The Rational Basis of Symbolic Racism

Modern political-science research on racial attitudes suggests that white conservatism stems from symbolism, prejudice and socialized resentment. The implication, sometimes made explicit, is that racial conservatism has no rational competitive basis; it does not grow out of the social structure of intergroup competition. Evidence for this claim usually appears in two sorts of analysis: (1) survey analysis connecting racial conservatism (e.g., opposition to affirmative action or busing) to anti-black “symbolic” value judgments, and (2) cross-level models showing that racial conservatism does not respond to measures of a white respondent’s “self interest.” The enclosed paper questions the extent to which racial conservatism can be passed off to mere psychological orientation. Using data from a particularly valuable racial-issues survey, I show that supposedly symbolic judgments in fact possess an underlying structural basis, one that I term Cultural Backlash. Other researchers missed the geographical structure contained in racial data because intergroup competition does not follow a simple pattern, as statistical models typically assume, but one that interacts with a community’s likely investment in white cultural capital. Paper prepared for the annual meeting of the Midwest Political Science Association, held in Chicago, IL, April 27-30, 2000. The author thanks Derek Bok and Bob Blendon for providing the survey data, the Social Choice Workshop at Binghamton University for advice on a previous draft, and the University of Kentucky College of Arts and Sciences for research support. All mistakes contained herein are the responsibility of the author. Voss The Rational Basis of Symbolic Racism 1 1 “Rationality” does not ascribe superhuman powers of reasoning to the subject, or even conscious deliberation. The outcome of rational, self-interested behavior may be far from satisfactory or optimal, depending upon the bounds within which decision making must take place. Modern political science has fallen heavily under the sway of economic reasoning. Researchers commonly assume that political actors will attempt to maximize their utility, based upon the particular incentives and information they face, and that the outgrowth of this optimization strategy will be behavior that the observer recognizes as fundamentally rational.1 This “social choice” perspective, whether accompanied by formal models or applied more loosely, has colored most political-science subfields (e.g., Fiorina 1977, 1978; Mayhew 1974; Peterson 1994; not to mention extensive work by scholars such as Terry Moe, Kenneth Shepsle and Barry Weingast). The notable exception, within American politics research, is the study of race. Indeed, the trend among those analyzing racial issues has cut in precisely the opposite direction. Early studies of racial conservatism emphasized the extent to which anti-black sentiments in the United States were embedded within the South’s political and economic institutions (Blalock 1967; Key 1949; Heard 1952; Glenn 1963). The most influential recent research, by contrast, operates within an explicit psychological paradigm (Bobo and Licari 1989; Carsey 1995; Gilliam 1996; Hurwitz and Peffley 1997; Kinder and Mendelberg 1995, 404; Kinder and Sanders 1996; McConohay 1982; Peffley, Hurwitz and Sniderman 1997; Sears et al 1980; Sears, Hensler and Speer 1979; Sniderman, Brody and Kuklinski 1984; Sniderman et al. 1991). Both inside and outside the discipline, writers portray conservative racial attitudes as a pathology–at best a product of extremely poor information, at worst an extreme mental illness (Page 1999). One school of researchers, for example, emphasizes the role that stereotype, rumor and abstract political ideology play in shaping issue attitudes (Feagin and Vera 1995; Hurwitz and Peffley 1997; Peffley, Hurwitz and Sniderman 1997). Fear of blacks grows out of prejudicial media communications and lack of cognitive sophistication (Bobo and Licari 1989; Carsey 1995; Entman 1990; Gilliam 1996; Kinder and Mendelberg 1995, 404). A second strain of scholarship recognizes white racial attitudes as a struggle to maintain group dominance (Blumer 1958; Bobo 1988; Giles and Buckner 1993; Giles and Evans 1985, 1986), but relaxes any requirement that group status actually be in jeopardy; the only claim is that whites respond to perceived threat (Bobo 1983, 1,200; Giles and Evans 1985, 58; Pettigrew 1957, 683) or their given level of racial sensitivity (Giles and Buckner 1996; Giles and Evans 1986, 471-72). Probably the most unadorned statement of racial conservatism’s irrationality, however, appears within the “symbolic politics” literature. Scholars bridging the boundary between political science and social psychology offer a long and vibrant research tradition claiming that racial resentment follows a psychological logic–dressed up with socialized values, thick with trite symbolic rhetoric, presumably alleviated by multicultural education or egalitarian contact with minorities. This line of research suggests that one’s position on progressive social policies responds to the assessment of why minorities have fallen behind in the first place. Liberals blame society, and therefore tend to desire institutional solutions to make up for discrimination; those who “blame the victim” see no such need. Presumably psychological orientations either are random or are