Field of Science

Atheists and rapists: you just can't trust 'em

Atheists are a pretty disliked bunch of people in North America. Most atheists will be aware of polling data that puts them at the bottom of the loathing pile.

Question is, what's driving that loathing? Will Gervais (University of British Columbia, Canada), who's previously published some fascinating research into this topic, is back with some more research (co-authored by another couple of names familiar to this blog: Azim Shariff and Ara Norenzayan).

Gervais' basic hypothesis is that prejudice against people who are not part of your group can be driven by different fears. For example, White Americans fear Black Americans, but view homosexual Americans with disgust. Gervais puts that together with another idea that many people have - that fear of supernatural punishment makes people more honest - to hypothesise that people dislike atheists specifically because they distrust them.

To test this, they took advantage of a clever psychological trick. Here is its original form (invented by Nobel prize-winning psychologist Daniel Kahneman), as described recently in The Guardian:

Linda is a single 31-year-old, who is very bright and deeply concerned with issues of social justice. Which of the following statements is more probable: a) that Linda works in a bank, or b) that Linda works in a bank and is active in the feminist movement? The overwhelming majority of respondents go for b), even though that's logically impossible. (It can't be more likely that both things are true than that just one of them is.) This is the "conjunctive fallacy", whereby our judgment is warped by the persuasive combination of plausible details. We are much better storytellers than we are logicians.

In Gervais' twist on this classic study, students at the University of British Columbia were told about Richard. Here's Richard's story:

Richard is 31 years old. On his way to work one day, he accidentally backed his car into a parked van. Because pedestrians were watching, he got out of his car. He pretended to write down his insurance information. He then tucked the blank note into the van’s window before getting back into his car and driving away.

Later the same day, Richard found a wallet on the sidewalk. Nobody was looking, so he took all of the money out of the wallet. He then threw the wallet in a trash can.
So, is Richard most likely to be a teacher, or a teacher and a Christian? What about a teacher and Muslim. Or a rapist? Or an atheist?

Well, the chilling results are shown in the graphic. Atheism was up there with rapist as an intuitive fit to Richard's character. Atheists? Don't trust 'em!

Gervais and co ran another study, in which half the students were given a different version of Richard. This Richard is not untrustworthy, but he is disgusting (with horrible, flaky skin and snot all over him).

They found that that the disgusting Richard was not associated with atheism (or, indeed, with homosexuality - even though they found in a different study that homosexuals evoke disgust).

What this and some other studies they did showed is that the reason atheists are disliked is specifically because they are distrusted.

They also found that the degree of this distrust is governed by the strength of belief that supernatural monitoring helps to enforce good behaviour. Those who believe this are most likely to distrust atheists.

So although lack of familiarity with atheists increases distrust, it seems that the root of this distrust is not simple fear of the unknown, or even fear about moral corruption, but rather a genuine and seemingly deep-rooted fear that people will not behave well unless they have an invisible policeman watching over them.

Which probably says rather more about these Christians than it does about atheists!
Gervais, W., Shariff, A., & Norenzayan, A. (2011). Do you believe in atheists? Distrust is central to anti-atheist prejudice. Journal of Personality and Social Psychology, 101 (6), 1189-1206 DOI: 10.1037/a0025882

Creative Commons License This article by Tom Rees was first published on Epiphenom. It is licensed under Creative Commons.


  1. There's a heck of a logic shift from concluding that atheists and rapists are not trusted to determining that Christians are only trustworthy because of a moral policeman. It rather mirrors the fallacious logic pointed out in the first example in the article.

  2. In one of their studies, they asked the participants whether they thought people behave better if God is watching them, and found that this belief moderated the effect ( from a statistical perspective). They did 6 studies in total, and they also spent a lot of time discussing alternative explanations, so what I've written does miss a lot of detail!

  3. I need to look into it more, but there is a theory that judicial gods partly came into being as communities got too big to monitor each other's behaviour. Hunter gatherers have few private moments. City dwellers have many. The judicial god is a metaphor for the community's over-sight of each other (and of course we no longer live in communities, were surrounded by strangers most of the time!). Swiss-Army knife Gods like Yahweh just combine this function into one convenient package.

    The Jain, Hindu and Buddhist ideas of karma take the place of judicial gods - karma is simply an impersonal design feature of the universe which ensures that what goes around comes around. Though Hindu's did start out with Varuṇa and Mitra to watch over them making sure they obeyed the law (ṛta) this seems have changed under the influence of Jainism and Buddhism before the common era. India is interesting because there is some documentation of the process of forming gods by people.

    And isn't it true that lot's of us don't play by the same rules when someone is watching as when they are? Is there research on public and private morality? Don't we even say that what goes on behind closed doors is nobody's business? We can do whatever we like when no one is observing us - and isn't that sometimes a relief?

    Unfortunately the whole thing breaks down because we know that wickedness is not confined to a particular group who believe in public retribution for private acts. Sinners arise in every community.

    I am an atheist (A Buddhist atheist with some reasonable doubts about traditional karma narratives) but this study does make me wonder about what keeps an atheist honest when no one is watching. I suppose the question broadens out into the reasons anyone is honest and orderly when no one is watching. Why is anyone moral? Is there some research on this you could recommend? I don't mean historical narrative or philosophical speculation, I mean recent science.

    Given how hostile prominent atheists are towards Christians, the kinds of things the write and say about them, I find I don't really trust them either. Especially in the USA the dialogue - or should I say the two parallel monologues - seem polarised, angry, and lacking any sense of empathy. Only conflict can result it seems to me. And both sides seem to welcome that fight. Or is it just primate posturing and dominance displays?

