income, gender, ideology, and experiences with discrimination in daily life.
Government Programs to Ensure Equality
The findings in Table 1 on support for government action in health care, education, jobs, and the administration of the law reveal the expected divergence between African Americans on the one hand and Latinos and Asian Americans on the other. For African Americans, higher economic status does not reduce support for government efforts to address racial and ethnic inequality. In contrast, higher economic status significantly reduces support for government action among both Latinos and Asian Americans. Perception of group opportunities has a significant effect in the expected direction on the policy preferences of individuals in all three groups. Among Latinos, education and conservatism increase opposition to government action. There is also a modest statistically significant relationship between conservatism and opposition to these measures among African Americans.16
The goodness of fit statistics for these models are characteristically low in the case of explaining African Americans' racial attitudes. The marginal distributions of their attitudes on the attention to race and government action items are highly skewed in the direction of support. Because there is little variance to explain in the models of African American attitudes, the R2 statistics for the estimations are low. Even if we control for additional variables (e.g., age, gender, and living in an integrated neighborhood) our R2 statistics only marginally improve to .07 and .04 for the “attention to race” and “government action” equations. These results are in line with previous studies of African Americans' attitudes toward racial policies. For example, Dawson's (1994, 195) model of the determinants of racial policy attitudes shows a similar low R2 of .07 despite using a 4-item policy index for his dependent variable.
The statistical estimates presented in Table 1 therefore confirm that support for group interests tends to be more strongly related to economic status among Latinos and Asian Americans than among African Americans even after controlling for education, ideology, perceptions of opportunities, and experiences with discrimination. As African Americans experience greater economic success, the salience of race for them tends not to decline. On the other hand, Latinos and especially Asian Americans are more likely to extrapolate from their individual success in judging race-based policies and the attention that should be given to race.
Although we presented the results separately for each issue and minority group, we also conducted an explicit test on the total sample of the interaction between economic status and minority group, using the group interests scale as the dependent variable. This test shows that economic status is unconnected to support for group interests among African Americans, but that a higher economic status significantly reduces support for group interests among Latinos and Asian Americans. A statistical test of the interaction between economic status and each minority group confirms that the effect of economic status is significantly different among Latinos and Asian Americans, relative to African Americans, at the .10 and .05 levels, respectively.
THE VARIED EXPERIENCES OF ECONOMIC STATUS
It appears from these results that support for racial and ethnic group interests is least responsive among African Americans to changes in personal economic status because of the broader consensus among African Americans that group opportunities and social conditions remain poor despite individual examples of success. This explanation can be tested more directly by tracing the experiences of higher status minorities that are relevant to their orientations toward political issues. A simple way to peer into the different experiences that accompany higher economic status among African Americans, Latinos, and Asian Americans is to explore the structure of correlations among economic status, reports of personal encounters with discrimination, and assessments of the opportunities enjoyed by group members. The correlation matrix in Table 2 for these variables illustrates the disparate experiences of higher status members of the three minority groups. As economic status increases among African Americans, there is no accompanying tendency for them to acquire a more positive outlook on American society. Higher status changes neither their appraisal of group opportunities nor the frequency of their personal encounters with discrimination.
Latinos and Asian Americans differ from African Americans in notable ways. Affluent Latinos provide a more positive assessment of group opportunities; they also report fewer personal experiences with discrimination. A similar pattern of correlations describes the experience of economic status among Asian Americans. A higher standard of living among Asian Americans is unrelated to perceptions of opportunities, but is accompanied by a significant reduction in reports of personal discrimination. Therefore improvements in living standards go hand in hand with more favorable assessments of social conditions among Latinos and Asian Americans, but they do not alter the generally negative outlook on society among African Americans.
SPECIFYING THE EFFECTS OF ECONOMIC STATUS
The effect of economic status varies among minority groups because the experiences accompanying that status are generally different depending on whether one is African American, Latino, or Asian American. African Americans place greater emphasis on race than do Latinos and Asian Americans, because they are on the whole more pessimistic about their socioeconomic prospects in American society and more likely to experience prejudice and discrimination. It should follow generally that individuals across minority groups who have not had the same bracing encounters with discrimination and who believe that minorities have a fair chance to better themselves should be more willing to downplay racial or ethnic group concerns as their own economic status improves.