handed down across generations, because they do not seem to have any basis in the respondent’s actual environment. Whites, that is, suffer an amorphous feeling of resentment with little structural basis aside from unfortunate historical residue. This argument appeared recently in an influential book by Don Kinder and Lynn Sanders (1996), but the fundamental perspective springs up in much of the popular commentary, from the claim that “Americans can’t think straight about race” (DeMott 1995, 179) to Gen. Colin Powell’s nonideological lament that this is a “racist society” Voss The Rational Basis of Symbolic Racism 2 2 By “rational” I do not mean that prejudices are correct or socially desirable. I mean only that they promote material goals, and represent one potentially successful means of achieving those goals. (Page 1999). This paper challenges the dominant paradigm, arguing that racial politics are not exceptional, that standards of instrumental rationality should not be foreign to the study of this crucial social problem, that racial conservatism cannot be passed off as mere psychological orientation. Rather, the alleged symbols dominating racial rhetoric are themselves in part an expression of real intergroup conflict. The apparent absence of “self interest” reflects a cursory analysis of where interest lies in contemporary struggles, and therefore an inadequate attempt to model that interest. Using data from a particularly valuable survey, the Kaiser Race Poll, I show that supposedly symbolic judgments in fact possess an underlying structural basis, one that I term Cultural Backlash. Other researchers missed the geographical structure contained in racial data because intergroup competition does not follow a simple pattern, as statistical models typically assume, but one that interacts with a community’s likely investment in white cultural capital. Looking Behind the Symbols Modern political-science research on racial attitudes suggests that white conservatism stems from symbolism, prejudice and socialized resentment. The implication, often made explicit, is that racial conservatism has no rational competitive basis; it does not grow out of the social structure of intergroup competition. Evidence for the claim usually appears in two sorts of analysis: (1) survey analysis connecting racial conservatism (e.g., opposition to affirmative action or busing) to anti-black “symbolic” value judgments, and (2) cross-level models showing that racial conservatism does not respond to measures of a white respondent’s “self interest.” Explaining attitudes with attitudes is always a risky business. It is very hard to know why whites might embrace a particular assessment of minorities. Travelers in the Old South were often struck by the negative stereotypes that planters had of their slaves and, later, their black wage workers (Howard 1917, 588). Yet few outsiders doubted that these negative impressions thrived precisely because such myths operated in service of rational incentives (Howard 1917, 582-84). Indeed, anti-black stereotypes only developed after slavery required an ideological defense (Wilson 1973, 79). Today’s symbols and stereotypes might be working in service of an equally “rational” goal.2 The symbols may not cause anything; they may be just another product of a racially stratified society. By accepting purely psychological explanations for racial resentment, simply because policy preferences coincide with acceptance of these symbols, researchers risk missing the competitive underpinnings of the social phenomenon they wish to study. As a thought experiment, we can consider what the “symbolic racism” methodology would have indicated about the slave system, since that institution was a rather clear case of group-based exploitation. What would antebellum public-opinion surveys have indicated about white racial attitudes? Judging from anecdotal evidence, whites considered slaves lazy, unprepared for democracy, and content with their condition–a myth shared between slave owners and other whites nationwide. Consider Thomas Jefferson’s relatively genteel “stereotypes” about his slaves (recorded in his Notes on the State of Virginia, Query 14): A black, after hard labour through the day, will be induced by the slightest amusements to sit up till midnight, or later, though knowing he must be out with the first dawn of the morning. They are at least as brave, and more adventuresome. But this may perhaps proceed from a want of Voss The Rational Basis of Symbolic Racism 3 3 I thank Bob Blendon and Derek Bok for making the survey data, codebook, and initial Kaiser report (Brodie 1995) available to me. 4 The exact figures vary depending upon how one treats black Latinos. 