Introduction

The affordable care act, nicknamed “Obamacare,” has been the subject of hostile and relentless criticism, including accusations of socialism and the establishment of “death panels.” Many defenders of “Obamacare,” insisted that the angry criticism was racially motivated. For example, Jim Winkler (2009), General Secretary of the United Methodist Church, noted that opposition to healthcare reform had “transmogrified into something far deeper, far more elemental.” “Anger,” he observed, was “its salient feature,” and “racism and fear is at the core of the anger.” Moreover, Winkler and others like him saw the “Obamacare” detractors as part of a larger pattern of racial backlash following Obama’s election.

Scholars have already demonstrated the impact of race on Obama’s 2008 victory. Segura and Valenzuela (2010) showed that Racial Resentment had a significant, negative effect on Obama’s candidacy. Similarly, Pasek et al. (2009) concluded that anti-African American racism reduced Obama’s vote share (see also Redlawsk et al. 2010). Furthermore, anti-Obama sentiment went beyond the color of his skin. Consistently, critics of Obama have portrayed him as “un-American” and “other.” For example, Obama’s heritage has sparked the consistent questioning of his American citizenship. These “birthers” view President Obama as a foreigner and his presidency as unconstitutional. In August 2010, a CNN poll found that forty-one percent of Republicans thought that President Obama was either “probably not born” or “definitely not born” in the USA, a finding consistent with most public polling since that time (Travis 2010).

Literature Review

Mechanisms of Racialization

In this article, we explore the ways in which attitudes toward healthcare reform have been racialized. There are three primary theoretical mechanisms. First, research shows that some non-racial political issues can become increasingly seen as “racial” via the effective use of “racial cues.” For example, the general public has come to view a connection between race and public policies designed to decrease poverty, such as welfare, or reduce crime. This link can be established by mass media or in general communication about the issue and can have significant implications (Winter 2008). For example, experimental evidence demonstrates that racial cues, such as “inner city,” or images of dark-skinned people, cause racial attitudes to become a more central predictor of respondents’ evaluations of government efforts to address various public problems (Hurwitz and Peffley 2005; Peffley et al. 1996).

Another plausible explanation for the racialization of healthcare is the result of policy positions of the two major political parties. For many years, particularly since the Civil Rights Movement, the Democratic and Republican Parties have become increasingly divided by race (Carmines and Stimson 1989; Layman and Carsey 2002; Valentino and Sears 2005), with the Democratic Party more liberal on racial issues. This racial division may influence attitudes about other policy positions, even if they are only indirectly related to race. Since past healthcare initiatives have been associated primarily with Democrats (e.g., Medicare and Medicaid under President Johnson in the 1960s and universal coverage under President Bill Clinton in the early 1990s), attitudes about the Affordable Care Act may have become racialized simply because the reform is supported by the Democratic Party.

The third possible mechanism for how healthcare and race may have become interconnected emphasizes the “personal cue”—in this case, President Obama himself. Research suggests that when leaders take stands on political positions, the groups that the leaders represent often become associated with the particular position or proposed solution. In other words, if a specific governmental program, or reform effort, becomes associated with, for example, a highly visible female, then it is possible for people to begin associating the program or reform effort with attitudes toward women in general (Tesler 2012). This mechanism has been shown to activate attitudes toward race and crime (Hurwitz and Peffley 2005) as well as religion and support for the Iraq War (Jacobson 2007).