In this section we test our second hypothesis that the effect of economic status depends on individual perceptions of opportunity and reports of discrimination.17
See the Appendix for details on the construction of the composite measure of perceptions of opportunities and reports of discrimination.
This hypothesis is once again tested on all minority groups for each of the three measures of group interest. The identical regression model (equation 2) is specified for each group. There are three coefficients corresponding to the effect of economic status: (1) the baseline coefficient estimates the effect of economic status among those who believe they face unequal opportunities and report high levels of discrimination; (2) the first multiplicative term (“mixed” experiences *economic status) reflects the incremental effect of economic status among those who either perceive equal opportunities or experience low levels of discrimination (but not both); and (3) the second multiplicative term (“positive” experiences * economic status) captures the incremental effect of economic status among those who both perceive equal opportunities and report experiencing low levels of discrimination.18
The unweighted percentage and number of each minority group in the “negative,” “mixed,” and “positive” categories of the combined index of opportunities and discrimination is as follows: among African Americans, 61% (193) are negative, 29% (92) are mixed, and 10% (33) are positive; the comparable figures for Latinos are 28% (88), 38% (117), and 34% (104); for Asian Americans, the comparable figures are 30% (74), 41% (102), and 29% (70).
The results of this test provide substantial support for the hypothesis that the effect of economic status is conditional on perceptions and experiences of social and economic equality. In most cases, as predicted, the baseline effect of economic status on racial attitudes among those with “negative” experiences is negligible; higher economic status does not diminish support for group interests among those who continue to feel social and economic discrimination. However, the interaction term combining economic status and “positive” perceptions of opportunities and discrimination tends to be negative and statistically significant.
Table 3 contains the statistical estimates for each of the three measures of group interest, calculated separately for each minority group. In addition, we provide a graphical summary of the key coefficients in Figure 1.19
In estimating equation (2), we centered the continuous “economic status” variable by subtracting its mean from each observation to create a new variable with a mean of zero. Centering the variable in this manner reduces multicollinearity between economic status and the measure of perceived opportunity and reported discrimination. See Aiken and West (1991) and Friedrich (1982) for discussions of the benefits of centering in models containing interaction terms.
Predicted Support for Racial and Ethnic Group Interests by Economic Status
Attention to Race
Consider the first item regarding the emphasis that respondents place on race in public affairs. The pattern of responses to this item across minority groups provides a remarkably good fit to the theory. In all three groups, the effect of economic status on the amount of attention that respondents feel should be devoted to race is negligible among those who say they are treated unequally. But among those who report more favorable experiences in society, the effect of economic status is consistently strong and statistically significant. As economic status increases in the latter group, there is a sharp decrease in the tendency to call for more attention to be paid to race. Moreover the size of the coefficient is comparable across all three groups. This is indicated in Figure 1 (panels a–c) by the similarly angled, downward sloping line in all three groups among those who provide a positive assessment of their opportunities and treatment by others.
The significant interaction between economic status and evaluations of opportunity and discrimination also means that the effects of opportunity and discrimination increase in magnitude as economic status increases. There is little difference among those with an extremely low standard of living between individuals who perceive unequal conditions and individuals who perceive equal conditions. The tendency of practically all such individuals is to call for greater attention to race. But with the exception of this small group of respondents, reports of equal opportunity and treatment lead people to downplay racial considerations in public policy.
Affirmative Action
Similar findings emerge on the issue of recruiting minorities in education and employment. The baseline effect of economic status among those with negative evaluations of society is negligible among African Americans and Latinos. The interaction terms for those with positive experiences are also in the expected negative direction for these two groups, although it is statistically significant only among African Americans. Therefore, increased economic status significantly reduces support for affirmative action among African Americans who perceive equal opportunity and report few personal encounters with discrimination. The lack of confirmation among Latinos on this issue is consistent with the positive relationship between economic status and support for affirmative action presented earlier in Table 1.