5 Even the rival “Group Threat” approach, while emphasizing perceived group incentives, nonetheless makes no specific claim to an individual’s real interests (Bobo 1983, 1,206). In other words, the competing claim is not that racial polarization comes from self-interest, only that it stems from rational motives (c.f., Hardin 1995, 46-49). forethought, which prevents their seeing a danger till it be present . . . They are more ardent after their female: but love seems with them to be more an eager desire, than a tender delicate mixture of sentiment and sensation. The likely result of compiling and analyzing these self-reported attitudes, therefore, would be an argument very much like the “racial resentment” analysis popular today. Antebellum racism would have seemed psychologically rooted, built on inherited symbols and socially constructed myths, rather than driven by self interest. One doubts that this explanation could have had much predictive leverage over the instrumental power relationships clearly driving the system. This paper’s purpose, therefore, is to analyze a few racial resentment variables directly (rather than the issue positions they are supposed to produce). I turn an eye on the symbols and myths that supposedly fuel politics in place of self-interested preferences: beliefs about why blacks are poor, about who gets which social resources, about whether racism is a social problem. My expectation, ultimately confirmed, is that acceptance of these symbols is not random, but a probabilistic function of one’s interests in the contemporary struggle over race. The Henry J. Kaiser Family Foundation funded a particularly valuable survey in 1995, in collaboration with the Washington Post and researchers at Harvard University (Morin 1995).3 The survey was part of a larger study on the role information and misinformation play in determining policy preferences, but this particular portion focused almost exclusively on race. It offers an extensive battery of race-based questions, including informational questions that allow tracing the genealogy of political preferences. Another virtue of the survey is that Chilton Research Services, who actually implemented the Kaiser Foundation Race Poll, recorded the county of residence for most respondents. Having this geographical information allows linking the individual-level data to aggregate contextual variables likely to influence racial views, in particular the estimated black density in each respondent's county. The main drawback, for these purposes, is only the small sample size among whites. The Kaiser poll included an oversample of minorities among its 1,970 respondents–roughly 474 African, 353 Asian and 252 Latino Americans.4 However, the white sample is almost uniformly large enough to explore the survey instruments used in Symbolic Racism research, and the large sample of African Americans will prove useful for one portion of the below analysis. Race Through a Contextual Lens The most influential social psychology research seems to indicate that self-interest plays little direct role in racial attitudes, at least as represented by the density of racial outgroups. The Symbolic Racism school emphasizes this disparity between self-interest and professed beliefs directly (Kinder and Sanders 1996, chap. 4; Sears, Hensler and Speer 1979, 374), in a few cases purportedly rebutting the rational “racial threat” approach (Kinder and Sears 1981, 421-22).5 Glaser (1994) also finds little connection between prejudice and racial context. These results concur with older research on urban Voss The Rational Basis of Symbolic Racism 4 6 Kinder and Mendelberg (1995, 420-21) seem to contradict this dominant interpretation, despite the lead author’s prominence in the Symbolic Racism literature. They dismiss the backlash logic for political expressions of prejudice, but curiously wish to retain the theory for prejudice itself. 7 Turner (1988, 16) actually uses the term “symbolic” to refer to cultural conflict. I avoid this term to prevent confusion with the “symbolic racism” literature. attitudes, which determined that a black community presence managed to ignite conflict yet still shatter stereotypes (Lipset and Raab 1978, 341; Pettigrew 1980, 9).6 In every case, the analysis operationalized “self interest” in an extremely limited manner: either residence in racially diverse communities or, in the case of research on school desegregation, having children who attend schools that might diversify. My intent is to revisit the search for proximity effects, armed with a more complex understanding of self interest than the mere expectation that more blacks means more white hostility. Tracing particular narratives about racial politics to observable (and therefore testable) implications risks a couple of pitfalls, however. The first mistake would be to presume that any geographical pattern in racial attitudes proves that polarization is rational, let alone self-interested. Even purely psychological phenomena can take on a systematic geographical expression over time. For example, migration patterns might result in racially sensitive whites cloistering themselves in homogeneous communities. Similarly, if the antidote to an irrational impulse is exposure to countervailing evidence, then life experiences in an integrated community might mitigate racial conservatism eventually. Another pitfall would be to assume that rational intergroup conflict must follow a strict territorial logic. Certainly ethnic groups sometimes fight for control of a particular territory, as they did during the segregation struggles of the 1960s or as they sometimes do in urban neighborhoods today (Green, Strolovitch and Wong 1998). But conflict can revolve around different stakes, some territorial and some not. Rather than assume that proximity spawns antipathy, that “familiarity breeds contempt,” it is better to differentiate among different forms of conflict and the specific geographical patterns of racial conservatism they might imply. Most forms of social exchange have tangible stakes attached, after all; any might invoke intense feeling from those with vested interest in preserving