INTRODUCTION
In the run up to the 2008 presidential election, pundits, academics, and others trumpeted the possible dawn of a “postracial” United States, citing Obama's success as both an example and a symbol of more to come. Others worried that Obama's success or failure might either mask hidden racist attitudes or undermine attempts to address racial inequality. Most of this discussion was based on speculation and only minimal empirical evidence.
We begin this article by proposing that although Obama ultimately won, we cannot reject that race—and in particular racism—played a significant role in the outcome. During the campaign, race was a consistent narrative used by Obama's opponents. His primary opponents, particularly Hillary Clinton, and Republicans in the general election used racial references to attack the Illinois senator, citing him for his perceived inability to connect to “real working Americans” (Bazinet and McAuliff, Reference Bazinet and McAuliff2008; Canellos Reference Canellos2008; MacGillis Reference MacGillis2008). A Republican in Georgia used the term uppity to describe Obama, a clear racial reference (Los Angeles Times 2008). Even the infamous “Joe the Plumber” charged Obama with seeking to redistribute wealth, raising age-old stereotypes of African Americans as radical, welfare dependent, and not as hardworking as the White working class. In short, he accused Obama of seeking to take money from hardworking “real Americans” to give it to “those people” (Rohter Reference Rohter2008).
In addition to race, we must consider the role of patriotism. After all, some individuals at campaign rallies, and even Clinton herself, suggested Obama was not a “real” American and questioned his patriotism. Some suggested that his middle name, Hussein, branded him a terrorist (Rich Reference Rich2008; Hosseini Reference Hosseini2008). The infamous July 2008 cover of the New Yorker that depicted him as a terrorist and Michelle Obama as a Black radical did him no favors. Obama, moreover, was pilloried by the political right, on at least one occasion, for declining to wear a flag pin and castigated for failing to salute the flag (Tichy Reference Tichy2009). His wife also drew fire from the right for feeling belated pride in her country, for stating she had only recently felt proud of the United States. Why had she not always taken pride in the country (Payne Reference Payne2008; Stanley Reference Stanley2008)? During the campaign, moreover, Republican standard-bearers often invoked references to patriotism among working-class voters as a means of identifying “us” from “them.” In fact, the blogs were afire, with one in five posts highlighting perceived differences between the Democratic nominee and the “real” Americans (Tichy Reference Tichy2009).Footnote 2
Alternatively, it's not at all clear that race or racism ultimately had anything to do with the outcome of the campaign. With a single exception, when he was forced to deal with the issue in March 2008 following the wake of Reverend Wright's comments, Obama insisted on steering clear of race, and for good reason: the injection of race almost always damages the Black candidate in a biracial contest (Becker and Heaton, Reference Becker and Heaton1967; Citrin et al., Reference Citrin, Green and Sears1990; Hajnal Reference Hajnal2007; Kinder and Sears, Reference Kinder and Sears1981; Reeves Reference Reeves1997; but see Colleau et al., Reference Colleau, Glynn, Lybrand, Merelman, Mohan and Wall1990). Obama, therefore, ran a deracialized campaign—that is, one in which he avoided overt appeals to racial issues. Doing so, according to the theory, permits a Black candidate in a biracial contest to capture more of the White electorate (McCormick and Jones, Reference McCormick, Jones and Persons1993).
This article explores the following questions: did racism affect the extent to which Whites supported Obama? If so, was it because the prospect of a Black president posed a perceived threat to White dominance? Or, given the symbolism associated with the presidency—that is, the way the president personifies the government and the nation, and is the preeminent political authority in the land—perhaps it was too much for Whites to tolerate a Black man in the office. Put differently, did Whites' habits of mind, in which many believed Blacks to be in violation of cherished U.S. values learned in childhood, allow them to countenance a Black man in the White House? Did patriotism have a bearing on whether Whites embraced Obama? What, if any, role did class play in the election?
To investigate these questions, we draw on two theories of racism to explore how race may have affected White support for Obama: laissez-faire racism and symbolic racism. In the case of the former, racism is believed to be a function of Whites' “sense of group position” vis-à-vis Blacks in society. We argue that a Black man in the White House may legitimize Blacks as a group, posing a threat to Whites' group dominance. The latter approach suggests that racism is fueled by the combination of dislike for Blacks and a belief that Blacks fail to observe cherished U.S. values. We believe both factors may have played a role and can be used to make sense of the campaign rhetoric. They can also provide a general view for thinking about the relationship between race and (non)support for Obama. Some may think it odd that we include patriotism alongside racism, but we do so for good reason. First, love of country was a mainstay of the campaign, as seen in questions over Obama's or his wife's patriotism, and Republicans' attempts to differentiate “real” Americans from implied “fake” Americans. Second, for reasons we detail below, patriotism often behaves much like racism.