On the other hand, higher economic status tends to reduce support for affirmative action among Asian Americans, regardless of their perceptions of opportunities and experiences with discrimination. However, the negative effect of higher status is larger among those who have a positive appraisal of opportunities and discrimination. Figure 1(f) shows that as economic status increases among Asian Americans who report equal opportunity and treatment, support for affirmative action drops precipitously from overwhelming support to overwhelming opposition.
Government Programs to Ensure Equality
A similar specification by perceptions and experiences is observed among African Americans, Latinos and Asian Americans on the index of questions about the responsibility of the federal government to ensure that minorities enjoy equality in their jobs, health care, schools, and treatment by the police and courts. In all three groups, among those who offer a negative assessment of opportunities and discrimination, higher economic status does not affect support for federal action to guarantee equality in these realms. But among those who provide a positive evaluation of opportunities and levels of discrimination, affluence is accompanied by greater reluctance to pay higher taxes for government programs to ensure equality.
Nevertheless the magnitude of the interaction between economic status and positive experiences varies across groups. For Latinos and Asian Americans, the interaction is statistically significant and comparable in size, as reflected in the roughly parallel downward sloping solid lines in Figure 1 (panels h and i). The more gently sloped solid line in Figure 1(g) illustrates that the preferences of African Americans on the set of government action items deviate somewhat from those of Latinos and Asian Americans. For African Americans, the negative direction of the interaction between economic status and positive experiences is consistent with the prediction of the theory, but the size of the interaction effect is not statistically significant.
The regression model on this set of issues also provides little explanatory power in the case of African Americans (R2=.02) because aggregate support for government action is so high that it is difficult to differentiate among the preferences of African Americans. The government action items therefore prove “too easy” for black respondents, although this easiness can also be seen to reflect the stronger orientation of African Americans, compared to Latinos and Asian Americans, toward government efforts to address inequality. Some of these residual group differences may be attributable to unmeasured factors such as the intense socialization processes within the African American community that reinforce perceptions of a racial group interest (Allen, Dawson, and Brown 1989; Dawson 1994).
In sum, the effect of economic status on support for racial and ethnic group interests is statistically insignificant as predicted in eight out of nine tests (three measures of group interest by three groups) among individuals who face frequent discrimination and believe they do not enjoy equality of opportunity. The lone exception is that higher status leads to stronger opposition to affirmative action among Asian Americans even if they hold negative evaluations of society. When we combined the measures of opportunity and discrimination to identify those individuals who perceive both equal opportunities and report infrequent discrimination, the interaction between this measure of socioeconomic acceptance and economic status is negative and statistically significant, as predicted, in seven out of nine tests. The interactions are also negative in the remaining two cases, but do not achieve statistical significance.
The observed patterns are sufficiently consistent across racial and ethnic groups that a common model using the group interest scale as the dependent variable describes the dynamics of opinion in all three groups. We pooled together the three groups and constructed a saturated model that tested for two-way interactions between economic status and experiences, economic status and minority group, and experiences and minority group, and a three-way interaction among experiences, economic status, and minority group. The results show that economic status significantly weakens support for group interests among all individuals who perceive equal opportunities and report infrequent discrimination; but that the effect of economic status is reduced to insignificance among those who perceive unequal opportunities and report frequent discrimination. However the three-way interaction among economic status, assessment of opportunities and discrimination, and minority group is not statistically significant, meaning there are no differences in the effects of changing economic status among the three minority groups once we take into account the conditional effects of the experiences of status.
Thus the overall results provide strong evidence for the proposition that higher economic status diminishes support for racial and ethnic group interests only among those who believe they can be socially mobile in American society. On average, Latinos and especially Asian Americans are more sanguine than African Americans about the opportunities available to their own groups, a distributional difference that explains our earlier results (in Table 1) showing that economic status tends to have a larger impact on the attitudes of Latinos and Asian Americans than of African Americans. But these are only group tendencies; within each group, there are economically successful individuals who downplay the significance of race or ethnicity because they perceive social and economic boundaries to be fluid.