the status advantages they convey. Following Max Weber (Bendix and Lipset 1966), and tracking the terminology of Turner (1988, 15-16),7 a simple model would be to break racial conflict into the battle for three sources of social status: • Economic: The more blacks in a labor market, for example, the more they seem to threaten white job opportunities (Blalock 1967, 147-50). The more blacks able to move into a particular neighborhood, the more they seem to threaten white property values. The more blacks enrolled in social welfare programs, the more that local government spending saps the economy and decimates the paycheck. It is not quite clear who would exhibit the most extreme racial conservatism in a struggle over economic resources, the whites who profit most from discrimination or the whites whose social class leaves them perilously close to poverty, but few would deny the political importance of a large minority underclass. • Political: The more blacks in an election district, the more discrimination and white mobilization needed to maintain white power (Key 1949; Matthews and Prothro 1963; Blalock 1967, 150-54). If black voters dominate an election, that could mean fewer spoils of office for whites, less attentive police protection, less responsiveness from government bureaucracies, and less influence over legislation. It also might spur government to address social discrepancies among different races (Fording 1997), a condition from which whites otherwise benefit (Glenn 1963). Voss The Rational Basis of Symbolic Racism 5 8 One ironic fact: Blacks from the West Indies enjoy a comparable economic position as whites, despite the double whammy of race and immigrant status (Loury 1995, 137). By coming from the outside, they are insulated from the institutional barriers that limit black achievement. I also suspect that the cultural barriers work in their favor, such as how whites perceive Jamaican accents compared to black ghetto accents. • Cultural: Racial conflict might revolve around cultural exclusiveness intended to reinforce status distinctions (Bourdieu 1986). Racial debates frequently emphasize “cultural capital”–the value attached to cultural traits that are unequally distributed across racial groups–rather than directly economic or political issues (Merelman 1994, Sowell 1994). Whites may wish to preserve dominance over the values promoted in cultural and educational institutions. They may be defending, or in some spheres trying to restore, the status advantages of whiteness. They may be fighting to preserve a cherished way of life, such as values that minorities seem to threaten (Kinder and Sanders 1996, 108-114; Wong and Strolovitch 1996, 4-6). This simple categorization is parsimonious yet useful, since many seemingly unrelated goods pass through some combination of these channels: the power to purchase, the power to appropriate, the power to legitimate. I am particularly interested in the third form of conflict, the struggle over cultural capital, because of the resonance that values enjoy in contemporary racial debate. The battle for political equality is mostly over; the ramparts of segregation and disenfranchisement have fallen; the forces of reaction have retreated. Economic conflict is also less prominent these days. White and black financial resources remain stubbornly far apart (Oliver and Shapiro 1995), but black worker incomes have been rising steadily while white incomes have stagnated. To the extent black poverty is a political problem, it cannot be solved through attempts to remove price discrimination, which skilled black labor no longer faces to any appreciable degree (Loury 1995, 97).8 Rather, lingering racial differences result from the statistical impact of a large, marginalized black underclass that does not find employment at all (Chandra 2000; Loury 1995, 138-39). Breaking apart this concentrated ghetto population, either through housing desegregation or through redistribution of financial resources, has not been a serious political goal of activists on the right or the left. Not even civil-rights leaders, who are predominantly middle class, show much interest (Marable 1995). To the extent economic issues do appear in the contemporary debate over race, it is in the guise of affirmative action, which primarily benefits higher-status African Americans (Loury 1995, 109). Affirmative action predominates not in working class job markets, but in professions such as academia, journalism, and management, and among businesses that accept government contracts. These occupational categories are not localized in a mobile society. Rather, they unite a regional, even a national, market–spreading “threat” around widely. An ambitious journalist who currently works in Gary, Indiana, is no more likely to be squeezed by the Washington Post’s affirmative-action policy than a similar reporter working in Missoula, Montana. The same is true with access into the professions. A white high-school student applying from South Carolina faces roughly as much “threat” from an elite university’s affirmative-action policy as one applying from Vermont. The commonly tested pattern of “rationality” should not really apply. More common is racial tension over the cultural battleground (Green, Strolovitch, and Wong 1998, 375; Wong and Strolovitch 1996, 6). Whites perceive a threat to cherished values, including those central to their vision of American society (Kinder and