Gender Differences in Whites' Racial Attitudes - CiteSeerX

0 downloads 0 Views 510KB Size Report
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of ... Department of Sociology, Virginia Polytechnic.
Gender Differences in Whites' Racial Attitudes: Are Women's Attitudes Really More Favorable? Author(s): Michael Hughes and Steven A. Tuch Source: Social Psychology Quarterly, Vol. 66, No. 4, Special Issue: Race, Racism, and Discrimination (Dec., 2003), pp. 384-401 Published by: American Sociological Association Stable URL: http://www.jstor.org/stable/1519836 Accessed: 27/02/2009 15:06 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=asa. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to Social Psychology Quarterly.

http://www.jstor.org

SocialPsychologyQuarterly 2003,Vol.66,No. 4,384-401

Gender Differences in Whites' Racial Atliludes: Are Women'sAttitudes Really More Favorable?* MICHAEL HUGHES VirginiaPolytechnicInstituteand State University STEVEN A. TUCH George WashingtonUniversity Studies of gender differences in orientation toward others have found that women are more strongly concerned than men with affective processes and are more likely to be other-focused, while men tend to be more instrumental and more self-oriented. Recent research has extended this finding to include gender differences in racial attitudes,and reports that women also are more favorable than men in their racial outlooks. In this study we examine differences between white men and white women across a broader spectrum of racial attitudes with more diverse national samples than were employed previously, including the 1988-2000 General Social Surveys and the 1988-1994 American National Election Studies. Wefind that gender differences in racial attitudes are small, inconsistent, and limited mostly to attitudes on racialpolicy. Ourfindings are consistent with the views that white women's and white men's racial attitudes are rooted in their shared sense of group position, and that gender-differentiated value socialization plays only a small role in racial attitudeformation.

Does women's prosocial orientation lead them to hold more positive racial attitudes than men? A gender socialization argument,

based on the assumption that gender-differentiated value socialization exerts an important influence on attitudes in general (Beutel and Marini 1995; Cross and Madson 1997) * A previousversionof thispaperwaspresentedat and on racial attitudes specifically (Johnson the annual meetings of the American Sociological and Marini 1998), strongly suggests that it Association,held in Chicagoin August1999.The data does so. This argument implies that an effecsets used in this study were made available by the tive mechanism for reducing the racial aniInter-UniversityConsortiumfor Politicaland Social and conflict that still deeply divide mosity Research.The data for the General Social Survey, NationalData Programfor the SocialSciences,origi- American society in the early twenty-first nally were collected by JamesA. Davis and TomW. century would be to socialize people, espeSmith of the National Opinion Research Center, cially males, to have the kinds of prosocial Universityof Chicago,and distributedby the Roper value orientations typical of females. Public Opinion Research Center, University of An alternative argument, based in the Connecticut.The data for the American National work of Blumer (1958), is that racialattitudes Election Study originally were collected by the Centerfor PoliticalStudiesof the Institutefor Social emerge from the structuralrelations between Research,Universityof Michigan,underthe direction groups,and that prejudiceis linked inseparaof WarrenE. Miller,Donald R. Kinder,and StevenJ. to what Blumer called a "sense of group Rosenstone.Neitherthe originalcollectorsof the data bly Because women and men in the position." bear for the nor the Consortium any responsibility analysesor interpretationspresentedhere.We thank same racial category occupy the same posiCarol A. Bailey, SandraL. Hanson, K. Jill Kiecolt, tion in the racial hierarchy, this argument Carolyn J. Kroehler, James W. Michaels, Rachel predicts that women's and men's racial attiParker-Gwin,Lee Sigelman,and MaryleeTaylorfor tudes should show little difference,regardless helpfulcomments.The authorsshareequal responsibilityfor the conception,design,andexecutionof this of any gender difference in prosocial orientapaper. Direct correspondenceto Michael Hughes, tion. In the present study we evaluate these Department of Sociology, Virginia Polytechnic argumentsusing data from two large, nationInstituteand StateUniversity,Blacksburg,VA 24061the General or StevenA. Tuch,Departmentof ally representative surveys, 0137,[email protected], and the American Social (GSS) Survey Sociology, George Washington University, National Election Study (ANES). Washington,DC 20052,[email protected]. 384

