Democrats May Overestimate Republican Prejudice in Elections

Notes on new research.

Posted Mar 12, 2020

Are Republican voters as closed-minded in their electoral decisions as Democrats think they are?

In a recent paper, the authors (Mercier, Celniker, & Shariff, 2020) describe three empirical studies examining Democrats’ estimates that Republicans would be willing to vote for candidates from different demographic categories. The studies examined several interesting hypotheses about Democrats’ beliefs about Republican biases as well as how beliefs about specific biases were related to Democrats’ beliefs about their favored candidate’s electability.

This is not an exhaustive review of their findings. There were many specific hypotheses regarding Democratic estimates of Democratic candidates from different categories that I do not discuss here. For example, the authors tested the perceived electability among Democrats of people from specific demographic categories and Democrats’ perceptions of Elizabeth Warren, Bernie Sanders, and Pete Buttigieg. In this post, I outline some of the findings that were most interesting to me.

The paper was published online prior to inclusion in a journal and has not yet been formally peer-reviewed. As always, I encourage readers to read the entire original article themselves and form their own opinions of the data — and explore the results I do not discuss here.

Data for Study 1 were collected from an online sample of 728 participants (76% White, 13% Black, 7% Hispanic, 6% East Asian; 56% male, 44% female; average age 35.75). Participants were asked about their willingness to vote for political candidates of various demographic groups and their estimates of how Democrats, Republicans, and all Americans would respond to the same questions (on a 0-100% scale). There were 369 Democrats, 175 Republicans, and 167 Independents in the sample.

As a baseline against which to compare estimates from participants, the researchers used data from a national Gallup poll that showed estimates of willingness to vote for a particular demographic category. The national Gallup data had previously shown that Republicans said they were most willing to vote for the following groups: Catholic (97%), Black (94%), Jewish (94%), Hispanic (92%), evangelical (92%), or a woman (90%).

On average, Democrats in the sample misestimated many categories. This includes average Democrat estimates that Republicans would indicate a willingness to vote for a candidate who is Catholic (70%), Black (40%), Jewish (45%), Hispanic (37%), evangelical (76%), or a woman (43%).

The national Gallup data had previously shown that Republicans were reportedly least willing to vote for the following groups: socialist (19%), Muslim (38%), or atheist (42%). Democrats substantially missed the mark on two of these three, given the average Democrat estimates that Republicans would indicate a willingness to vote for a candidate who is Muslim (21%) or atheist (29%).

Thus, Democrats overestimated the negative response of Republicans toward the categories of Catholics, Blacks, Jews, Hispanics, Evangelicals, and women, with a particularly incorrect assessment of Republican bias against Hispanics. That is an interesting misconception of Republicans given that Hispanic candidates Marco Rubio and Ted Cruz were two of the major challengers in the 2016 GOP presidential primary.

Democrats also overestimated Republican rejection of a candidate who is Muslim or atheist. Additionally, Democrats overestimated how many Republicans would be willing to vote for a candidate over 70 years old or a socialist. Given that the three top contenders for the Democrat and Republican nominations are all over 70 years old (Biden, Sanders, Trump), as a socialist, Sanders may have the most to lose in terms of national electability. Additional data in Study 1 showed that Republicans were more accurate in predicting Democrat willingness to vote for candidates than Democrat predictions were of their own party. This could be due to Republicans being more attuned to the current splits in the Democratic Party than Democrats are in making their estimates.

Data for Study 2 were collected in January 2020 from an online sample of 597 participants. It only surveyed Democrats and added questions about how much contact the participant has with Republicans. To me, the most interesting finding in Study 2 was that the more regular contact that Democrat participants had with Republicans, the more accurate their estimates of the willingness of Republicans to vote for a candidate of a particular demographic. This result seems to underscore the need to get out of our echo chambers and talk to each other.

Data for Study 3 were collected in February 2020 from an online sample of 930 participants. It was the same as Study 2, except that it had an experimental manipulation: Participants were either given base rate information on the true percentage of Americans willing to vote for a candidate from a particular demographic group or they were not given such information. Being provided with base rate data did lead to Democrats estimating higher electability of a candidate who is atheist, Black, female, gay, Hispanic, Jewish, or Muslim, and lower electability of a candidate who is Catholic, evangelical, Christian, socialist, or over 70 years old.


The authors of the reviewed research conducted three studies that show how Democrats perceive Republicans and provides insight into how electability, attitudes towards groups, and perceived attitudes of others towards groups can influence a person’s support for a particular candidate. It is appropriate for political strategists to use this psychological science to determine their strategies. Most importantly, it gives basic researchers in psychological science insight into how attitudes influence the current political climate.


Mercier, B., Celniker, J. B., & Shariff, A. F. (2020). Overestimating explicit prejudice causes Democrats to believe disadvantaged groups are less electable. PsyArXiv Preprints.

The Economist. (2020, March 9). Americans overestimate voters’ prejudices against women and ethic minorities. The Economist (online).