Education and Social Desirability Bias: The Case of a Black Presidential Candidate n Jennifer A. Heerwig, New York University Brian J. McCabe, New York University Objective. Survey research consistently reports a positive association between ed- ucational attainment and socially tolerant attitudes, but critics hold that respon- dents with high levels of education may simply purport to hold attitudes seen as socially desirable. In this article, we seek to adjudicate between the claim that the association between education and social tolerance is simply an artifact of sophis- ticated social desirability reporting on the part of well-educated respondents and the competing theory that education has a real impact on increasing forms of social tolerance. Methods. Using support for a black presidential candidate as our mea- sure of social tolerance, we utilize an innovative online list experiment to test whether high levels of support are inflated because of social desirability reporting among the educational elite. Results. We find no evidence of systematic overre- porting of support for a black presidential candidate among respondents with high levels of education, and note that social desirability bias declines as educational attainment increases. Conclusions. This research bolsters arguments about the lib- eralizing effect of education on socially tolerant attitudes, and challenges evidence that attributes this relationship to high levels of social desirability bias. Survey research consistently reports a positive association between edu- cational attainment and socially tolerant attitudes, but critics argue that this association may be the result of high levels of social desirability reporting among the educational elite, rather than a true measure of social tolerance. One commonly measured estimate of social tolerance is support for a black presidential candidate, an item included in the General Social Survey (GSS) since the early 1970s. The GSS reports that support for a black presidential candidate has grown from 75 percent in 1972 to 92 percent in 1996—the year the question was terminated. As Figure 1 illustrates, support has been n Direct correspondence to Jennifer A. Heerwig, 295 Lafayette St. 4th Fl., New York, NY 10012 h jah321@nyu.edu i . Author names are listed in alphabetical order; both contributed equally to the article. Jennifer Heerwig and Brian McCabe will make available all data and coding information to those wishing to replicate the study. The experimental data used in this article were collected by Time-Sharing Experiments for the Social Sciences, NSF Grant 0094964, Diana C. Mutz and Arthur Lupia, Principal Investigators (Mutz and Lupia, 2003). The authors thank Vivek Chibber, Dalton Conley, Amy LeClair, and Jeff Manza for their thoughtful comments. SOCIAL SCIENCE QUARTERLY, Volume 90, Number 3, September 2009 r 2009 by the Southwestern Social Science Association