Drawing on a fresh set of data, collected on the eve of the 2008 election, we find both patriotism and symbolic racism dampened support for Obama on the eve of the election. We also find that the effect of racism and patriotism on support for Obama was on some occasions conditioned by class. This article promises at least two interventions. First, we examine the effect of race in a biracial contest at the national level and for the highest office in the land. As we have indicated, the question of whether race affects candidate favorability and ultimately political choice in biracial elections isn't a new one. In most cases, actual and experimental, racism dampens support for the Black candidate. However, with the exception of David Sears and colleagues' work (Reference Sears, Citrin, Kosterman, Moreland, Steed and Baker1987), these evaluations are based upon subnational outcomes, contests confined to state- and local-level politics. Because the president is the personification of the government domestically, the face of the United States to foreign audiences, and the figure of authority for the country, one cannot overestimate the symbolic importance of the office (Binkley Reference Binkley1952; Easton and Dennis, Reference Easton and Dennis1969; Greenstein Reference Greenstein1960). Obama's presidential bid, in other words, was a world apart from the mayoral, gubernatorial, and even senatorial contests in which Black and White candidates have faced off.
A second intervention we offer, and a new wrinkle we add to the patriotism literature, is the extent to which patriotism affected candidate favorability and political choice. Others have explored circumstances under which patriotism influenced political engagement and activism (Huddy and Khatib, Reference Huddy and Khatib2007; Schatz et al., Reference Schatz, Staub and Lavine1999). John Sullivan and colleagues (Reference Sullivan, Fried and Dietz1992), in the only other study that examines the effect of patriotism on presidential choice, explored how patriotism affected the outcome of the 1988 presidential campaign. However, Republicans in that case didn't question whether Governor Dukakis was indeed a “real” American as they did with Obama. Rather, they simply questioned his political views. The article opens with a brief review of the relevant theoretical approaches, after which we specify the hypotheses we plan to test. We then introduce the data, proceed to hypothesis testing, and discuss the results. We close the article with a brief discussion of the implications of our findings.
RACISM AND PATRIOTISM
There are at least two theories of racism, as well as one theory of patriotism, that suggest race may have affected evaluations of Obama and whether one ultimately cast a vote for him. The first theoretical approach indicates that racism is a significant factor in politics that not only explains attitudes toward Blacks but also structures other kinds of political behavior, such as partisanship and preferences for various policies such as welfare. The second theory suggests racism is motivated by group identities and competition over resources. The first has its roots in social psychology, the second in sociology. Symbolic patriotism, in which love of country is expressed through attachment to national symbols, is, we think, a third theory of racism.Footnote 3 All three should help us understand some of the campaign rhetoric, as well as to provide a general view to think about the relationship between race and (non)support for Obama.
We begin with the more social psychological approach. Some scholars have argued that old-fashioned racism—in which Whites viewed Blacks as biologically inferior and, therefore, worthy of subordination—has been replaced by a more modern form of racism. Symbolic racism, as it has come to be known, suggests that rather than overtly arguing for the inferiority of Blacks in the United States, Whites claim that Blacks violate cherished U.S. values (Kinder and Sanders, Reference Kinder and Sanders1996; McConahay Reference McConahay1982; Sears Reference Sears, Wheeler and Shaver1983). Unwillingness to work, lack of thrift, criminality, and welfare are code words for race, used to express racist sentiments.Footnote 4 Imagine the president, a figure of unparalleled authority in the United States, an image inculcated since childhood (Easton and Dennis, Reference Easton and Dennis1969; Greenstein Reference Greenstein1960), is now a Black man. Under these circumstances, we think it is possible that Obama, as a Black man and symbolic leader of Black people, will lose support from Whites. Whites may see him as a symbol of all Blacks and therefore think him unworthy of the office, ultimately rejecting him.
Hypothesis 1: In this account, racism matters, and Whites will fail to embrace Obama because they will view him as undeserving of the office. As a result, they'll resent him and ultimately fail to support him.
Another approach to racism, one that shifts the focus from the individual to the group, emphasizes threat and entitlement. Laissez-faire racism, according to Lawrence Bobo and colleagues (Reference Bobo, Kluegel, Smith, Tuch and Martin1997), also follows the decline of what they term Jim Crow or old-fashioned racism. As the economic and political power of southern planters declined, U.S. cultural views became more progressive, and as the Black community continued to assert itself, Jim Crow racism ceased to function as a valid ideology. It was, according to Bobo and his colleagues (Reference Bobo, Kluegel, Smith, Tuch and Martin1997), supplanted by a form of racism fueled by perceived threats to what Whites believed to be their rightful possession of greater resources and social status. As a means of justifying their “sense of group position,” this approach indicates, Whites ascribe negative stereotypes to Blacks. In other words, Whites have “a tendency to blame Blacks, themselves, for the Black-White gap in social economic status” (Bobo et al., Reference Bobo, Kluegel, Smith, Tuch and Martin1997, p. 16). Whites, then, may continue to justify their dominant group position by subscribing to stereotypes in which they perceive Blacks as lazy, violent, and unintelligent (Bobo and Kluegel, Reference Bobo, Kluegel, Tuch and Martin1997). Our contention is that on the eve of the election, President Obama posed a threat to Whites' group position—both materially and in terms of social status.
Hypothesis 2: With an emphasis on power and perceptions of ingroup-outgroup difference, this theory suggests that Whites will express fear of a Black presidency because it would represent some decline in power, prestige, or material circumstances for their group.