GENDER AND RACIAL AT'I''I'UDES BACKGROUND Genderand Normative Orientations Substantial evidence indicates that women are more strongly concerned about others and about social relationshipsthan are men. For example, Cross and Madson's (1997) review of research on gender and selfconstrual shows that social relationships and connections to others are more likely to be a part of women's than of men's self-concepts; as a consequence, women's moral decisions are more likely to reflect others' needs as well as obligations to others. These differences are consistent with findings in the literature on gender differences in behavioral and normative orientations (Beutel and Marini 1995) showing that women have closer relationshipswith others than do men, are more deeply involved in caregiving,are more prosocial and altruistic, and are more compassionate and empathetic. These conclusions are consistent with a wide array of findings on gender differences in orientations toward violence showing that men are far more likely than women to support violent or forceful options across an arrayof social control,foreign affairs,and law enforcement situations (Smith 1984). Research on policy preferences reported in many national studies follows a similar pattern:women are more likely than men to prefer policies that oppose violence and support compassion, regulation, and protection (Shapiro and Mahajan1986). Overall these gender differences in normative orientations strongly suggest that women would be more likely than men to support interracial interaction and policies promoting racial equality,and to reject dehumanizing racial stereotypes-in short, that women's racial attitudes would be more favorable than men's. Not every study, however, reports that women's typical value pattern is expressive and communal and that men's is typically instrumentaland autonomous.The review in Prince-Gibson and Schwartz (1998), for example, shows that studies of attachment values and traditional work values reveal no gender differences. Prince-Gibson and Schwartz's own (1998) study uncovered no gender differences in either value priorities

385

or the meanings of values across 10 value types: power, achievement, hedonism, stimulation, self-direction, universalism, benevolence, tradition, conformity, and security. These findings suggest that if gender differences are present in racial attitudes, they would not be due to values. If racial attitudes are linked strongly to values, however, these findings would suggest little or no gender difference in racial attitudes. Genderand Theoriesof Whites'Racial Attitudes Theories of racial attitudes (for a review, see Sears, Hetts, Sidanius, and Bobo 2000) have not explicitly considered gender as an importantvariable.Yet theories of racial attitudes that implicate personality dimensions as determinants of such attitudes are consistent with the idea that gender differences in personality could lead to gender differences in these attitudes.Symbolic racismtheory,for example, which emphasizes individualism and antiblackaffect (Kinder and Sears 1981), authoritarian personality theory, which emphasizes authoritarianismand aggressiveness (Adorno et al. 1950; Altemeyer 1994), and social dominance theory, which emphasizes social dominance orientation (Sidanius 1993), all focus on personality dimensions that the literature on gender differences in self-construal (Cross and Madson 1997) and value orientations (Beutel and Marini 1995) suggests would be more prominent in men than in women. Social structural theories, on the other hand, do not imply that men and women should have different racial attitudes.These include realistic group conflict theory (Bobo 1983;D. Campbell 1965), group position theory (Blumer 1958;Bobo, Kluegel, and Smith 1997), and various approaches that combine race and class (Bonacich 1980). Although these approaches differ widely in their specifics, they are generally consistent with Blumer's (1958) idea that racial prejudice and antagonism are reflections not of individual personality traits, but of competition and conflict between groups for material rewards, power, and status in a racialized society. A social structural approach would suggest that because white men and white

