A new study challenges the long-held notion that American voters categorically reject candidates who identify as atheists. Across several survey experiments, the researchers found no sweeping penalty for atheist candidates once results were averaged across the full sample. Although Republicans tended to respond negatively and Democrats showed more support, those opposing and favoring attitudes often balanced one another in ways that did not doom the candidate overall. The findings have been published in the Journal for the Scientific Study of Religion.
People assume that voters in the United States will never vote for a candidate who openly identifies as an atheist. This belief has deep historical roots, from accusations against Thomas Jefferson at the dawn of the Republic to whispers during modern presidential campaigns. Yet the nation’s religious landscape is shifting, with fewer people belonging to organized religions and more describing themselves in secular or nonreligious terms.
“Geoff Layman and I (along with John Green) have published the book Secular Surge: A New Fault Line in American Politics, in which we examine both the causes and consequences of growing secularization in America,” explained David E. Campbell, the Packey J. Dee Professor of American Democracy at the University of Notre Dame.
“In that book, we report on experiments testing Americans’ reactions to candidates of varying degrees of secularity (e.g. candidates who describe themselves as ‘not particularly religious’ vs. ‘an atheist.’ In our book, we report that there is more support for secular candidates than suggested by conventional wisdom. In this paper, we joined with our then-graduate student (now an assistant professor at the University of Tennessee) to further explore how American voters react to atheist candidates.”
The researchers conducted four separate survey experiments between 2020 and 2023. These experiments involved a total of 8,869 participants recruited through online platforms from three different survey firms: YouGov, Prolific, and Lucid Theorem. These samples were designed to be representative of the United States population in terms of key demographics like gender, age, race, and education level.
In each experiment, participants were asked to read what appeared to be a news article about a hypothetical candidate running for political office. The researchers randomly assigned participants to different versions of these news stories. In some versions, the candidate was identified as an atheist, while in others, there was no mention of the candidate’s religious beliefs.
In some variations, the candidate’s patriotism was emphasized by describing them as a decorated war veteran. The researchers varied the office the hypothetical candidate was running for in one of the studies to see if the office type affected voter preference. After reading the news story, participants were asked to rate how likely they would be to vote for the candidate on a scale from zero to 100. This setup allowed the researchers to measure how the information about the candidate’s atheism, and other factors, influenced voters’ stated likelihood of support.
The study’s findings revealed a nuanced picture of voter reactions to atheist candidates. Overall, when considering all participants together, identifying a candidate as atheist had little to no negative impact on their perceived electability in most of the experiments.
This overall finding, however, masked significant differences based on political party affiliation. The researchers discovered a clear partisan divide in reactions to atheist candidates. Democrats showed a tendency to be more supportive of a candidate described as atheist, while Republicans tended to be less supportive. This partisan split resulted in the overall effect appearing minimal when looking at the entire sample.
“Contrary to the common belief that identifying as an atheist spells political doom for a candidate, we find that the label of atheist is a political wash overall–because the negative reaction among Republicans is counterbalanced by a positive one among Democrats,” Campbell told PsyPost. “From our previous research, we had reason to believe that atheists face less stigma today than in the past, but we were nonetheless surprised that the ‘atheist penalty’ is so minimal.”
When examining reactions to different political offices, the researchers found that the type of office the candidate was seeking did not significantly change voter reactions. “Whether it is the school board, mayor, governor, senator, or president, voters respond the same way to an atheist. In other words, it is not that voters are especially concerned about an atheist governing their schools, or in the executive vs. legislative branch, or that the presidency is treated differently than other offices.”
Both partisanship and personal religiosity independently influenced voter reactions. Within both Democrats and Republicans, varying levels of personal religious commitment played a role in how they viewed an atheist candidate. Specifically, among Republicans, those with higher levels of religiosity reacted more negatively to an atheist candidate. However, even when considering religiosity, the partisan divide remained evident, suggesting that both religious beliefs and party loyalty independently shape voter attitudes.
Finally, the researchers found that framing an atheist candidate as a patriotic war hero did somewhat lessen negative reactions, particularly among highly religious Republicans. However, this patriotic framing did not entirely eliminate the negative impact of atheism for this group, and the effect was not consistently observed across all groups or when combined with partisan cues.
“Voters’ negative perception of an atheist candidate can be partially assuaged by describing him as patriotic,” Campbell said. “One reason for some Americans’ concern about atheists is a perception that they are not ‘true Americans.’ When we counter that image by describing an atheist as a military veteran, it partially–but, importantly, does not fully–temper anti-atheist sentiment.”
While these experiments suggest that atheists may have more electoral potential than once assumed, the researchers note that their work has some limitations. Survey participants were reacting to short news-style descriptions in a controlled setting. Real campaigns play out over weeks or months, involve multiple debates, and feature many sources of information.
“Our paper is based on a series of experiments, in which subjects read descriptions of candidates,” Campbell noted. “This can tell us how people respond in the moment to information about someone running for office, but it cannot fully replicate a political campaign.”
“Our hope is that our paper illuminates the political implications of the secular population, which has grown dramatically over the last 30 years. While there are few candidates who describe themselves in secular terms, and fewer still who are atheists (or at least admit to being an atheist), it is likely that we will see more such candidates in future.”
“If readers are interested in knowing more about the political characteristics of the secular population–including how secularism’s growth is largely a backlash to the Religious Right–we would encourage them to read our book Secular Surge: A New Fault Line in American Politics.”
The study, “Will Americans Vote for an Atheist?” was authored by David E. Campbell, Geoffrey C. Layman, and Wayde Z. C. Marsh.