There is yet another approach to racism, one that connects national pride to racial antipathy: patriotism. For the uninitiated, it may seem strange to compare patriotism to racism. Patriots—especially U.S. patriots—are supposed to root for freedom and equality. After all, these are among the foundational values on which the country was established (Berns Reference Berns2002; Dietz Reference Dietz and Primoratz2002). So, one may ask, to borrow a turn of phrase from Paul Gomberg (Reference Gomberg1990), how is “patriotism like racism?” Here's how: insistence upon universal freedom and equality, for which democracy is known, is restricted to conationals (Viroli Reference Viroli1995). In the United States, this means, of course, those who are perceived as American. That, in turn, raises the following question: who is considered American? Recent scholarship has suggested that Whites are more closely associated with the U.S. ideal than any other racial group (Devos and Banaji, Reference Devos and Banaji2005).
Since they are perceived as prototypical Americans, Whites represent the baseline from which other racial groups presumptively depart. Thus, it should come as no great shock that Whites appropriate symbols of U.S. patriotism as a means of perpetuating domination. They do so because the public institutions responsible for the maintenance of White domination are represented by national symbols (Sidanius and Petrocik, Reference Sidanius, Petrocik, Asmore, Jussim and Wilder2001). Indeed, symbols such as the flag and the Constitution, at least in the United States, are considered representations of the political community (Baas Reference Baas1979; Elder and Cobb, Reference Elder and Cobb1983). Perhaps this is why Whites tend to identify more with, say, the flag than with subordinate groups (Sidanius and Pratto, Reference Sidanius and Pratto1999). For these reasons, Social Dominance Theory suggests that patriotism, and the symbols of the nation associated with it are commensurate with negative feelings toward the subordinate group (Sidanius and Pratto, Reference Sidanius and Pratto1999). Put differently, “Love of country … implies the endorsement of [the country's] hierarchical ethos, at least among dominant groups” (Sidanius and Pratto, Reference Sidanius and Pratto1999, p. 240), something that transforms patriotism into racism.
Hypothesis 3: Since patriotism is used as a means to justify their position in the social hierarchy, Whites will reject Obama because he isn't perceived as part of the dominant group.
Another pair of hypotheses revolves around the prospect that the effects of patriotism and racism are contingent upon class, that the working-class Whites tend to cling to their social position at least in part through racism. Historically, they attempted to extract psychological benefits from their racial group membership when Blacks were perceived to threaten their material superiority (Roediger Reference Roediger1991), something that continued from the 1960s to the 1980s (Durr Reference Durr2003). Furthermore, as Monica McDermott (Reference McDermott2006) has illustrated, working-class Whites' antipathy toward African Americans may now be driven by other factors. The White working class may come to resent Blacks, for instance, if they are forced to reside in close proximity to majority Black areas. According to McDermott, shrinking residential distance suggests failure on the part of working-class Whites. Other times, working-class Whiteness is rooted in ethnic pride in which power and privilege come with the territory (McDermott Reference McDermott2006). This situation suggests that a sense of group position may lead to racism (Bobo and Hutchings, Reference Bobo and Hutchings1996; Bobo Reference Bobo1999). Of course, this implies that more educated Whites will feel less threatened and more tolerant (Bobo and Licari, Reference Bobo and Licari1989; but see Jackman Reference Jackman1978). For these reasons, we believe that the effect of racism is contingent upon class. Likewise, because working-class Whites generally have a narrow vision of who is or isn't American (Gerstle Reference Gerstle2002), we think it likely that the effect of patriotism on support for Obama is also moderated by class.
Hypothesis 4: Because the White working class tends toward insecurity vis-à-vis Blacks, we believe the effect of racism on support for Obama will be conditioned by class status among Whites.
Hypothesis 5: Given working-class Whites' narrow conception of U.S. identity, we believe the effect of patriotism on support for Obama is moderated by class status.
There are two perspectives that suggest race may not have played a role in the 2008 campaign. One suggests that by not emphasizing racial appeals, Black candidates can run a “deracialized” campaign. Thus, by using universal rhetoric and not appealing directly to racial solidarity among Blacks, Black candidates can avoid being branded as racial candidates and thus appeal to White voters (McCormick and Jones, Reference McCormick, Jones and Persons1993). The perspective suggests that the universal rhetoric of Obama should have inoculated him from racist attacks, and therefore race should not have played a significant role in the campaign. The other perspective, proposed by Paul Sniderman and his colleagues, suggests that conservative concerns about Blacks and opposition to affirmative action are driven by “principled conservatism” (Sniderman and Hagen, Reference Sniderman and Hagen1985; Sniderman and Piazza, Reference Sniderman and Piazza1993; Sniderman et al., Reference Sniderman, Tetlock and Carmines1993; Sniderman Reference Sniderman1996; Sniderman and Carmines, Reference Sniderman and Carmines1997). Unlike proponents of symbolic racism, Sniderman and others have suggested that programs such as welfare and affirmative action have created a perception that African Americans violate conservative norms of hard work and meritocracy. Thus, conservative preferences, rather than anti-Black racism, for nonintervention in economic and social life drive their feelings toward these policies and candidates, and ultimately Blacks. Since Obama made no broad appeal on issues such as affirmative action, welfare, or other explicitly racial issues, race might not have played a role in his victory.
Hypothesis 6: To the extent that Obama largely avoided explicit appeals to race, we think it possible that the role of race and racism in the outcome will have been minimal.