386

SOCIAL PSYCHOLOGY QUARTERLY

women share the same racial position, they would hold similar racial attitudes.Indeed, if such attitudes reflect a sense of racial group position, then white women's position as whites should be considerably more important than their experience as women in determining their racial attitudes. Researchon Genderand RacialAttitudes Surprisinglyfew studies have examined gender differences in racial views. Most of the research on whites' racial attitudes, including several influential early studies (e.g., Greeley and Sheatsley 1971; Hyman and Sheatsley 1956, 1964; Taylor, Sheatsley, and Greeley 1978) as well as more recent work (Firebaugh and Davis 1988; Schuman, Steeh, and Bobo 1985),basicallyignored gender. Other studies employ gender as a control variable but show little concern with its effects; they include little or no comment on the substantiveimplications of gender differences (Bobo and Kluegel 1993;Hughes 1997; Kinder and Sanders 1996; Thch 1987; Tuch and Hughes 1996a, 1996b). A few studies have described and commented briefly on the association of gender with racial attitudes (Baxter and Lansing 1983; Schuman et al. 1997;Steeh and Schuman 1992). Studies that do present data on white men's and women's attitudes about race only partiallysupport the theoretical expectations suggested in reviews by Cross and Madson (1997), Beutel and Marini (1995), and Shapiro and Mahajan (1986). Studies of social distance, for instance, show that men are more accepting of close social relations with members of other groups than are women. Bogardus reported this gender difference in his original studies (e.g., Bogardus 1928) and also in his later research (Bogardus 1959). Using a national sample, Owen, Eisner, and McFaul (1977) found the same pattern in their replication of Bogardus's work. Muir's and McGlamery's studies of social distance on a college campus also uncovered a pattern in which men reported greatertolerance of close interracialrelationships than did women (Muir 1990;Muir and McGlamery 1984). Both Bogardus's and Muir and McGlamery's studies include data on the acceptability of each of several kinds

of social relationshipsacross a range of closeness; they show that men are more likely to accept persons of other races in more intimate relationships (e.g., to marry,date, have as a "chum,"or room with), whereas women are more likely to accept less intimate interracial relationships (e.g., as neighbor or coworker,or to sit next to or eat with). Another early work, Angus Campbell's (1971) study of whites' attitudes toward blacks in 15 American cities, was conducted in the wake of the civil disordersand violence in Detroit, Newark, and other urban areas in 1967. This work uncovered very little in the way of substantively meaningful gender differences across an array of attitudes including interracial contact, perceptions of discrimination, sympathy with blacks' protest, support for civil rights legislation, and improving blacks' living conditions. Although only about 10 percent of respondents gave negative responses to all four of the interracialcontact questions asked in the survey, white women were almost twice as likely as white men to do so (12% vs. 7%). Poole and Ziegler's (1985) study of gender differences in political attitudes showed little difference between men and women in the salience of the issue of government help for minorities; among women, this was the least salient of any issue. In addition,men and women were virtually identical in their support for government help for minorities.The Schlozman et al. (1995) study of gender and political participation uncovered very little difference in patterns of political participation: men were slightly more likely than women to participate in political activity. Among those who were motivated to participate because of concern about particular issues, white men and white women were equally likely to say that they were motivated to do so by concern for civil rights or minorities (1% of both men and women). In Bobo and Kluegel's 1993 study of racial policy attitudes, white women were more supportive than white men of policies that would enhance opportunities for blacks, but were not more likely than men to support increased spending to assist blacks or government help to improve blacks' standardof living.According to studies byTuch and Hughes (1996a) and by Hughes (1997) of various