On the other hand, African Americans, Latinos, and Asian Americans who attain material security but experience hostility toward their success and continuing discrimination are loath to relinquish racial or ethnic considerations in their political decisions. These results both corroborate and extend previous research on African American racial consciousness by identifying the general conditions under which support for group interests among all minorities will either be strengthened or weakened by changes in individual life circumstances.
SUMMARY AND CONCLUSION
The study of racial politics in the United States can no longer be confined to analyzing contrasts between Blacks and Whites. Population changes in the last 40 years resulting from the massive influx of immigrants from Latin America and Asia after the Immigration Act of 1965 have created a new demographic context for the politics of race and ethnicity. Although Whites still constitute a numerical majority, minority groups now wield considerable potential power in electoral politics because of their high concentration in several parts of the country.
As we shift our attention to this new multiracial environment, an important question is whether individuals in these groups will share a political outlook and exercise political power as members of racial and ethnic groups. Our analysis shows that intergroup differences in attitudes toward racial and ethnic policies are explained by varying experiences among minority groups rather than by different theoretical processes. Vital differences between African Americans and other minority groups in their experiences of economic status affect their respective tendencies to embrace a racial or ethnic identity and pursue group interests in public policy. Among Latinos and Asian Americans, economic status is correlated with more favorable assessments of race relations. In contrast, middle-class African Americans retain negative assessments about group opportunities and report more personal discrimination than lower status African Americans. Therefore, racial consciousness tends to remain stronger among affluent African Americans, whereas successful Latinos and Asian Americans place less emphasis on racial or ethnic considerations in their political attitudes and policy preferences.
But these findings also suggest the general proposition that the effect of economic status on the group consciousness of all minorities depends on the experiences accompanying that status. Indeed, we found that African Americans who have more positive experiences of middle-class status pay less attention to race and show less support for race-based public policies. In general, for all minority individuals who perceive equal opportunity and experience social acceptance, an improved standard of living tends to lead to a weaker focus on race and ethnicity. On the other hand, higher economic status fails to diminish the salience of race and ethnicity among those who encounter frequent discrimination. Nonetheless, there remain residual differences across minority groups even after opportunities and discrimination are controlled, indicating that group identification is somewhat stronger among African Americans.
An important lesson from this analysis is that support for racial and ethnic group interests is strengthened by the failure of society to provide equality of opportunity and weakened by favorable experiences of economic status.20
Alba and Nee (2003, 278) similarly argue that the assimilation of contemporary immigrants depends on creating democratic institutions that protect the civil rights of minorities, allow for socioeconomic mobility, and eliminate racial discrimination.
People emphasize or downplay their racial and ethnic group memberships partly depending on the utility of those identifications. When minority status carries a stigma and presents an obstacle to personal achievement, individual group members may try to differentiate themselves from the group by avoiding the attributes and behaviors associated with it. This is a form of exit from the group. When we speak of social mobility, we normally think of socioeconomic or class mobility, but we can also include mobility between ascriptive categories. People can take measures to heighten or conceal their racial and ethnic backgrounds or their sexual identities by modifying their behavior and appearance (Chong 2000).
But individuals cannot totally control how they are defined, for their social identities are affected by how others perceive them. Some individuals may prefer not to be categorized by their race or ethnicity, but others may treat them stereotypically and thereby increase the salience of these group identifications. If race and ethnicity restrict opportunities for social mobility, minority group members may not have the option to choose other forms of social identification. Instead of reducing their racial or ethnic ties, they may alternatively embrace their group and pursue various means to improve its status. These strategies may include collective action to achieve political goals; systematic efforts to increase the value of group attributes that have been underappreciated by the majority culture; and social comparisons with less (rather than more) powerful groups in society, a strategy that may heighten competition and conflict among socio-economically disadvantaged groups (Chong 2000; Hogg and Abrams 1988).
In contrast, individuals who feel less constrained by their minority status are more likely to downplay racial or ethnic considerations when their economic situation improves. As their need for group identification diminishes, such individuals give greater weight to their own life circumstances as opposed to the group's condition, and are more inclined to evaluate government policies in terms of how those policies will impinge on themselves.21
Minority individuals may nevertheless retain racially or ethnically based identities and affiliations for social and cultural reasons that are not explicitly instrumental to economic and political outcomes (Alba 1990).