Sanders 1996, 108-14). Consider how frequently racetinged debates ultimately revolve around cultural stakes: Whites resent collective economic strategies, Voss The Rational Basis of Symbolic Racism 6 9 Just to be extra clear, I am not claiming that blacks genuinely pose a cultural threat, or otherwise endorsing the perspective summarized by these two paragraphs. Nor do the scholars who have popularized the “symbolic politics” approach necessarily see a genuine cultural conflict, aside from a serious divide over how to deal with race in America (Kinder and Sanders 1996, 29-30). which they perceive circumventing traditional paths to prosperity–citing the example of parents or grandparents who were poor but “never asked for handouts from anyone.” Blacks supposedly undermine “middle-class values,” such as hard work and financial self-sufficiency, by seeking institutional favoritism. Blacks seem to threaten “family values,” breeding out of wedlock at a faster rate than they can support offspring, and with multiple partners, an irresponsible behavior that (infuriatingly) does not lessen the sex appeal attached culturally to urban black males. Their influence in schools substitutes the “three R’s” with multiculturalism, patriotism with cynicism, objective tests with feel-good ideological screening, and the classics with trendy leftist scribblings such as those by Franz Fanon. Middle-class blacks use their influence in academia and journalism to undermine American prestige, “turning everything into a racial issue,” and defining themselves as an oppressed caste despite professional success that many Americans can only envy (Loury 1995, 54). Their influence in Hollywood and in the music industry results in violent and sexist entertainment, including vulgar rap music, which corrupts otherwise nice white (and black) kids. Disproportionate black criminality and hatred of whites makes cities and towns unsafe, sharply curtailing the freedom of action that law-abiding Americans once enjoyed.9 Given the prominent role of values-based language, it is worth considering whether cultural conflict does not veil an underlying struggle over instrumental power (Finkenstaedt 1994, 19). White resolve to participate in this cultural struggle, and to view it in racial terms, should not necessarily respond to proximity. Enough has changed, rhetorically at least, to warrant a new assessment of the old proximity logic. In particular, if interest in “white middle-class values” is not shared by all whites equally, any attempt at cultural exclusion will reverberate primarily where whites have something to defend. This rational form of symbolism might follow a backlash pattern outside urban areas, but within the modern metropolis it is in the all-white suburbs where we find America’s “keepers of the culture.” Thus I would expect the conventional backlash pattern to decline or even reverse in metropolitan areas. Based upon the above reasoning, I have derived a handful of testable, interactive geographical hypotheses, each corresponding to a particular narrative about contemporary race relations: 1. Naive Symbolic Politics Approach–Racial conservatism does not grow from objective conditions, but from stereotypes and symbols picked up during childhood. Some whites see African Americans as a threat, while others do not, and individual perspectives scatter haphazardly across the population because they have no rational basis. Racial polarization thus should not follow any meaningful geographic pattern, with relation to black proximity, at any level of aggregation (Berard 1998). 2. Contact Hypothesis–If racial conservatism is fundamentally irrational, then it must stem from ignorance and unchecked prejudice. Stereotypes cannot persist where African Americans are present in sufficient numbers that regular social interaction will educate whites. Proximity thus would lessen racial conservatism as reality imposes itself, shattering prejudices (Carsey 1995). In keeping with refined versions of the concept, we might expect proximity will do the most good where conditions are more equal between two groups (Miller and Brewer 1984, 2). Also, proximity would only be beneficial if the black population has been in place for some length of time; new arrivals have not enjoyed sufficient contact to overturn stereotypes and may even excite white fears (Wong and Strolovitch 1996). 3. Migration Hypothesis–Even if racial conservatism grows out of childhood socialization, rather than any Voss The Rational Basis of Symbolic Racism 7 10 The Authoritarian Personality appears most commonly among lower-status whites, and therefore when applied to geographical data is simply an alternate description of the same argument. Lower-status whites are supposed to prefer order and a simplified world because of their marginal status, and one appealing way to achieve these psychological benefits is by categorizing people racially. The end result is the same: material deprivation makes racial conservatism more appealing. genuine contextual influence or competitive threat, this perspective will shape individual decisions later in life. In particular, racially prejudiced individuals will gravitate toward communities where their disliked group is in short supply, whereas those for whom race is less salient will pick communities based upon other criteria. The end result of this process would be similar to the contact hypothesis, in cross-sectional data: proximity will appear to produce lower racial conservatism. However, the result should be strongest where more whites are migrants, and could sort themselves out. 4. Traditional White Backlash Hypothesis (Politics)–Racial conservatism represents an effort by whites to maintain political control as an empowered majority (Key 1949). Polarization will surge as the AfricanAmerican population increases in density, with an exceptionally large expression of racial conservatism where blacks approach the point of deciding local elections. This pattern would be especially strong at a level with politically meaningful borders: counties, legislative districts, states. 