Racial Attitudes, Healthcare Reform and Existing Evidence

Recent scholarship suggests that opposition to healthcare reform is driven primarily by partisanship and age (Gelman et al. 2010; Kriner and Reeves 2011). Further, the “personal cue” concept has been shown to influence attitudes about both gender and race in the area of healthcare reform (Knowles et al. 2010; Winter 2008). And a growing body of evidence suggests that President Obama’s African American heritage has caused attitudes toward healthcare reform to become racialized. For example, using the American National Election Survey (ANES) from 2008–2010, Byrd et al. (2011) found that “whites who were racially resentful were less likely to support the healthcare reform law.” Similarly, Knowles et al. (2010) found that when participants read about a healthcare reform proposal that was presented as “Obama and the Democrats approach to healthcare reform,” as opposed to “Bill Clinton’s 1993 healthcare reform,” participants with negative racial attitudes evaluated the reform more negatively when framed as Obama’s rather that Clinton’s. Similarly, Henderson and Hillygus (2011) examined panel data and opinion change from 2008 to 2010 and found that racial attitudes were a strong predictor of change in opposition to healthcare reform.

In addition to evidence suggesting attitudes toward healthcare reform are increasingly related to racial attitudes, there is some evidence suggesting that Ethnocentrism may also play an important role. After all, not only President Obama represents the first African American President, but to many he represents “otherness.” For example, despite President Obama’s continued public professions of adherence to Christian beliefs and being born an American, a sizable percentage of the public continues to believe that he is not only foreign born, but is also secretly a Muslim (Pew Research Center 2010). In fact, Tesler and Sears (2010) found that attitudes about Muslims were not only a significant predictor of voting intention in the 2008 presidential election but also support for private health insurance. As they argue “[l]ike the spillover of racialization, this result suggests that President Obama may also make ethnocentric opinions about out-groups a more important factor of partisan political decision making in the years ahead” (Tesler and Sears 2010, p. 17).

We expand on existing research in several ways. First, while some research has included measures of Racial Resentment (Byrd et al. 2011; Henderson and Hillygus 2011; Tesler 2012) and Tesler and Sears (2010) examine the effects of attitudes toward Muslims, scholars have yet to fully investigate the effects of Racial Resentment and Ethnocentrism. In some previous research, attitudes about Muslim’s have been used as a proxy for Ethnocentrism (Tesler and Sears 2010), but the theoretical and empirical work by Kinder and Kam (2009) provides much greater leverage for uncovering any relationships between Ethnocentrism and attitudes toward healthcare reform.

In addition, while Byrd et al. (2011) took advantage of the ANES, their investigation did not include controls for important variables such as religion, living in the southern region of the USA or political sophistication. Given that southerners and respondents who are very religious are also generally conservative, omitting controls for these variables could have biased conclusions regarding influence of both ideology and Racial Resentment on support or opposition to healthcare reform. In addition, most existing research examines attitudes toward a specific aspect of healthcare reform (Tesler and Sears 2010) or an overall evaluation of reform efforts (Byrd et al. 2011; Henderson and Hillygus 2011; Tesler 2012). To allow for a more developed understanding of opposition to healthcare reform, we follow the approach by Knowles et al. (2010) and include several questions about healthcare reform, ranging from concerns about euthanasia, socialism and benefits to undeserving people.

Racial Resentment and Ethnocentrism

Racial Resentment, similar to its predecessor symbolic racism, reflects the changing nature of racial attitudes in the USA. Whereas Jim Crow racism was based on the fundamental belief in the necessity of racial separation and black political oppression, Racial Resentment is based on “a blend of anti-black affect and the kind of traditional American moral values embodied in the Protestant Ethic” (Kinder and Sears 1981, p. 416). Sniderman, Piazza, Tetlock and Kendrick surmise that the transformation of old-fashioned racism, or Jim Crow racism, into Racial Resentment hinges on the “perception of blacks as ‘violating cherished values’” (1991, p. 424; McConahay and Hough 1976, p. 39). Henry and Sears (2002) contend that Racial Resentment is cultivated in early life when negative ideas of African Americans are held in conjunction with conservative moral values (Sears et al. 2000; Tarman and Sears 2005).