GENDER AND RACIAL A'II'UDES racial attitudes, white women were significantly more favorable than white men only on affirmative action. In Steeh and Schuman's 1992 study, women were more favorable than men on two of three items concerning affirmative action but on only three of nine other racial policy attitudes. Analyzing gender differences in whites' racial attitudes across a number of attitudinal dimensions,Schuman et al. (1997) concluded that women were more favorable than men on most social distance and racial policy attitudes. On some issues, however, they found no difference. In the most intimate areas of racial contact, women were less favorable than men, a finding that echoes the results of Bogardus (1928, 1959) and of Muir (Muir 1990; Muir and McGlamery 1984) reported above. In a study using the 1996 GSS that examined both gender and racial differences in gender and racial attitudes, Kane and Kyyro (2001) reported no significant gender differences in whites' racial attitudes about racial segregation or affirmative action. They found, however, that women were more likely to name discriminationrather than lack of willpower as the reason for African Americans' poor life chances. Overall, Kane and Kyyrofound that racial differences were more pronounced than gender differences across the various attitudes they studied, particularlyfor racial attitudes. Taken together, these studies of gender and racial attitudes support the conclusion that gender has very little relation to racial attitudes. This is what we would expect if racial attitudes reflected a sense of group position, as argued by Blumer (1958); it would suggest that the attitudes of white women in the United States regarding African Americans are influenced very heavily by their position as whites in a society that privileges whites over African Americans. Gender,Socialization,and RacialAttitudes A fairly recent study by Johnson and Marini (1998) is the only research to date that employs a large national sample in examining the nature and meaning of the association between gender and racial attitudes. Exploring social distance attitudes

387

among black and white high school seniors in data collected from 1976 to 1992, these authors found that women in both racial categories were more likely than men to agree that interracial contact was desirable. This relationship was not affected by controls for political ideology, religiosity, nor interracial friendships. Figure 1 shows gender differences among whites at the zero-order level for each item making up Johnson and Marini's social distance index for the most recent data collection period covered by their study (1991-1992).' (See Appendix 1 for the item wording and response categories used in this index.) Johnson and Marini explain women's more positive racial attitudes on the basis of differential gender socialization: they argue that women are socialized to be more concerned for others and more strongly focused on the importance and meaning of social relationshipsthan are men, and conclude that there is "a potential for improving race relations through socialization of males" (1998: 256). The Johnson and Marini (1998) study is important because it suggests that the difference between women and men in racial attitudes may be even greater than indicated by other studies. It is limited, however, because of the sample (adolescents), the racial attitude measure (a unique social distance index),2 and the analytical strategy, 1

Data and items used by Johnson and Marini (1998) are taken from the 1976-1995Monitoringthe Future survey (conducted by Jerald G. Bachman, Lloyd D. Johnston,and PatrickM. O'Malleyof the Survey Research Center, Institute for Social Research, University of Michigan). To construct Figure 1, we calculatedmeans from the 1991-1992 data. 2 SinceBogardus(1928),most socialdistancescales havebeen basedon the idea thatpeople tolerateclose socialrelationshipswithmembersof othergroupsless readily than more impersonal relationships. For instance,people are more likely to accepta member of another group as a co-workerthan as a spouse. Itemsin socialdistancescalesaskrespondentsa number of questionsaboutthe acceptabilityof intergroup socialrelationsvaryingin the degreeof closeness.The closestdegreeof socialrelationshipthatis acceptable is taken as the measureof social distance.A person who will accept members of another group as coworkers,but nothingcloser,expressesgreatersocial distance than a person who finds close friendship

SOCIAL PSYCHOLOGY QUARTERLY

388 4-

3.5

3.42

3.319

3.31

3.3

3.22

3.22

3-

2.5 -

2-

1.5 -

.

.

I

,

I

1. Close Friend

2. Next-Door Neighbor

3. Some Neighbors 4. Child's Friends

I1

Female ?Male

5. In Child's School

6. Co-Worker

7. Job Supervisor

|

Figure 1. Mean Scores for Items in Johnson and Marini's (1998) Social Distance Scale: White High School Seniors, 1991-1992