DATA
The data used in the ensuing analysis were from the Washington Poll, a nonpartisan, academic survey research project sponsored by the Washington Institute for the Study of Ethnicity, Race and Sexuality (WISER), a research center at the University of Washington in the School of Social Sciences. The survey was administered by telephone, by Pacific Market Research, in Renton, Washington, using randomly selected phone numbers from a list of registered voters, including an oversample of African Americans. However, for our immediate purposes, we use only the White sample. The survey was in the field from October 19, 2008 to November 6, 2008, and surveyed 1203 respondents, 891 of whom were White, from within the state of Washington (see Table 1).
Table 1. Selected Sociodemographic and Political Variables
Source: Washington Poll (2008)
A quick overview of the data revealed the following: 52% of the respondents were male, and 49% had college degrees. Additionally, 47% of the respondents resided in the Puget Sound region, and 31% reported household incomes of $60,000–$99,999; 34% of the respondents were Democrats, 29% Republicans, and 37% identified as Independent or other. The majority of respondents (39%) identified themselves as conservative. The respondents' ages were distributed as follows: 8% between the ages of eighteen and thirty-four, 17% between thirty-five and forty-nine, 40% between fifty and sixty-five, and 35% of the respondents age sixty-six or older.Footnote 5
RESULTS
What was the extent to which White Washingtonians supported Obama? As Figures 1–3 make clear, overall, we found broad-based support for Obama. Nevertheless, when moving from what his candidacy meant to what one's intention was to vote for him, support varied.
Fig. 1. Obama's Candidacy Makes Me Proud to Be an American
Fig. 2. Obama's Favorability
Fig. 3. Presidential Vote Intention, Obama vs. McCain
We first questioned respondents about whether Obama's candidacy made them “proud to be an American.” The response was captured by a four-point Likert-scaled item ranging from strongly agree to strongly disagree.Footnote 6 Almost three-quarters (72%) of Washington's White electorate answered that Obama's candidacy for the highest office in the land made them feel proud to be Americans. We used a similar question to ask respondents about favorability—that is, how one feels about the candidate, not how the candidate makes one feel about the country. We told respondents, “I'd like to ask you about some people who have been mentioned in the news lately. Please tell me whether you have heard of the person, and if your impression is very favorable, somewhat favorable, somewhat unfavorable, or very unfavorable.” When we shifted from asking about pride in the country to asking about the favorable impression of the candidate, Obama's support among respondents dipped a bit, to just below two-thirds (64%); of course this means that approximately 36% of respondents failed to have a positive opinion of him. Support for Obama declined more, still, when respondents were asked whether they would cast a vote for him. At this point, his support slipped to 53%, with 47% declining to vote for him.
How did patriotism and racism affect support for the president? Is it possible that either or both might have accounted for some of the observed disparity in support? We used a three-item index to measure symbolic patriotism, the extent to which one feels attached to the nation by way of its symbols. Symbolic racism is represented here by a three-item index. Laissez-faire racism is represented by a two-item index of racial stereotypes. (Please see the Appendix for coding and question wording.) Judging by the preliminary results in Figures 4–6, we found symbolic and laissez-faire racism and patriotism did affect support for Obama.
Fig. 4. Obama's Candidacy Make Me Proud to Be an American
Fig. 5. Obama's Favorability
Fig. 6. Intention to Vote for Obama
We began by examining the extent to which Obama's candidacy permitted one to take pride in their American identity. Patriotism, it seemed, produced more consensus than did symbolic racism and laissez-faire racism. Roughly eight percentage points (77% to 69%) separated strong patriots in their pride in Obama's candidacy from those who were less attached to the nation (χ2 = 46.8; p < 0.01). Negative attributes of Blacks widened the gap between those who took pride in Obama's candidacy and those who did not. Only 48% of those who attributed negative stereotypes to Blacks were proud of his candidacy versus 77% of those who saw Blacks more favorably (χ2 = 28.2; p < 0.01). Variation in symbolic racism produced similar results: 77% of those who were not resentful of Blacks took more pride in their country as a result of his candidacy versus 46% of those who were resentful (χ2 = 31.0; p < 0.01).
As we moved to examine bivariate results for Obama's favorability, we saw comparable differences between those who were strong patriots and had negative images of Blacks and those who were less attached to the United States and had a favorable image of Blacks. For instance, only 54% of strong patriots had a positive impression of Obama on the eve of the election versus 71% of those less attached to the nation (χ2 = 25.4; p < 0.01). We obtained similar results among the racially intolerant: 57% of those who subscribed to negative stereotypes of Blacks viewed Obama favorably versus 67% of those who held relatively positive views of Blacks (χ2 = 22.3; p < 0.01). The difference between those who believed Blacks should try harder and Whites less inclined to think in the same way was even greater. Only 32% of strong symbolic racists had a favorable impression of Obama; this was in contrast to 70% of those who failed to share their belief (χ2 = 72.7; p < 0.01).
