Character  &  Context

Angels and Demons: Differences in Mental Images of Believers and Atheists

Image of expressionless gray masks on a gray wall background

Atheists are considered one of the most untrustworthy social groups in the U.S. People tend to imagine that atheists are morally uninhibited; capable and willing to cheat, steal, and even murder. This assumption that atheism signals immorality may be based in the premise that religious belief is a prerequisite to moral behavior. That is, a moral person is a religious person, and an immoral person must be an atheist.

An immoral person cannot be trusted. Given that human beings live in cooperative, social groups, we are deeply motivated to identify trustworthy and untrustworthy individuals. Religiosity has become a cultural signal of trust. For example, previous research has found that those who signal their religious belief, such as by fasting for a religious holiday, regardless of what religion they identify with, are trusted more than those who do not signal their religious belief. Atheism has become a cultural signal of extreme distrust. For example, previous research has found that anti-atheist prejudice and discrimination is driven by stereotypes associating atheists with being untrustworthy.

Our research investigated whether these cultural stereotypes of theist trust and atheist distrust alter people’s mental images of theists and atheists. That is, do people spontaneously generate positive mental images of theists and negative mental images of atheists? To investigate this question, we used a technique to visually estimate people’s mental images of theists and atheists. Specifically, participants were presented with pairs of “fuzzy” images across four hundred trials. These images consist of a background image of a person and random visual noise added on top of the background image. The visual noise slightly distorted the features of the pictured person on the background image. Then, in one trial, participants were asked to determine which image among the pair looked most like a theist (in one condition) or atheist (in another condition). After participants completed this task, we took all of the images that were selected in the theist condition and averaged the visual noise together to create an average image of a theist. And, we did the same process to create the average image of an atheist (see average images below).

Next, a second sample was asked to rate these images on several dimensions, including perceived religiosity and trust. Importantly, this second sample was not aware of how these images were generated, so any effects would emerge spontaneously. Consistent with our hypothesis, the average theist image was rated as more trustworthy and religious than the average atheist image. And, generally speaking, the average theist image was associated with more positive attributes (e.g., moral, competent, warm, likeable, etc.) than the average atheist image. Further, participants believed that the pictured person in the average atheist image was more likely to behave immorally (e.g., kick a dog for no reason), than the pictured person in the average theist image. Together this suggests that mental images of theists are associated with more positive attributes than images of atheists, and these mental images influence perceptions of potential moral and immoral conduct.

Our initial research raises some interesting questions. For example, the majority of the image generation sample was religious. It may be that both type of religious affiliation and extent of devotion influence people’s mental images of atheists and theists. If theological belief is represented psychologically as a continuum from lack of belief to extreme belief, we would expect this continuum to be represented in people’s mental images. That is, participants may create extremely trustworthy images when they imagine devout theists and trustworthiness would fade as a function of the reduction in devotion to the belief in God. And, our research was conducted in the U.S. using convenience samples. We expect that mental images of atheists and theists would change based on the cultural context. For example, in countries with little anti-atheist prejudice, we would expect mental representation of atheists to be less representative of negative attributes. But, so far, the research presents a clear picture: People tend to imagine devilish atheists and angelic theists.

Two images of faces, left image "Average Atheist Image," right image "Average Theist Image"


Jazmin L. Brown-Iannuzzi is an associate professor of psychology at the University of Kentucky. Her research seeks to understand why social group-based disparities may persist and, in some cases, grow.

Stephanie E. McKee is a graduate student at the University of Kentucky. Her research investigates the psychological mechanisms and consequences of economic inequality.

Will M. Gervais is an associate professor of psychology at the University of Kentucky. He is the director of the Beliefs and Morality (BAM!!!) Lab and studies the psychology of religious belief and disbelief, as well as research methods

Blog Category: 

About our Blog

Character & Context is the blog of the Society for Personality and Social Psychology (SPSP). With more than 7,500 members, SPSP is the largest organization of social psychologists and personality psychologists in the world.   

Learn More ›

Questions ›

Writing Resources ›

Contribute to the Blog ›

Get Email Updates from the Blog