which allowed no test of the socialization hypothesis. THE PROBLEM Blumer's (1958) definition of racial prejudice as a sense of relative group position locates the source of racial antagonism not acceptable. In addition, traditional measures of social distance possess the scalability properties of a Guttman scale (Upshaw 1968), such that, by knowing the closest relationship acceptable, we can predict quite accurately that all of the more distant relationships are also acceptable. The index used by Johnson and Marini, however, is not based on measures that reflect this scalability assumption of traditional social distance measures. As Figure 1 shows, analysis of data used in their study indicates that people rate close relations with members of other groups (e.g., a close friend) more positively than they rate distant relations (e.g., neighbor or co-worker), the opposite of the pattern found in studies by Bogardus (1928,1959) and Muir (Muir 1990; Muir and McGlamery 1984). This may be the case because their measure taps attitudes about persons of"a different race," not about persons of specific racial or ethnic groups, as in Bogardus and Muir. Also, and importantly, the response category reflecting the least social distance in Johnson and Marini's items taps desirability of social relations. This is a significant departure from the categories used in traditional social distance measures, which are limited to acceptability of contact. We return to this issue below.

only in the negative racial affect displayed by many whites toward blacks, but also, and more generally,in perceptions of competition between groups in a racializedsociety (Bobo 1999). According to this argument,prejudice reflects majority-group members' sense of entitlement to valued resources such as jobs and housing. If Blumer's analysis were correct, we would expect gender differences in racial attitudes to be small and inconsistent. On the other hand, if a value socialization model is correct-if gender differences in socialization produce women who are more concerned with affective processes and who tend to be "other-focused"and men who are more instrumental and more self-orientedthen we would expect consistent and fairly substantial gender differences in racial attitudes.We also would expect that measures of concern for others would substantially explain gender differences in racial attitudes. In the present study we use data from two large, nationally representative omnibus surveys conducted in the United States to examine gender differences across a wide array of attitudinal dimensions including social distance, symbolic racism, racial policies, stereotypes, and perceptions of discrimination, and to determine whether any

GENDER AND RACIAL A'l'TUDES differences we observe can be explained by the gender socialization argument. DATA AND METHODS Data We use the 1988-2000 General Social Surveys (GSS) and the 1988, 1990, 1992, and 1994 American National Election Studies (ANES) to examine gender differences in racial affect. We combine several years from each data set to ensure sample sizes large enough to sustain our analyses.As described below, the use of two data sets provides the opportunity to test the association between gender and racial attitudes with multiple measures of the empathy orientation that presumablyunderlies this relationship. General Social Survey. The GSS is a nationally representative survey of noninstitutionalized adults in the contiguous United States. (For a complete discussion of sampling methodology, see Davis and Smith 2000.) From the GSS data we derived measures of 26 racial attitudes spanningsix broad categories: racial stereotypes, social distance attitudes,attitudes toward racialpolicies, perceptions of discrimination,symbolic racism, and feelings of sympathy and admiration toward blacks.Appendix 2 contains the complete wordings of questions for all items. American National Election Study. The other data set we use is the American National Election Studies (ANES) for 1988 through 1994.These biannual surveys collect data from national probability samples of respondents;like the GSS, they are designed to be representative of the noninstitutionalized adult population of the United States. (For more information on these data and on sampling methodology, see Miller 1987; Miller, Kinder, and Rosenstone 1992, 1993.) The ANES includes 13 variables tapping respondents' attitudes about racial policy, beliefs in racial stereotypes, and symbolic racism.(For question wordings,see Appendix 2.) Analyses of racial attitude items in the ANES after 1994 are not included here because the full complement of dependent and control variables was not available (see below).