So far, we have considered how variations in patriotism and racism promoted pride in Obama's candidacy and the extent to which each contributed to his favorability rating. In both cases, as we progressed from one indicator of support to the next, variations in patriotism and racism produced increasing divergence. This pattern was repeated once White Washingtonians were queried about their commitment to vote for Obama. For instance, only 41% of strong patriots intended to vote for the president versus 60% of those who weren't as attached to the nation (χ2 = 28.3; p < 0.01). Differences between those who held negative stereotypes of Blacks and those who did not were even larger, where only 29% of the former group, but 56% of the latter, planned to vote for Obama (χ2 = 26.5; p < 0.01). Finally, among the Washingtonians who doubted the earnestness of Blacks' efforts to succeed, only 19% intended to vote for Obama versus 59% of those who disagreed with that sentiment and believed that Blacks are really trying (χ2 = 65.8; p < 0.01).
Up to this point, our data support Hypotheses 1–3, which predict racism and patriotism will affect overall support for Obama on the eve of the election. With the exception of his candidacy's ability to inspire national pride, patriotism behaved very similarly to both types of racism. To be sure, the gaps separating strong and weak patriots were narrower than, for instance, the observed differences between symbolic racists and their counterparts. In spite of the smaller differences that accrue to patriotism, the results for strong patriots mirror the results for Whites who perceive Blacks as encroaching on their group's position and those who believe that Blacks simply refuse to try hard enough—all failed to embrace Obama. Did racism and patriotism serve divisive purposes during the campaign, affecting support for the Democratic nominee and evaluations of him in the same way? To fully assess the veracity of this claim, we had to move beyond simple cross-tabulations, for we also needed to account for the effects of other factors that might have affected support for Obama, including the effect of class and principled conservatism, Hypotheses 4–6.
Models of Support for Obama
Modeling support for Obama required us, first, to control for key sociodemographic factors. These included age, gender, income, and education. We also had to account for the effects of manifest political factors, at least two of which were especially relevant in this case. Since the president is viewed as the face of the government domestically and the embodiment of the United States to foreigners, a Black man in the Oval Office represents monumental change. As such, we believed that the image of a Black man guiding the United States would have activated one's partisan predispositions as well as engaged one's ideological proclivities (Citrin et al., Reference Citrin, Green and Sears1990; Sears Reference Sears, Iyengar and McGuire1993; Sears et al., Reference Sears, Citrin, Kosterman, Moreland, Steed and Baker1987; Sniderman and Piazza, Reference Sniderman and Piazza1993). We measured this with dummy variables where Republicans served as the reference group for Democrats and those who claimed to be political Independents. For ideology, we used a seven-point self-placement item in which conservative = 1.
Economic issues, at both the personal and national level, have often informed political choice on Election Day (Kinder and Kiewiet, Reference Kinder and Kiewiet1979; Kiewiet Reference Kiewiet1983; Lewis-Beck et al., Reference Lewis-Beck2008). Given the economic crisis, these issues may have been especially relevant in the most recent election cycle. Hence, we also had to account for economic anxiety. To do so, we used a three-item index in which economic anxiety = 1. (Please see the Appendix for coding and question wording.) We estimated two models to fully capture support for Obama. The first model estimated the effect of patriotism and racism, controlling for key sociodemographic, political, and economic factors on support for Obama, where:Footnote 7
Y = level of support for Obama
b 0 = intercept
b 1 = level of patriotism
b 2 = level of symbolic racism
b 3 = level of laissez-faire racism
b 4–7 = sociodemographic controls (age, income, level of formal education, gender)
b 8–11 = political controls (Democrat, Independent, Conservative, and level of economic anxiety)
A second model tested the hypothesis that the effects of patriotism and racism on support for Obama will be conditional on class membership. We did so, in chief, because much has been made of White working-class antipathy toward Obama during both the primary and the general election campaigns. For this reason, we specified a product term in the second model in which education was a crude proxy for class.Footnote 8 Thus, the following was added to the prior equation, where:
b 12 = the effect of patriotism moderated by a respondent's level of formal education
b 13 = the effect of symbolic racism moderated by a respondent's level of formal education
b 14 = the effect of laissez-faire racism moderated by a respondent's level of formal education
We first examined how patriotism and racism contributed to the meaning of Obama's candidacy. More specifically, what accounted for the ways in which his candidacy affected how people felt about the United States? We expected racism to do so, but what about patriotism? Table 2 reports the results.
Table 2. Determinants for American Pride in Obama's Candidacy
Note: Cell entries are ordered logistic regression coefficients. Coded such that 1 = pro-Obama.
* p < 0.05 = one-tailed test.
To our surprise, patriotism increased the probability of Whites who took pride in Obama's candidacy by 17%. Likewise, Democrats and those who claimed Independent status increased the probability of his nomination making them feel good about being an American by 33% and 13%, respectively, compared with Republican partisans. Differences emerged, however, upon considering the effect of symbolic racism, where symbolic racists were 33% less likely than others to take pride in Obama's candidacy. In the second model, the one in which we hypothesized that class will moderate the effect of patriotism and racism on support for Obama, the results were telling. For those who were part of the White working class, increasing symbolic racism as well as patriotism reduced the probability of feeling proud to be an American because of Obama's candidacy by 11% and 10%, respectively (the effect of patriotism is insignificant). The results for White Washingtonians who were middle-class or higher were a bit more complicated. Patriotism and racism with this group operated quite differently vis-à-vis the working class. One difference was that the effect of symbolic racism was even greater among this group, dampening support for the candidate by an additional 33%. But we observed something different when the focus shifted to patriotism. In this case, the probability of feeling good about Obama's candidacy on the eve of the election increased by 25% beyond that which was felt by working-class Whites, for whom there was no effect for patriotism. See Figures 7 and 8.