The transformation of racial attitudes has important implications for support and opposition to various public policies. Recent work on Racial Resentment has demonstrated that white resistance to public policies results from the activation of racial attitudes, rather than a realistic threat to white self-interest. Studies have revealed that Racial Resentment is closely related to white opposition to various public policies, often policies that are indirectly linked to race, such as housing, busing and crime (Gilliam and Iyengar 2000; Peffley and Hurwitz 2007; Peffley et al. 1996).

Though Racial Resentment is a potentially powerful explanation for some of the outcomes in the 2008 presidential election, as well as some of the policy outcomes under the Obama administration, the theory has faced challenges. The most significant concern is that Racial Resentment is not clearly distinct from old-fashioned racism or conservative ideology. While these measures are clearly correlated, Sears et al. (1997) argue that levels of Jim Crow racism have been consistently declining while levels of Racial Resentment remain substantial. Further, factor analysis and structural equation models demonstrate that political conservatism and Racial Resentment are distinct and Racial Resentment contributes significantly to multivariate models above and beyond traditional measures of political conservatism (Hutchings and Valentino 2004; Krysan 2000; Sears and Henry 2003, 2005; Sears et al. 1997; Valentino and Sears 2005; Tarman and Sears 2005). In many previous studies, the Racial Resentment scale has been demonstrated to be a reliable valid scale and possesses “construct, predictive and discriminant validity” (Sears and Henry 2005, p. 258).

In addition to the importance of Racial Resentment as a fundamental aspect of contemporary political attitudes, recent research also suggests that Ethnocentrism is also an important part of understanding contemporary opinions (Kam and Kinder 2012). The argument, as Kinder and Kam note, can be traced back to William Graham Sumner who concluded that Ethnocentrism was the “technical name for this view of things in which one’s own group is the center of everything” ([1906] 2002, p. 13; Kinder and Kam 2009, p. 1). In this article, we employ Kinder and Kam’s measure of Ethnocentrism created from the 0 to 100 point “feeling thermometer” scales. They argue that this measure proves significant in evaluating such diverse categories as “the war on terrorism, humanitarian assistance to foreign lands, immigration and citizenship, the sanctity of marriage, social security and welfare reform, and school desegregation and affirmative action” (2009, p. 3).

Measuring how respondents’ feel toward their in-group compared to various out-groups is particularly useful in assessing whether Obama’s “otherness” shapes views on healthcare reform, particularly since African American candidates may seem less American to many Caucasians (Devos et al. 2008). Moreover, Kay and Mayer (2010) have noted the ramifications for policy views, concluding that immigration attitudes in Virginia have been triggered by Obama’s “otherness.” And Cheryl Kaiser et al. (2009) and her colleagues have concluded that in the post-Obama era, there has been less support for public policies that are designed to address racial inequalities. If government support and dollars are perceived as a zero-sum game, then competition between in-groups and out-groups (Ethnocentrism) could be on the rise. Moreover, it could trigger opposition to healthcare reform if it is associated with alleviating racial inequalities.

Expectations

Because health care is a public policy related to attitudes about race and ethnicity, we expect individuals who embrace greater levels of Ethnocentrism and Racial Resentment to express lower levels of support for healthcare reform. Further, the fact that the first African American President has been such a strong proponent of healthcare reform renders attitudes toward race even more closely linked to attitudes about healthcare reform. Further, we expect that these attitudes will play a significant role in opposition to healthcare reform even when controlling for the effects of other potential factors.

H1 :

We expect that both Racial Resentment and Ethnocentrism will have a direct influence on healthcare attitudes. We expect that increasingly resentful and ethnocentric individuals will dislike healthcare reform more than those who are less resentful and less ethnocentric

H2 :

We expect respondents who are asked to evaluate “President Obama’s healthcare reform” to be less supportive than respondents who are asked to evaluate “recent healthcare reform.”

H3 :

We expect that the importance of Racial Resentment and Ethnocentrism will be greater among those respondents asked to evaluate “President Obama’s healthcare reform” compared to those who are asked to evaluate “recent healthcare reforms.” We expect respondents who are increasingly resentful and ethnocentric to evaluate healthcare reforms more negatively if they are asked to evaluate “President Obama’s healthcare reform.”