Although we found considerable evidence of a common theoretical process affecting all minorities, there remain hints in the data that support for group interests is more robust among African Americans. Compared to higher status Latinos and Asian Americans, economically secure African Americans were more likely to maintain support for government action to obtain racial equality in employment, education, health care, and the administration of the law even when they evaluated socioeconomic conditions favorably. As we noted earlier, such residual contrasts between African Americans and other minorities may reflect some of the unmeasured contextual factors not accounted for in our model. Racial consciousness among African Americans may be sustained, despite improvements in living standards, because African American communities, more so than Latino and Asian American communities, assign greater utility to collective action and contain institutions, such as African American churches and mass media, that promote perception of a racial group interest (Allen, Dawson, and Brown 1989; Dawson 1994; Tate 1993). Neither Latinos nor Asians have experienced a watershed event such as the Civil Rights Movement that fortified African Americans' belief in the instrumental value of group solidarity as a political resource (Chong 1991; McAdam 1982; Morris, Hatchett, and Brown 1989). Therefore even middle-class African Americans who perceive progress toward racial equality and diminished prejudice against them may continue to draw lessons from recent political history and believe their individual interests are furthered through political solidarity with fellow African Americans.
Finally we do not rule out the possibility that individualistically oriented minorities can be mobilized to adopt a group or collective frame of reference. Those who are individualistic in some circumstances can be group oriented in others depending on the nature of the issues and the local political conditions and contexts in which they arise (e.g., Cohen and Dawson 1993; Gay 2004; Junn 2003). Although the material and educational successes of Asian Americans would seem to undermine their incentives to adhere to a racial group identity, both the record of discrimination against Asian Americans and their historical exclusion from politics and economic activities (e.g., Ancheta 1998; Chang 2001) remind Asian Americans that their life chances continue to be defined in part by their race. Asian Americans therefore are more ambivalent between individualist and collective approaches to politics and less likely than either African Americans or Latinos to pursue a political strategy built around racial identity. But the racial identity of Asian Americans nevertheless may be invoked effectively in specific contexts or on issues that directly address their collective interests.
Likewise, ethnic divisions and generational differences among Latinos may hinder political mobilization around ethnic identity, but contemporary political conflict on salient issues such as English-only legislation and immigration and social welfare reform may prove to be a unifying force (Cain, Citrin, and Wong 2000; de la Garza 2004).22
de la Garza et al. (1992) found that Mexicans, Puerto Ricans, and Cubans are more likely to identify with separate ethnic groups than with a common pan-ethnic group.
APPENDIX
The survey analyzed in this paper was sponsored by the Washington Post, Kaiser Family Foundation, and Harvard University, and conducted by International Communications Research during March 8–April 22 and May 16–20 in 2001. The data are available from the Roper Center for Public Opinion Research. The total sample size is 2,717. In the first round of the survey, 1,709 respondents were interviewed. Because White respondents were not asked about their direct experiences of discrimination in the first survey, a brief additional survey of 1,008 respondents was conducted to gather information on these questions (Q57, 58, 59(a–d), and 59a1(a–b)). In our research, we used only the respondents who answered the first survey. Thus, the sample size for our study is 1,709. We used Stata version 8.0 statistical software for our analysis.