389

Analytic Strategy Our analysis is limited to white respondents in the GSS and ANES data sets. We present unstandardizedOLS regression estimates or (for dichotomous dependent measures) binary logistic regression estimates for the racial attitude items derived from the GSS and ANES data.We report either two or three sets of coefficients for each dependent variable. First, the zero-order effect of gender is displayed. Next, we report a set of coefficients that are net of controls which may mediate gender's effect on racial tolerance: education, measured in years of schooling completed; age, in years; political ideology, measured on a seven-point scale from extremely liberal to extremely conservative; religiosity, tapped by a two-item additive index consisting of frequency of attendance at religious services,which ranges from never to several times a week, and frequency of prayer, which ranges from never to several times a day;3 region (south vs. nonsouth); urban-nonurbanresidence;4and (if data on a dependent variable are available for more than one survey year) year of interview. In cases where the effect of gender remains significant net of controls,we report the results of incrementingthe model with a measure of respondents' orientations toward helping the poor and/or disadvantaged. If (as Johnson and Marini argue) the effect of gender on racial attitudes is explained by women's greater tendency toward empathetic and caring mindsets, then controlling on a helping variableshould reduce or eliminate the effect of gender on racial attitudes. Each data set offers a different measure of a helping or caring orientation.In the GSS this orientation is measured with a question 3 In the ANES, religiosity is measured by church attendance only. 4 Urban is defined as residence within an SMSA and a large central city (over 250,000) or suburb, a medium-sized central city (50,000 to 250,000) or suburb, an unincorporated area of a large central city, or an unincorporated area of a medium-sized central city. Nonurban is defined as residence outside an SMSA and in a small city (10,000 to 49,999), a town or village (2,500 to 9,999), an incorporated area with a population of less than 2,500, an unincorporated area with a population of 1,000 to 2,499, or open country.

390

SOCIAL PSYCHOLOGY QUARTERLY Otherwise this is an appropriatemeasure of what Beutel and Marini (1995) describe as women's greater sense of responsibility toward the less privileged and greater support of economic policies designed to help those individuals. The ANES analyses use a "feeling thermometer" index to measure sympathy and compassion. The index consists of four variables, each measuring how positive or how negative the respondent feels about particular groups that are oppressed and might benefit from some care and concern from others: poor people, people on welfare, illegal immigrants, and homosexuals (i.e., gay men and lesbians). Respondents were asked to rate how favorable they felt about these persons and groups; higher ratings indicated more positive affect. Ratings of 50 to 100 indicated favorable and warm feelings; ratings of 0 to 50 indicated that the respondent did not feel favorable and did not care much for the person or group.We summed the ratings into a feeling thermometer index. Alpha reliability coefficients ranged from .70 to .74 across the four survey years.

about improving the living standards of all poor Americans: "I'd like to talk with you about issues some people tell us are important. Please look at this card [respondent is shown a card]. Some people think that the government in Washingtonshould do everything possible to improve the standardof living of all poor Americans;they are at point 1 on this card. Other people think it is not the government's responsibility, and that each person should take care of himself [sic];they are at point 5. Where would you place yourself on this scale, or haven't you made up your mind on this?" The card shown to respondents contains responses rangingfrom "I strongly agree that the government should improve living standards" (coded 1) to "I strongly agree that people should take care of themselves" (coded 5). One might argue that the reference to the "government in Washington"makes this GSS question about helping a less inclusive measure by reducing the likelihood of agreement among those who have a helping orientation but also a principled opposition to specifically government intervention. Another problem with this measure is that some respondents may interpret government action to improve the living standardsof the poor as something that would be likely to benefit people of color. Although research has shown that whites' attitudes about helping the poor and helping blacks are distinct (Bobo and Kluegel 1993), we suggest caution in using this variableto interpretthe findings.

FINDINGS Stereotypes The first panel of Table 1 displaysregression estimates for the effects of gender and the control variables on five racial stereotypes. Three of the stereotypes-views that blacks are lazy,violence-prone, and unintelli-

Table1. OLS andLogisticRegressionEstimatesfor the Effect of Genderon RacialAttitudes,1988-2000 GeneralSocialSurveysand 1988-1994AmericanNationalElectionStudies Variables - A Dependent

Model 1 Coeff. R2

Stereotypes GSS .001 .000 Lazy .204* .003 Violent .133 .002 Unintelligent .034 .000 Unpatriotic -.147 .002 Welfare-dependent ANES .037 .000 Lazy .044 .000 Violent .053 .000 Unintelligent SocialDistance GSS -.293 .002b Favorsegregatedschoolsa Againstinterracialmarriagea .130 .001b

Model3

Model2 n

Coeff.