Methods

Participants

The data used in this analysis comes from a national internet survey fielded immediately following the November 2010 midterm elections and was administered by Knowledge Networks (www.knowledgenetworks.com). Knowledge Networks proprietary database features a representative sample of Americans, including representation of the roughly 30 % of US households that do not have internet access. In addition, through address-based sampling, the database covers the growing number of cell phone-only homes, recently estimated at 23 % of all households.

The survey was conducted online in both English and Spanish, and was assigned to 5,844 participants in the Knowledge Networks panel; 3,406 respondents completed the survey, a 58 % cooperation rate. The survey took an average of 21 min to complete and included a total sample of 3,406 individuals who were 18 years and older. Included in this large survey were 1,649 white, non-Hispanic respondents. Throughout the analysis below, we restrict the sample to only adult Caucasian participants.

Before answering questions about various aspects of healthcare reform, participants were randomly assigned into two groups. Prior to the questions about reform, all participants were presented with the statement, “There has been a great deal of discussion recently about healthcare reform.” Then, participants were asked to respond to a general question about healthcare reform: In general how do you feel about [the recent/President Obama’s] approach to healthcare reform? The control group was asked to evaluate “recent” approaches to healthcare reform, while the experimental group was asked to evaluate “President Obama’s” approach to healthcare reform. Participants were asked to evaluate the above question on a 5-point scale ranging from “strongly oppose” to “strongly approve.” While reference to either “recent approaches” versus “President Obama’s approach” to healthcare reform was a subtle manipulation, any difference between groups demonstrates the importance of even a brief mention of President Obama’s name in relation to healthcare reform.

Measures

Healthcare Reform

First, respondents were asked to respond to an overall evaluation of healthcare reform. The overall evaluation item was introduced by the following: “There has been a great deal of discussion recently about healthcare reform.” Then, participants were asked to respond to the following: In general how do you feel about [the recent/President Obama’s] approach to healthcare reform? A random sample of respondents were presented with the question phrased as “recent” approach to healthcare reform, while the other half were presented with “President Obama’s” approach to healthcare reform.

Following the response to the overall question, respondents were asked to evaluate more specific concerns about healthcare reform. The items related to more specific concerns were introduced by the following: “To what extent are you concerned that recently proposed reforms may lead to the following?” Participants were asked to respond to each of potential outcomes along a 5-point scale ranging from “not at all concerned” to “extremely concerned.” Responses to these questions were recoded so that higher scores reflect negative evaluations of healthcare reform and lower scores reflect increasingly positive evaluations of healthcare reform.

As shown in Table 1, these seven questions refer to specific concerns that have been expressed by opponents of healthcare reform, such as reduced services, lower quality and the reforms leading to socialism. These seven questions were recoded to range from liberal to conservative. In other words, higher scores reflect more negative evaluations of healthcare reform and lower scores reflect increasingly positive evaluations. Throughout the rest of this analysis, we analyze responses to the overall healthcare evaluation question separately and refer to this question as the “overall evaluation question.” In addition, we analyze responses to the seven more specific concerns separately as a scale. We summed responses to these seven specific questions and refer to this index as the healthcare reform scale. These questions were summed to produce a seven-item scale with range of 0–32, a mean value of 24.5, standard deviation of 8.9 and an alpha of .93.Footnote 1 The healthcare reform scale is correlated with the overall evaluation question at .65.Footnote 2

Table 1 Means, standard deviation and item correlations for healthcare reform index
Full size table
Table 2 Means, standard deviations and item correlations for Racial Resentment scale
Full size table
Table 3 Descriptive statistics for racial, demographic and political variables
Full size table

Racial Resentment

To measure Racial Resentment, we relied on the validated scale developed by Henry and Sears (2002). Participants were asked the following questions

  • Irish, Italians, Jewish and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.

