Is President Obama’s Race Chronically Accessible? Racial Priming in the 2012 Presidential Election

Abstract

A vast literature indicates that racial animosity has a pervasive influence on the public’s evaluations of U.S. President Barack Obama. Can political communications enhance and/or defuse the link between White Americans’ racial attitudes and evaluations of Barack Obama? In this article, we report the results of an experiment conducted in the midst of the 2012 presidential campaign which examines the effect of political rhetoric on the extent to which evaluations of Barack Obama are racialized. Drawing from research on attitude strength and pretreatment effects in experimental studies, we argue that the use of racial appeals in the pretreatment environment and the strength of citizens’ preexisting attitudes toward the incumbent president may produce a downward bias in average estimates of racial priming effects toward President Obama. After accounting for individual differences in the propensity to form strong attitudes with need to evaluate, we observe substantial effects of campaign rhetoric in priming racial attitudes toward President Obama, especially among individuals who are low in the need to evaluate and who tend to have more malleable political attitudes. We conclude by discussing implications for research on racial priming and the politics of racial intolerance in evaluations of Barack Obama.

Keywords:

• Barack Obama
• presidential campaigns
• pretreatment effects
• racial priming
• voting behavior

Acknowledgments

We thank Gene Borgida, Leonie Huddy, Howie Lavine, participants at the 2012 CSPP election conference, and our anonymous reviewers for helpful comments on earlier drafts of this article.

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s website at: http://dx.doi/org/10.1080/10584609.2016.1166168.

Notes

^1. Although the importance of the pretreatment environment for experimental research was clarified by Druckman and Leeper (2012), acknowledgment of the problem of pretreatment dates back to at least Hovland (1959).

^2. These strategies can also be considered part of President Obama’s efforts to run a deracialized campaign. As noted by McCormick and Jones (1993), deracialized campaigns avoid explicit racial references while also emphasizing issues considered to be “racially transcendent” (see also Yon, 2010; Frasure, 2010; and McIlwain, 2010).

^3. For example, Sides and Vavreck (2013, p. 207) use one of Romney’s ads on Medicare as stimulus materials.

^4. Also making exposure to similar messages in the pretreatment environment more likely is the fact that our sample is highly knowledgeable about and interested in politics relative to the population. Indeed, we find no differences in our effects across levels of political sophistication.

^5. Although 1,800 respondents participated in the first wave of the survey, our analyses only included White respondents (N = 1,443). In the analysis of vote choice, we also dropped respondents who did not indicate an intent to vote for either Barack Obama or Mitt Romney, further reducing our sample size in those analyses.

^6. In addition, our sample characteristics are similar to those seen in other MTurk studies. In particular, we found similar values for respondent age, median income, and gender to those reported in Berinsky and colleagues (2012).

^7. Our randomization appears to have been successful. We find no significant differences across conditions in terms of ideology, partisanship, racial attitudes, need to evaluate, or demographic characteristics.

^8. This quote is just one of many used by Romney on the campaign trail designed to highlight Obama and his economic policies in particular.

^9. Both of our measures focus on the explicit racial attitudes of our respondents. Although implicit racial attitudes have also been shown to influence attitudes toward Obama and his policies (Greenwald, Smith, Sriram, Bar‐Anan, & Nosek, 2009; Knowles, Lowery, & Schaumberg, 2010), we chose to focus on the explicit attitudes that have been the focal point of studies showing the chronic accessibility of racial negativity.

^10. The distribution of both racial resentment and overt racism in our samples are similar to other samples; however, our Mechanical Turk sample reports slightly lower levels of racial prejudice on both measures.

^11. Specifically, on a 0–1 scale, need to evaluate in our study has a mean of 0.591. This is similar to other estimates obtained on Mechanical Turk (e.g., Berinsky and colleagues [2012] report a mean in their MTurk sample of 0.628), and is also similar (though slightly higher) than means reported in national samples (e.g., in an Internet sample obtained via the American National Election Studies [ANES], Berinsky and colleagues [2012] report a mean of 0.579. while a face-to-face sample mean is 0.558).

^12. For example, some studies conduct a tertiary split on need to evaluate (e.g., Jarvis & Petty, 1996) while others divide low from high need to evaluate on the median (e.g., Druckman & Nelson, 2003).

^13. Like Druckman and Nelson (2003), we also find similar, though less consistent, results using the continuous measure of need to evaluate (see supplemental Appendix).

^14. We find that low and high need to evaluate respondents are quite similar demographically and politically (see supplemental Appendix). There are small correlations between need to evaluate and need for cognition, education, racial resentment, and overt racism (Pearson correlation of 0.10–0.125). Those high on need to evaluate enjoy cognitive engagement, are somewhat higher in levels of racial animosity, and are somewhat less educated relative to those low on need to evaluate.

^15. While there is a small positive correlation between need to evaluate and racial negativity (see supplemental Appendix), we find that there is full variation across both racial attitudes measures among both low and high need to evaluate respondents (see supplemental Appendix).

^16. Analyzing how need to evaluate moderates framing and priming effects on political issues versus candidates represents an important avenue for future research.

^17. All of our results are robust when ran across our entire sample (i.e., when including non-White respondents; available upon request).

^18. We also ran our analyses separately by low and high need to evaluate respondents (see supplemental Appendix). Results from this analysis are identical to those reported in the following 3-way interaction: The choice message reduces the impact of racial attitudes among individuals low in need to evaluate, but not among individuals high in need to evaluate.

^19. We tested whether or not need to evaluate provides a significant improvement in model fit using goodness of fit tests. We found that adding need to evaluate does not provide a statistically significant improvement in model fit. However, our focus in this article is not on need to evaluate as a predictor of vote choice; rather, it is about the differences between people high and low on need to evaluate in response to our experimental treatments. Thus, the lack of improvement in model fit with need to evaluate as a predictor variable does not undercut the focal point of our analysis—the different effects of choice and referendum frames on racialized views of Barack Obama among low need to evaluate respondents.

^20. Our results also suggest that the extent to which racial attitudes are accessible varies along a continuum. As racial attitudes become relatively less accessible, other considerations may become more accessible and “override” racial attitudes in candidate or issue evaluation. We look forward to future research in this vein.

Additional information

Notes on contributors

Matthew D. Luttig

Matthew D. Luttig is a Postdoctoral Scholar in the Department of Political Science, University of Chicago. Timothy H. Callaghan is a PhD candidate in the Department of Political Science, University of Minnesota.

Timothy H. Callaghan

Matthew D. Luttig is a Postdoctoral Scholar in the Department of Political Science, University of Chicago. Timothy H. Callaghan is a PhD candidate in the Department of Political Science, University of Minnesota.