Fig. 7. Change in Likelihood of Pride Due to Obama's Candidacy
Fig. 8. Predicted Change in Pride Due to Candidacy
Next, we explored Obama's preelection approval. This was a fundamentally different question than the historical significance of Obama's candidacy because it required respondents to assess the candidates on their own terms, not according to how it made them feel about the country. Did patriotism and racism affect the extent to which Obama achieved a favorable impression among Washington's White electorate? Table 3 reports the results.
Table 3. Determinants of Obama's Favorability
Note. Cell entries are ordered logistic regression coefficients, coded such that 1 = pro-Obama.
* p < 0.05 = one-tailed test.
Indeed, there was an inverse relationship between symbolic racism and favorable impressions of Obama which dampened his favorability by 15%. Beyond that, variation in neither laissez-faire racism nor patriotism affected support for the Democratic nominee. For political factors, it was clear that Democratic partisans and Independents were both significantly more likely than Republican identifiers to have a good impression of Obama, by a margin of 46% and 26%, respectively. Economic anxiety also contributed to how people felt about Obama, insofar as it increased the probability that they would have a more favorable impression of him, by 18%, compared to those not concerned about the economy. Political conservatives, on the other hand, were 48% less likely than avowed liberals to embrace him. In the second model, patriotism was the only determinant of substantive interest to have been affected by class differences, albeit in opposite directions. For working-class Whites, increasing patriotism diluted support for Obama by 34%. In contrast, among the middle-classes, patriotism increased the chance that they would see the first Black president in a favorable light by 55%, relative to their counterparts in the working class. See Figures 9 and 10.
Fig. 9. Change in Likelihood of Approval for Obama
Fig. 10. Predicted Change in Approval for Obama Due to High Levels of Patriotism
Did patriotism and racism affect Whites' willingness to vote for Obama? Table 4 indicates they both did. For instance, when holding the other variables in the model at their respective means, patriotism, symbolic racism, and laissez-faire racism all decreased the likelihood of casting a vote for Obama by 35%, 40%, and 37%, respectively, as did political conservatism. As one might suspect, identifying as a Democrat increased the likelihood of voting for Obama, in this case by 70%, compared to identifying as Republican. Identifying as Independent also wound up benefiting Obama: 32% were more likely to vote for the Democratic standard-bearer than for Republican partisans. Further, as economic anxiety increased, so too did the probability of voting for Obama by 40% (see Figure 11). The second model revealed a different pattern, one in which class fails to moderate the effect of patriotism. Substantively this means, according to Figure 12, that strong, working-class patriots were 47% less likely to have cast a vote for Obama than members of the White working class with a relatively weak attachment to the country. The same pattern emerged for laissez-faire racism. More to the point, those in the White working class who subscribed to negative racial stereotypes of Blacks were 38% less likely to support Obama at the polls than those who refused to accept stereotypes. In the absence of significant interactions in the second model, we can only conclude that middle-class and working-class White Washingtonians were like-minded in that increasing patriotism and laissez-faire racism dampened the likelihood of voting for Obama.
Table 4. Determinants for Intention to Vote for Obama
Note. Cell entries are logistic regression coefficients, coded such that 1 = pro-Obama. Diagnostics reveal no multicollinearity.
* p < 0.05 = one-tailed test.
Fig. 11. Change in Likelihood of Vote for Obama
Fig. 12. Predicted Change to Vote for Obama among Working-Class Whites
DISCUSSION
Broadly speaking, patriotism and racism, as anticipated, affected support for Obama on the eve of the election. Upon a more discriminating examination, however, it is clear that some of the hypotheses failed to hit their respective marks. Support for laissez-faire racism was rather limited as were hypotheses in which class was predicted to moderate the effects of racism. Symbolic racism performed well across the board. And, though it failed to always yield results in the hypothesized direction, it's hard to question the predictive power of patriotism, including the ways in which its effects were often contingent upon class. This suggests that the hypothesis that race did not matter in the campaign can no longer be taken seriously, in light of the aforementioned evidence.
Patriotism, for the most part, performed as expected. For instance, it failed to permit Whites attached to U.S. national symbols, those who claimed to love the country, the latitude to vote for the first Black president. These findings agree with Jim Sidanius and colleagues' work (Reference Sidanius and Pratto1999; Reference Sidanius, Petrocik, Asmore, Jussim and Wilder2001), among others, which indicates that symbolic patriotism dampens support for racial progress. Yet, these findings also depart from the literature in at least two important instances. Strong symbolic patriots were proud that a Black man had secured the nomination of a major party. In a similar vein, Whites who were objectively middle-class or higher were significantly more likely than their working-class brethren to both embrace the historic nature of his candidacy and have a favorable view of Obama. Perhaps it's the case that those who've benefited from a better education are also likely to have a greater appreciation for the values represented by the flag and national anthem. In any case, the latter findings comport with other work which indicate increasing patriotism doesn't necessarily lead to the derogation of outgroups (Brewer Reference Brewer1999; Citrin et al., Reference Citrin, Wong, Duff, Ashmore, Jussim and Wilder2001; De Figueiredo and Elkins, Reference De Figueiredo and Elkins2003; Li and Brewer, Reference Li and Brewer2004). These findings are also in concert with Christopher Parker's (Reference Parkerforthcoming) work in which he concludes that strong symbolic patriots embraced some racial and religious minorities.