  • Over the past few years, blacks have gotten less than they deserve.

  • It is really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.

  • Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.

Responses ranged from 1 = strongly agree to 5 = strongly disagree. All responses were recoded to range from lower scores indicating an individual who is racially unresentful to higher scores representing an individual who is racially resentful. As reported in Table 3, the scale ranged from 4 to 20 with a mean value of 14.1, a standard deviation of 3.35 and an alpha of .78. This measure of racial animosity was purposefully chosen because it has been shown to be internally consistency, have strong construct validity, discriminant validity, generalizability and predictive validity (Henry and Sears 2002; Sears and Henry 2003, 2005; Tarman and Sears 2005). Additional descriptive statistics associated with this scale are presented in Table 3.

Ethnocentrism

To measure white Ethnocentrism, we followed Kinder and Kam (2009) who subtracted the average of feeling thermometer score (0 = most negative feelings/“cold feelings” to 100 = most positive/“warm feelings”) of the in-group (in this case, whites) and the average score of out-groups (African Americans and Latinos). In effect, this provides a measure of the difference a respondent feels between their own racial group and other groups. To more closely approximate normality, we transformed the measure of Ethnocentrism by adding 74 to each score and taking the square root. The measure is coded so that higher scores represent greater Ethnocentrism. The transformed measure has a mean of 9.4, a standard deviation of 1.05 and a range of 0–13. Additional information is presented in Table 3. This measure of Ethnocentrism has been used in previous research and found to have construct validity (Kam and Kinder 2012; Orey and Park 2012).

Attitudinal Measures: Party Identification, Ideology and Fundamentalism

We measured party identification with the standard seven-point scale with 1 = strong Democrat, 2 = Democrat, 3 = Independent leaning Democrat, 4 = Independent, 5 = Independent leaning Republican, 6 = Republican and 7 = strong Republican. Similarly, we measured ideology along a seven-point scale with 1 = strong liberal, 2 = liberal, 3 = moderate, leaning liberal, 4 = moderate, 5 = moderate, leaning conservative 6 = conservative and 7 = strong conservative. We measured religious fundamentalism by asking respondents to indicate which response more closely reflects their views of the Bible. We gave the value of one to respondents who indicated that they believed “The Bible is the actual Word of God and is to be taken literally, word for word.” Respondents were coded zero if they indicated that they believed that “The Bible is the inspired Word of God, but not everything in it should be taken literally,” or, “The Bible is a book written by men and is not the Word of God.”

Political Sophistication

Following the work of Delli Carpini and Keeter (1993), the knowledge based variable, political sophistication, was coded as a sum of the number of correct answers to the following three questions: (1) Do you happen to know who Eric Holder is? (2) What job or political office does Joe Biden now hold? (3) What job or political office does John Roberts now hold? This measure runs from 0 to 3. Approximately 22 % of respondents did not answer any question correctly, while approximately 41 % answered one question correctly, 16 % answered two questions correctly and 22 % answered all three questions correctly. Additional descriptive information for all variables is presented in Table 3.

Demographic Measures: Education, Age, Gender, Region and Employment

Education was measured with four categories ranging from 1 = less than high school, 2 = high school, 3 = some college, to 4 = bachelor’s degree or higher. Gender was coded one for female participants and zero for male participants. Similarly, participants living in the south were coded one, while non-southern respondents were coded zero. Participants who were unemployed were coded one, while respondents who were employed, disabled or retired were coded zero. In addition, we included age in our analysis as well as age squared because older respondents are expected to be increasingly opposed to healthcare reform, but the relationship changes and opposition attenuates among those in the oldest age categories.