Dependent Variables
Attention Paid to Race
Q12_a. Is there too much, too little, or about the right amount of attention paid to race and racial issues these days? (Too little = 3, About the right amount = 2, Too much = 1)
Recruiting Qualified Minorities
Q51. Do you favor or oppose employers and colleges making an extra effort to find and recruit qualified minorities? (Favor = 1, Oppose = 0)
Support for Government Action to Ensure Equality
A 4-item scale was created from the following items: Q12a–d. Do you believe it is the responsibility or isn't the responsibility of the federal government to make sure minorities have equality with Whites in each of the following areas, even if it means you will have to pay more in taxes? Making sure minorities have:
a. Jobs equal in quality to Whites; b. Schools equal in quality to Whites; c. Health care services equal to Whites;
d. Treatment by the courts and police equal to Whites
Independent Variables
Economic Status
A 6-item scale was created from the following items:
Q3. Would you describe the state of your own personal finances these days as excellent, good, not so good, or poor? (Excellent = 4, Good = 3, Not so good = 2, Poor = 1)
Q61a–c. For each of the following, please tell me whether or not it is something you and your family have had to deal with recently:
a. You have had problems paying the rent or mortgage for yourself or your family; b. You have delayed or had trouble getting medical care for yourself or your family;
c. You have been unable to save money for future needs.
(Have had to deal with = 1, Have not had to deal with = 0)
Q62. Do you own stocks, bonds, or mutual funds—either directly or through a 401K plan? (Yes = 1, No = 0)
D11, D11a–b. Your total annual household income from all sources and before taxes.
Cronbach's alpha for the scale is .732.
Perception of Opportunities
Q6a–c. Do you feel that [a. African Americans; b. Hispanic Americans; c. Asian Americans] have more, less, or about the same opportunities in life as Whites have? (More opportunity = 3, About the same = 2, Less opportunity = 1)
Experience of Discrimination
A 7-item scale was created from the following items: Q 58. During the last 10 years, have you experienced discrimination because of your racial or ethnic background, or not? (Yes = 1, No = 0)
Q59a–d. In your day-to-day life, how often do any of the following things happen to you because of your racial or ethnic background? Would you say very often, fairly often, once in a while, or never?
a. You are treated with less respect than other people;
b. You receive poorer service than other people at restaurants or stores; c. People act as if they are afraid of you; d. You are called names or insulted.
(Very often = 4, Fairly often = 3, Once in a while = 2, Never = 1)
Q59a1a–b. Have you ever been (items a, b) because of your racial and ethnic background?
a. Physically threatened or attacked; b. Unfairly stopped by police.
(Yes = 1, No = 0). Cronbach's alpha for the scale is .789.
Perception of Opportunities and Experience of Discrimination
“Perception of opportunities” was dichotomized by recoding “more opportunity” and “the same opportunity” to 1 = positive perception of opportunity; and “less opportunity” to 0 = negative perception of opportunity. Similarly, the scale variable “experience of discrimination” was dichotomized by dividing the scale at the 65th percentile for the population. The values below 65% of the distribution are recoded to 0 = low levels of discrimination, and above the cut point to 1 = high levels of discrimination. The cut point was adjusted to ensure a meaningful distribution of African Americans across high and low categories. There is still a skewed distribution of African Americans even after we use this cut point: about 28% of African Americans fall in the low category, whereas 72% fall in the high category. We combined the two dichotomized variables to create three ordered categories: “negative” experiences of opportunity and discrimination = 0, “mixed” experiences (i.e., either positive assessments of opportunities or infrequent encounters with discrimination, but not both) = 1, and “positive” experiences of opportunity and infrequent encounters with discrimination = 2.
Education
D09. What is the last grade or class that you completed in school? (Responses range from none to post-graduate training or professional school after college.)
Ideology
D02. Would you say your views in most political matters are liberal, moderate, conservative, something else, or haven't you given this much thought? (Liberal = 1, Moderate = 2, Conservative = 3, Something else = 4, Haven't given this much thought = 5). In the analysis, those who said “something else” or “haven't given this much thought” were recoded as “moderate” to reduce the number of missing cases.
We received valuable comments on this paper from Jack Citrin, Jamie Druckman, Claudine Gay, Jane Junn, Taeku Lee, Reuel Rogers, Michael Wallerstein, Yael Wolinsky, Cara Wong, and the Editor of the APSR and its anonymous referees. We also thank the participants in workshops in the Department of Government at Cornell University, the Institute for Policy Research at Northwestern University, the Eagleton Institute of Politics at Rutgers University, the Institute of Governmental Studies at the University of California, Berkeley, and the Department of Political Science at the University of Chicago, where earlier versions of the paper were presented.