3,053 1,659 2,308 670 690

-.027 .129 .117 .026 -.137

.070 3,053

1,894 1,877 1,880

-.013 .006 .004

.033 1,894 .033 1,877 .046 1,880

5,371 4,268

-.268 -.031

.047b 5,371 .202b 4,268

R2

.035

n

1,659

.010 2,308 670 .045 690 .072

Coeff.

R2

n

GENDER AND RACIALA'TITUDES

391

Table1. Continued. DependentVariables

Model 1 Coeff. R2

Favorsegregatedneigborhoods -.043 .001 -.121* .004 Againstlivingin half blackhalf whiteneighborhood -.218* .003 Do not feel close to blacks PolicyAttitudes GSS

Model2 n

Coeff.

R2

Model3 n

3,016 2,392

-.068* .140 3,016 -.124* .082 2,392

1,366

-.093

.071 1,366 .075b 5,192 .081b 3,359

Coeff.

R2

n

.032 .171 -.119* .097

378 1,313

Against black presidenta Against open housinga

-.193* .001b -.204 .000b

5,192 3,359

-.150 -.039

Againstspendingto improve blacks'standardof living No governmentresponsibility for helpingblacks Againstaffirmativeactionin employment Againstspecialfavorsfor blacks

-.060* .002

3,508

-.056* .071 3,508

-.019

.119

1,978

-.134* .003

4,723

-.099* .070 4,723

-.043

.164

4,298

-.066* .002

3,082

-.055

-.092* .002

3,210

-.085* .113 3,210

-.053

.145

1,571

Against businga

-.175* .002b

5,119

-.186*

-.142

.068b 2,499

.037 3,082

.048b 5,119

-.063 .001 358 -.001 .071 358 Againstenterprisezones -.083 .001 1,688 -.120* .166 1,688 -.119 .172 967 Againstextrafundingfor schoolsin black neighborhoods 364 -.181 .067 364 Againstspecialscholarships -.222* .011 ANES Againstspendingto improve -.055* .002 5,197 -.052* .052 5,197 -.026 .130 4,816 blacks'standardof living No governmentresponsibility -.232* .005 5,209 -.225* .060 5,209 -.152* .125 4,926 for helpingblacks -.102* .002 4,903 -.082* .039 4,903 -.040 .071 4,880 Againstaffirmativeactionin employment -.258* .008 4,033 -.250* .044 4,033 -.203* .079 4,014 Againstaffirmativeactionin college admission No governmentrole in school -.108* .005 3,434 -.103* .031 3,434 -.081* .048 3,417 integration No governmentrole in .046 .001 3,216 .036 .044 3,216 employmentequality Perceptionsof Discrimination GSS Perceiveno discriminationin -.052 .001 698 -.067 .046 698 jobs Perceiveno discriminationin -.033 .000 690 -.051 .046 690 housing SymbolicRacism GSS Blackshave too muchinfluence -.073 .002 668 -.081 .087 668 Whiteslose jobs or promotions .013 .000 2,399 .014 .036 2,399 to equallyor less qualified blacks ANES -.268* .001 4,886 -.321* .111 4,886 -.133 .192 4,865 Symbolicracismindex SympathyandAdmiration GSS Feel no sympathyfor blacks -.214* .018 731 -.141* .128 731 -.127 .178 356 Feel no admirationfor blacks -.212* .018 727 -.140* .105 727 -.074 .128 355 Notes:Coefficientsare unstandardized. Model 1 reportsthe bivariateregressionof racialattitudeson gender. Model 2 addscontrolsfor education,age,politicalideology,religiosity,region,urban-nonurban residence,and year of interview.Model3 controlsfor the helpingvariable(GSS) or the feelingthermometer(ANES). a Binarylogisticregressionestimates. bPseudoR2= modelchi-square/model chi-square+ n * -

p