Results

Descriptive Findings

Looking first the two primary independent variables, Racial Resentment and Ethnocentrism, we find that the average respondent expressed slightly ethnocentric and resentful responses. For example, the mean score for Racial Resentment among Caucasian respondents was 14.28 with a standard deviation of 3.37. Since the range was 4–20, the average response was slightly resentful. Similarly, the mean for Ethnocentrism was approximately 9.4 with a standard deviation of 1.05. The range for the Ethnocentrism was 0–13, so the typical response was in the ethnocentric direction. Looking next at our two dependent variables, the overall healthcare question and the healthcare index, we find that the mean for the overall question was 3.6 with a standard deviation of 1.28. The mean of the healthcare index was approximately 25 with a standard deviation of 8.05 and a range of 7–35. Since the midpoint of this scale is 21, a mean of 25 indicates that the typical participant was slightly opposed to healthcare reform.

Bivariate Relationships

In Table 3, we present the bivariate correlations between the variables included in our analysis. We do not include the treatment variable in this table because it was uncorrelated with any of the other variables. Looking first at the correlation between the overall healthcare question and the remaining variables, we see significant bivariate correlations with both Racial Resentment (.49) and Ethnocentrism (.17). The correlation between the overall healthcare question and Racial Resentment is more than double the correlation between the overall healthcare question and Ethnocentrism, but both correlations are positive and significant. In addition, the overall healthcare question is significantly and positively correlated with party identification, ideology, living in the south, fundamentalism and political sophistication. It is negatively and significantly correlated with being female and education. It is not significantly correlated with being unemployed or age. Looking next at the healthcare reform scale, we see similar patterns. We find significant and positive correlations between Racial Resentment (.55) and the healthcare reform scale as well as significant and positive correlations between Ethnocentrism and the healthcare reform scale (.24). Again, we find a greater correlation between the healthcare reform scale and Racial Resentment compared to the correlation of the scale with Ethnocentrism. In addition, we find significant and positive correlations between the healthcare reform scale and living in the south, age, party identification, ideology and fundamentalism. Finally, we find significant and negative correlations between education, being unemployed and political sophistication (Table 4).

Table 4 Bivariate correlations (n = 1,649)
Full size table

Looking next at Racial Resentment, we find significant and positive correlation between Racial Resentment and Ethnocentrism (.36), as well as living in the south, party identification, ideology and fundamentalism. We find that Racial Resentment is significantly and negatively correlated with being female, education and political sophistication. In a similar manner, Ethnocentrism is significantly and positively correlated with party identification, ideology, being fundamentalist, being unemployed and age. It is significantly and negatively correlated with education and political sophistication.

Multiple Regression Analyses

Looking first at model 1 in Table 5, we include only demographic variables as predictors of the overall healthcare question. In this model, the only variables that reach conventional levels of statistical significance are education and being female. Both of these variables are negative indicating that increases in age and being female are related to more positive evaluations of healthcare reform. Looking next at model 2, we include additional attitudinal control variables. In model 2, we see that education is still a negative and significant predictor, but being female is no longer significant. The effects of party identification and ideology, however, are positive and significant. Predictably, being republican and conservative are associated with more negative evaluations on the overall healthcare reform question. We note also that the model now explains significantly more variation with .39 increase in the R square in model 2 compared to model 1. Looking next at models 3 and 4, where we now include a dummy variable for the experimental treatment (1 = “President Obama’s reforms”, 0 = “recent reforms”), as well as Racial Resentment and Ethnocentrism. Here, we find a positive and significant effect associated with the treatment condition as well as Racial Resentment. In this case, we find support for hypotheses one and two. In other words, those participants who were randomly assigned to the group presented with “President Obama’s” recent healthcare reforms provided significantly more negative evaluations compared to the group that was presented with “recent” healthcare reforms. Racial Resentment is also positive and significant indicating that increasingly resentful respondents were significantly more negative in their overall evaluation of health care. Despite our expectation, however, the effect of Ethnocentrism is not significant in this model. Looking next at model 5, we include interaction terms between the treatment condition and both Racial Resentment and Ethnocentrism. In this model, we find that the interaction term between the treatment condition and Racial Resentment was not significant. The interaction term between the treatment condition and with Ethnocentrism was also not significant. Therefore, we fail to find support for hypothesis 3.