5. White Colonizer Backlash Hypothesis (Economic)–Whites enjoy the fruits of anti-black economic discrimination when surrounded by a large, marginalized African-American population (Glenn 1963). They are able to construct a split labor market, and profit therefrom (Bonacich 1972, 1976). Proximity thus leads to greater racial conservatism. However, in locales where white and black resources become more similar, the labor market is undifferentiated, so whites face little incentive to engage in race-based collective action. Proximity will not matter when conditions are relatively equal. 6. Hard Times Backlash Hypothesis (Economic)–A retrenchment against civil-rights gains set in when national finances turned sour around 1968. Americans were happy to see racial progress while the economy was booming, but once job competition became a zero-sum game then repairing statistical disparities among racial groups required sacrificing particular whites. Lean prospects encourage whites to defend historical group privileges, especially when racial advantage would give them leverage against nearby blacks. Poor and uneducated individuals may compensate for material sacrifices by stressing status advantages attached to their group membership (Bettelheim and Janowitz 1964). Proximity thus leads to greater racial conservatism, the more so as economic conditions tighten (e.g., higher unemployment, lower income) or likelihood of a job threat becomes more severe (e.g., white and black socioeconomic resources become more similar). 7. Symbolic White Backlash Hypothesis (Cultural)–Whites struggle with the African-American population to define the nation’s social and political discourse. They fight to preserve a “way of life” that appears threatened (Wong and Strolovitch 1996). They also struggle to determine the value of “cultural capital,” or the social value placed on traits that vary across ethnic groups and across geographical locales (Merelman 1994, 3-4). Cultural struggle may revolve around what many commentators call “institutional racism,” or facially race-neutral judgments that nevertheless are unnecessary and have a racially disparate impact. Contrary to the “naive Symbolic Racism” approach, this perspective acknowledges that conflict over social values represents a genuine power struggle, and the threat does not strike all whites similarly. Proximity may produce greater racial conservatism in small-town or rural America, because it forces whites to struggle with a black population to define local social institutions. This is especially true because rural areas tend to resist change, and therefore will have preserved historical proximity patterns that municipalities have cast off (Berard 1998; Ford 1960; Orum and McCranie 1970, 170). However, the proximity effect should decline or even reverse in more urban locales, where whites vary in their attachment Voss The Rational Basis of Symbolic Racism 8 11 These norms should not be mistaken for a real class-based identity. Most Americans consider themselves “middle class,” so these values are primarily an associational or ideological quality. 12 Wong and Strolovitch (1996) assume that conflict over “way of life” is intimately linked to “a struggle for territory and the privilege associated with it.” With New York hate crimes, their supposition apparently proves correct. This is not necessarily surprising, though, because hate crimes require not just the racial hostility but also the victim, which is why lynching studies have struggled to capture the probability of two races running into each other (Reed 1972). The territorial link need not apply to all forms of cultural backlash, so I do not incorporate it here. 13 Obviously the county is not the only possible measure of aggregation that might be relevant. However, no smaller level of aggregation was available for this study. Furthermore, the county is a politically and economically relevant areal unit (Glaser 1994), whereas smaller (and therefore more porous) units such as the precinct do not possess any clear meaning. 