The findings are also telling as they relate to racism. Laissez-faire racism, as it is represented by negative stereotypes of Blacks, failed to gain much traction. One explanation for this lies in our measurement: we were able to draw on only two of the typically five items (Bobo and Kluegel, Reference Bobo, Kluegel, Tuch and Martin1997).Footnote 9 Another, more likely explanation is simply that Obama defies the stereotypes often attributed to Blacks. After all, how could a man who has two Ivy League degrees conform to less than flattering racial stereotypes? It's also quite possible that Whites are more worried about their absolute, individual circumstances instead of their relative, group-based position. For, as the results indicate, the more economically anxious a respondent was, the more the respondent embraced Obama.
The results indicate a different story for symbolic racism, something that often militated against support for Obama. We contend that this result has much to do with symbolism. We reason that a Black president, as a political symbol (Edelman Reference Edelman1985), is too much for avowed racists to handle. The president personifies the United States abroad; is the face of the national government; and is perceived, beginning in childhood, as the chief political authority (Binkley Reference Binkley1952; Easton and Dennis, Reference Easton and Dennis1969; Greenstein Reference Greenstein1960). Perhaps it's the large symbolic load borne by the presidency that disallows those afflicted with racism, including those in the upper class, to reject the first Black president. We believe that Obama's background, which includes his spending part of his life in Indonesia, spurs some to see him as a foreigner (Tichy Reference Tichy2009), compounding the effect of symbolic racism. This squares with scholarship in which symbolic racism pushes Whites to reject perceived foreigners (Sears et al., Reference Sears, Citrin, Cheleden, van Laar, Prentice and Miller1999). This may also explain why white-collar Whites were less likely than blue-collar workers to take pride in Obama's candidacy. Maybe it's the case that this strata would have embraced the historical nature of his candidacy more had he a “traditional” African American background, someone with more organic ties to the Black struggle.
CONCLUSION
This article helps us understand how racism and patriotism shaped support or nonsupport for Obama. We can thus perhaps move beyond a discussion of whether racism matters, in some on/off fashion, to what effect racism had on voters relative to other factors that might have determined voting. We also assessed the extent to which patriotism affected evaluations of Obama, and whether it helped or hindered his quest for the presidency. Many pundits have concluded that since Obama won, we have now moved to a “postracial United States” or defeated racism, but they do not consider that some nontrivial segment of the broader public might have been influenced by racism when choosing their candidate.
The evidence clearly shows we have not moved on to a postracial era. In the 2008 campaign, the one in which Americans elected the first self-identified African American president, both race and racism played critical roles. We can further conclude, though not decisively, that some nontrivial portion of the White population used racist logic in evaluating the African American candidate, sometimes using narrow notions of patriotism and national identity to do so. Blacks and perhaps other groups such as Latinos and Asians still bear the burden of being considered by many Whites as either inferior or aliens, regardless of their accomplishments or the rhetoric of U.S. exceptionalism, which promises that believing in the civic ideals of the United States qualifies one for membership in the political community (Fuchs Reference Fuchs1990; Lipset Reference Lipset1996).
Having said this, we cannot help but think that, as in Cuba after the revolution, the interface between race, racism, and patriotism may change. It is entirely possible that while the campaign itself had lots of discourse on “old” and “new” racism, the symbolic value of an African American president may change notions of racial identity. However, as in Cuba and other societies, a more expansive notion of racial identity will not necessarily mean that racial prejudice will be any less salient for substantial portions of the population. Though this is speculation, we can reject out of hand the idea that President Obama's election means racism is either dead or an irrelevant factor in U.S. society, in general, and electoral politics, in particular. However, we also must recognize the complicated and powerful racial crosscurrents that will require us to sharpen our conceptual analyses of and modes of thinking about the roles that race and racism play and will play in politics.
It is worth noting, for instance, that the effect of race is not unambiguously bad. True, symbolic racism represented the view of roughly 42% of White Washingtonians. This means, of course, that 58% of White Washingtonians did not subscribe to such a view, suggesting that these people supported Obama. Thus, this portion of Whites, those who did not put much stock in the belief that Blacks did not try hard enough, felt pride in their American identity because of Obama's candidacy. Likewise, 54% of Whites did not accept negative stereotypes often associated with Blacks. We can, therefore, conclude that this group, Washington Whites who rejected laissez-faire racism, likely voted for Obama.
While our findings shed light on some of the relevant factors that shaped support for Obama on the eve of the election, they should be evaluated with the following in mind: almost half of our sample, drawn from the voting rolls of the state of Washington, earned college degrees, a figure at variance with the national levels (U.S. Census Bureau 2009). This is a valid concern, for it may bias our results. Yet, if anything, these results underestimate the effects of racism and patriotism on the 2008 campaign to the extent that education tends to undercut racism of all types (Bobo and Kluegel, Reference Bobo, Kluegel, Tuch and Martin1997; Hughes Reference Hughes, Tuch and Martin1997), as well as symbolic patriotism (Davis Reference Davis2007). Hence, there is the very real possibility that racism and patriotism had more to do with the outcome than is revealed in this article.