Table 5 Individual level determinants of overall healthcare evaluation question
Full size table

Presented in Table 6 is a similar set of multiple regressions predicting responses to the seven-item healthcare reform scale. As before, model 1 includes only demographic variables. Here, we see that education, being unemployed are negatively and significantly related to the healthcare scale. Being unemployed or having greater levels of education were associated with more positive evaluations on the healthcare scale. Living in the south and the respondents’ age, however, were significant and positive predictors, indicating increasingly negative evaluations on the healthcare scale. In model 2, we see that when we include politically relevant variables, the R square increases considerably and both party identification and ideology are positive and significantly related to values on the healthcare reform scale. More conservative and more republican respondents expressed greater concerns. Further, the effect of political sophistication is negative and significant, indicating that more politically knowledgeable respondents were less concerned about healthcare reforms. Also in this model, we find that the negative effect associated with education remains significant. Finally, we see in model 2 that when we control for other factors, the effect of being female is positive and significant. In this case, being female was associated with significantly more opposition to healthcare reform. It appears that once we control for the tendency of women to be more democratic and liberal, then they were more likely to express concerns about healthcare reform compared to their similarly situated male counterparts.

Table 6 Individual level determinants of healthcare reform scale
Full size table

In models 3 and 4, we include a dummy variable for the treatment condition (1 = “President Obama’s reforms,” 0 = “recent reforms”) as well as Racial Resentment and Ethnocentrism. In this case, there were no significant differences between the treatment and the control conditions. When looking at the seven-item healthcare reform scale, the mention of president Obama did not significantly alter respondents’ evaluations. Therefore, we find no evidence for hypothesis two in this model. The effects of Racial Resentment and Ethnocentrism, however, are significant and positive, indicating support for hypothesis one. Respondents who were resentful and ethnocentric expressed more concern over healthcare reform. Including these variables in the model also increased the explained variation (from .403 in model 3 to .508 in model 4). The effect of education and political sophistication both remains negative and significant. In model 5, we include interactions between the treatment condition and both Racial Resentment and Ethnocentrism. In this model, the predicted interactions were not significant, and therefore, we find no support for hypothesis three.

Overall, we found consistent evidence of support for the influence of Racial Resentment on responses to the overall healthcare question as well as the healthcare reform scale. We found evidence of support for the influence of Ethnocentrism on the healthcare reform scale, but not on the overall healthcare question. Therefore, hypothesis one was supported in three out of four instances. We found that referencing President Obama resulted in significantly more negative responses to the overall healthcare question, but was not significant in predicting the healthcare reform scale. Therefore, hypothesis two was supported in only one of two instances. Finally, we found no support for hypothesis three that the influences of Racial Resentment and Ethnocentrism would be greater among those respondents presented with a reference to President Obama.

Discussion

Healthcare reform remains a critical issue in American life and on the campaign trail, and it will continue to gain a great deal of public interest and media attention. Moreover, we contend that it must be considered a “racial issue,” not only because it involves providing healthcare benefits to all groups in the country, but also because the first African American president has argued so strongly for these reforms and has become so clearly identified with this policy issue. However, our knowledge about the complicated manner in which attitudes about other groups and attitudes about race influence support and opposition of healthcare reform is limited. We hope that scholars will continue to investigate the interplay of race and attitudes toward this important area of public policy, particularly by expanding investigations to different racial and ethnic groups.