14 A fair test of the traditional backlash hypothesis actually requires testing a parabolic model as well, which does not require new data, simply a squared version of the density variable. However, this term never achieved statistical significance in any of the tests reported below, and so does not appear in the discussion of results. to “white middle-class norms.” Those with the greatest cultural resources, who are more (sub)urbanized, will have the most to lose precisely if they reside in white enclaves where these norms are strongest. Testing each of these contextual implications requires linking aggregate data to the Kaiser survey, made possible because Chilton recorded the county in which each respondent resided.13 In each case, I will need a measure of black racial density, for which I use the black percentage of a county’s voting-age population. For the conventional hypotheses, numbers 1 and 4, this measure is enough; the first predicts a null coefficient on the measure (or any interaction using it) and the fourth predicts a positive coefficient regardless of interactions.14 The remainder require some sort of interaction term, using a second Census measure. The specific list of interaction terms used in this paper are as follows. • Hypotheses 2, 5 and 6: All require black density multiplied by a measure of status inequality. I tested several inequality measures in the analysis, but for purposes of reporting results I choose one, the difference between white and black per capita income, where a positive number means whites are wealthier. • Hypotheses 2 and 3: Both require black density multiplied by a measure of population instability. For the latter I use the percentage of the population that did not live in the county five years previously. • Hypothesis 6: This requires black density multiplied by a measure of white socioeconomic status, to permit lower-income white communities to exhibit a backlash stronger than found in wealthier locales. I used several distinct measures of white status, including education rates, but for purposes of reporting I use only white per capita income. • Hypothesis 7: The Cultural Backlash approach is most complicated. Surveys allow multiple, imperfect ways to get at “white middle-class” identification, but I do not wish to use explicit values questions that cannot be pulled apart from the social interests reflected in my dependent variables. The analysis requires more concrete proxies for interest or affiliation. Therefore I tested the idea from several angles, all of which required black density multiplied by some contextual measure of the likely prominence of “middle-class values,” the social investment in white cultural advantages (Merelman 1994, 2-5). Voss The Rational Basis of Symbolic Racism 9 15 While this attempt to demonize rhetorically those who disagree may seem excessive, scholars who use the label are certainly correct that racial conservatism “blames the victim,” and therefore implies a rather negative prejudicial judgment of African Americans as a group. First, I use the level of urbanization, since suburban isolation allows white control of cultural capital (Merelman 1994, 5). This test for suburban exceptionalism is a rather blunt measure, however, since counties can be large and diverse. Rather than simply use the county’s overall level of urbanization, therefore, I also tried a second approach based upon whether the respondent reported personally living in a metropolitan area. Finally, I tried a measure that would get past pure suburbanization, and look specifically at one component of white middle-class privilege: the percent of the labor force in professional occupations, since these draw heavily on “cultural capital.” These hypotheses are summarized in Figure 1 on the next page. Tacit Theories of Blame The socioeconomic status of African Americans is clearly lower than that of whites in the United States, a fact few dispute. Where people disagree is their assessment of why the races differ. Racial liberals attribute the gap to racism, discrimination, and inadequate government attempts at equalization. Racial conservatives emphasize faults within the black community. It is their desire to downplay systemic causes of inequality that some scholars dub symbolic “racism.”15 The Kaiser poll included a handful of questions asking why different ethnic groups lag in achievement, and therefore allows a formal exploration of whether these symbols are in fact unrelated to systematic interests. The wording on these questions tends to differ from that found in other surveys, and therefore offers a refreshing look at the issue. The survey includes two questions centered on black opportunity (MONEYBL and STDLIV). The first asks whether blacks have as much opportunity to be successful and wealthy. The second focuses on the African-American comparative standard of living. Tracing “blame” for economic results is an extremely tricky business; some polemics blame inner-city black culture while others point the finger at mainstream society. Where one falls out seems more a matter of faith than a matter of hard evidence, so I am not worried that this is actually a factual question with different correct answers based upon geography. This is a measure of ideology, of the sort scholars have chosen to call “symbolic racism.” I begin with the standard of living variable, since it was applied to the entire sample. It asked whether black Americans have “the same standard of living and opportunities as whites.” This is unfortunate wording, since it combines two judgments–a question of fact (“standard of living”) and one of potential (“opportunities”). However, I find no connection between actual black poverty in a county and the response that whites provide to the question, so presumably it evokes a broad social attitude rather than an empirical observation. It taps a judgment of what African Americans could accomplish on their own merits. The variable takes on two values: 1 if the respondent considered black opportunities worse, 0 Voss The Rational Basis of Symbolic Racism 10 Interactive Term Argument Proximity Number Hypothesis Polarization Motive for