APPENDIX
Study Description, Question Wording, and Coding
The data collected is from the Washington Poll, a nonpartisan, academic survey research project sponsored by the Washington Institute for the Study of Ethnicity, Race and Sexuality (WISER), a research center at the University of Washington in the School of Social Sciences. The survey was administered by telephone by Pacific Market Research, in Renton, Washington, using randomly selected phone numbers from a list of registered voters. The survey was in the field from October 19, 2008 to November 6, 2008, and surveyed 1203 respondents from within the state of Washington. The response rates are, according to AAPOR RR4, 19.3 at a statewide level and 23.3 for the African American oversample (which was the same as the RR3). The high response rate on the oversample seems surprising, but it is because a decent number of voters dialed in the African American sample were non-African Americans and, therefore, not included in the sample or the response rate. The cooperation rates (COOP3) are, at a statewide level, 94.3, with a COOP3 for African Americans of 87.5.
Dependent Variables
Presidential Vote Choice
Vote choice was measured by respondent's presidential vote: “If the election for president were held today, for whom would you vote?” Responses were coded 0 or 1, where 1 was either vote for Obama or vote for McCain, depending on which dummy variable (vote Obama or vote McCain) was being used.
Barack Obama Favorability
Favorability was measured by respondent's impression of Barack Obama: “Do you have a very favorable/very unfavorable impression of Barack Obama?” Favorability was coded on a four-point 0–1 scale, where very unfavorable = 0 and very favorable = 1.
Pride in Obama's Candidacy
A respondents pride in Barack Obama's candidacy was measured by the following statement: “Barack Obama's candidacy makes me proud to be an American.” The response options were strongly disagree/strongly agree, and were coded on a four-point 0–1 scale where strongly agree = 1.
Predictors
Economic Anxiety
Economic anxiety is indexed by the following items, and was transformed into an average score scaled from 0 to 1 (reliability: α = 0.63; pro-anxiety = 1).
1. “Now, thinking about your own economic situation. Some people are very anxious about their own economic situation while other people are not anxious at all. How anxious are you about your own economic situation—not at all anxious, a little anxious, somewhat anxious, or very anxious?”
2. “Now, NOT thinking about your own economic situation, but rather about the country's economy as a whole. Some people are very anxious about the economic situation in the country while other people are not anxious at all. How anxious are you about the economy—not at all anxious, a little anxious, somewhat anxious, or very anxious?”
3. “When you hear or read about the economy in the news, do you become less or more anxious about the economy?”
Symbolic Racism
Racism was indexed by the following items, and was transformed into an average score scaled from 0 to 1. Each item was coded on a four-point 0–1 interval from strongly disagree to strongly agree, with the exception of Item 3, which ranged from 0 to 1 where strongly agree = 0 and strongly disagree = 1 (reliability: α = 0.49; pro-racism = 1).
1. “While equal opportunity for Blacks and minorities to succeed is important, it's not really the government's job to guarantee it.”
2. “If Blacks would only try harder, they would be just as well off as Whites.”
3. “A history of slavery and being discriminated against has created conditions that make it difficult for Black people to work their way up.”
Laissez-faire Racism
This form of group-based racism was indexed by the following items, and was transformed into an average score scaled from 0 to1. Each item was coded on a four-point 0–1 interval from strongly disagree to strongly agree = 1 (reliability: α = 0.43; pro-racism = 1).
1. “Blacks are more violent than Whites.”
2. “Most Blacks who are on welfare programs could get a job if they really tried.”
Patriotism
The patriotism index is the average score of respondents who answered the following questions. Each item was coded on a four point 0–1 interval from strongly disagree to strongly agree (reliability: α = 0.77; pro-patriotism = 1).
1. “How good does it make you feel when you see the American flag flying—extremely good, very good, somewhat good, or not very good?”
2. “Do you consider your love of America to be strong or not so strong?”
3. “How proud do you feel when you hear the national anthem—extremely proud, very proud, somewhat proud, or not very proud?”
Education
These questions determined the respondent's highest level of education: “What is the highest level of education you completed?” Recoded as a dummy variable where college degree = 1 and else = 0.
Gender
This question determined the respondent's gender. It was coded either 0 or 1, where male = 1.
Age
This question asked for actual age in years, compressed into a four-point scale (0–1), where oldest cohort = 1.
Income
This question measured the household income of the respondent: “What was your total combined household income in 2007 before taxes.” Income was coded on a four-point scale (0–1), where lowest level (less than $40,000) = 0 and highest level of income ($100,000–$150,000) = 1.
Party ID
This question determined the respondent's party identification: “Generally speaking, do you think of yourself as a Democrat, a Republican, an independent, or what?” Coded into three different variables on a two-point scale (0–1), where for each dummy variable Democrat = 1, Republican = 0, and Independent/other = 1.
Ideology
This question determined the respondent's political ideology: “When it comes to politics, do you usually think of yourself as a liberal, a conservative, a moderate, or haven't you thought much about this?” Conservative = 1.