Our findings indicate that Racial Resentment plays an important role in shaping white Americans’ attitudes toward healthcare reform. Increased levels of Racial Resentment were associated with lower levels of support for healthcare reform among whites, even controlling for a range of alternative explanations. In fact, Racial Resentment was consistently a significant predictor of attitudes about healthcare reform, even when controlling for background factors and other attitudinal measures. This study provides additional evidence for the construct validity of Racial Resentment and importance of racial animus in contemporary health policy. While framing the healthcare reform as “President Obama’s reform” resulted in increased opposition to reform in the model predicting the single question asking respondents about their overall feelings about reform, there were no significant differences between groups when predicting the seven-item scale comprised of more specific concerns about healthcare reform. Further, while the effects of Ethnocentrism were significant in predicting respondents’ attitudes to the seven-item scale, it was not significant in predicting evaluations to the single overall evaluation question. Further, our analyses indicate that when predicting respondents’ evaluations on the seven-point scale, both education and political sophistication play an important role with respondents who are more educated and more sophisticated expressing fewer concerns with healthcare reform.

Limitations

While the analysis indicates that Racial Resentment plays an important role in attitudes about healthcare reform, even when controlling for party identification, ideology and a range of demographic variables, the subtle manipulation in question wording may not have been sufficient to provide consistent evidence for the role that President Obama himself plays in attitudes about reform. Alternatively, survey respondents may have become increasingly sophisticated in responding to questions about President Obama and may have even moderated their responses when confronted with difficult racial situations. The general public is increasingly sensitive about answering racial questions, even when the survey is conducted confidentially over the internet. To the extent to which respondents intentionally moderated their racial attitudes in order to provide more socially desirable answers, the less we will find relationships between attitudes about race and healthcare reform. Taken in this light, however, finding significant relationships between Racial Resentment and health policy even when respondents may be moderating their racial attitudes, indicates the prevalence of how much healthcare reform has become a racial issue. Yet another limitation is that this investigation represents only a single point in time, while the basis of Racial Resentment suggests that attitudes about race will continue to change and evolve as the general public is forced to face a president that confronts traditional stereotypes of African Americans. Finally, the psychological concepts that we explore in this investigation are very complex and sometimes interrelated. Attitudes and beliefs about Ethnocentrism, Racial Resentment, political knowledge and even religious fundamentalism are all extremely complex constructs, and our measures of these constructs are blunt instruments that may not fully capture the complexity of these constructs. In fact, our relatively blunt measures are most likely one of the reasons behind our mixed results. More encompassing measures of Ethnocentrism, Racial Resentment and religious fundamentalism may have provided more consistent findings.

Directions for Future Research

While we are beginning to understand the public’s support and opposition to healthcare reform, we show in this analysis that attitudes about race and evaluations of other groups are an important part of how white Americans evaluate healthcare reform. Future research, however, must consider more sophisticated ways to evaluate racial attitudes, such as implicit and unconscious measures of racial attitudes. Particularly when internet surveys are sophisticated enough to evaluate respondents’ reactions to various images and trait assessments, future research should investigate how the effect of a visual image of President Obama, or an immigrant, or an image of a Muslim, etc. influences respondents’ support or opposition to healthcare reform. Further, given the pressure of social desirability and racial attitudes, future research should consider the potential role of self-moderation and various information processing contexts. Respondents, for example, may be more likely to rely on stereotypical information processing strategies when distracted, pressed for time, or if they receive their information about healthcare reform from print, internet or radio sources. Political campaigns often specifically target various constituencies with messages designed to illicit particular responses and some of these messages are likely to be greater catalysts for information processing strategies that rely on racial attitudes to form the basis of opinions. In addition, our treatment manipulation, a simple and single reference to President Obama, may not have been strong enough to fully assess the influence of President Obama. In comparison, the national news and reports about healthcare reform generally include captivating images of President Obama. Finally, future research should continue to look at the evolving relationships between race, health care and public support. The unique situation of having a positive “stereotype exemplar” in President Obama provides an unprecedented opportunity for understanding how racial attitudes may change and shape the public’s evaluations of other public policies as well. Ultimately, the arguments surrounding Racial Resentments and Ethnocentrism are dynamic and are focused on how individual’s attitudes about race and “others” evolve and change in the face of societal change and increased interactions with minority groups.