  4. When you talk about a policeman, this is really too simplistic to account for what believers refer to as the "fear of God". Fear of God does not have a lot to do with fear, except for the fear of displeasing the Lord. Fear of God for a believer is probably very close to the feeling of a higher moral imperative that atheists experience when they are confronted to a moral choice. Moral atheists and moral Christians do feel accountable for what they do, right ? Feeling accountable for what I do is what I refer to by "fear of God". A Christian who trusts and distrusts Christians and atheists equally.

  5. Only a theist would read about a study suggesting that people think of atheists' trustworthiness the same way they think of a rapist's, and respond with "Wha wha, stop picking on us Christians! Your mildly snarky closing line very slightly misrepresents the subtleties of our mythology!" Shameful.

  6. The religious have done far more lasting damage in this world than atheists and rapists. We can at least trust the religious to continue the trend of wishing they were better than they are. Atheists will keep getting better because of reason and practice. Rapists will at some point be diminished to the point of obscurity as stricter enforcement is applied to the religious doctrines concerning the treatment and respect afforded women.

  7. I find the distrust conclusion plausible but just accepting that doesn't take us terribly far it seems.

    We do trust people or not based on a quick superficial impression from available data, such as group membership (and possibly perception of the way they think and what they believe?) but that obviously doesn't exhaust the concept of trust for us. So what we are really looking at here from my perspective is how we form powerful and persisting first impressions based on things like our assumptions about other people based on their beliefs. If that is indeed what is going on as it seems at first to be.

    I don't know if this data teases it out, but I suspect that atheists tend to mistrust each other relative to the average trust level between members of the same religious community. That points to something other than simple out-group prejudice at work, which is why I think the question "how do beliefs really enter into this at all?" comes into play.

    Do we somehow instinctively and below our awareness assume people act on their beliefs, and that
    we understand the relevant beliefs by knowing what religious community they do or do not belong to? That seems an odd instinct to have from an evolutionary perspective, though perhaps not impossible depending on the underlying nature of beliefs.

    I think we rely a lot on predicting each others' behavior by assuming we are all acting rationally on stable beliefs when we don't have more intimate knowledge of each others behavior patterns.

    Neither acting rationally nor having stable beliefs is neccessarily a reliable baseline for prediction except in aggregate I suspect. It may truly be an approximation. It represents our behavior much more closely when we have time and resources for reflection, and even then it is often a rationalization rather than the cause of our decisions. Plus leisurely detached reflection is not happening under the most common circumstances where we would want to predict each others' behavior.

    If you buy into the moral policeman idea, do even atheists tend to have this feeling that atheists are less trustworthy because they don't think they are being watched? Do we have an intuitive assumption that people have to be watched in order to be good and are we making that first approximation of rationality and assuming that atheists can't be trusted because they refuse to believe that they are being watched? Even when we know we don't believe it ourselves?

    Behavior consistent with moral principles clearly doesn't *depend* on literally believing we are being watched, since we find plenty of cases where people do it when they have no reason to think they are being watched. Still, we also know it is often a factor since most of us do act differently when we perceive we are being watched.

    It's at least an interesting guess to try to test. To me we need more of the individual and psychologically oriented research to tease these things out. Relying as heavily as we often do on sociological aggregate data seems to give rise to problems in biased interpretation that rely on gross approximate theories of human decision making like "rational choice."

  8. Jayarava said...

    ...this study does make me wonder about what keeps an atheist honest when no one is watching.

    As I understand it, altruistic behavior is widespread throughout the animal kingdom because it gives indirect evolutionary benefits--it is adaptive. In other words, being good is a human instinct, as is our propensity to see altruism as good. Unfortunately, dividing into groups and seeing insiders as good and outsiders as bad are also instinctive human behaviors. I hope most atheists don't see the religious a group of bad outsiders.

    Anonymous said...

    The religious have done far more lasting damage in this world than atheists...

    This is a weak argument because the religious outnumber atheists by a large amount, and are more powerful, so it is a statistical certainty that they will have committed more crimes. The problem with religion is not that religious people are worse. One of the conclusions I draw from what I have read on Epiphenom is that there is little difference.

  9. @ Tom Rees

    Has work been done to show that humans actually are more dangerous when they don't feel they are being watched by real people? I have always intuitively assumed they are.

    Pascal Boyer talks about the phenomena of ancestor spirits (not goods) used as threats of punishment for bad behavior as populations grew.

  10. People certainly behave nicer if they think they are being

    watched by other people - and that includes acting more

    trust-worthily. Simply sticking a pair of eyes above an

    honesty box
    is enough to do the trick.

    As to whether that also applies to supernatural watchers,

    well that's where it gets interesting.

    Jesse Bering did a study in which he told people the lab

    was haunted - they behaved more honestly as a result. And

    if you give people religious prompts, they also behave

    more decently (this applies to religious and non-religious).

    But when you look at societal level, well the countries with the the most atheists have the highest levels of trust.

    Now, there's all sorts of reasons why this might be the case. It might partly be the case that people don't think about god most of the time. It might be that people don't view themselves as cheaters - even when they are.

    But I think that the belief that people are inherently bad and will only behave when they are being watched is corrosive.

    It makes you think that everyone else is cheating - and one of the things that's most effective in encouraging cheating is the belief that everyone else does it. I don't think even a supernatural watcher would stop that (because you can rationalize that you are justified, under the circumstances, so the supernatural watcher would forgive you).

    So I think that encouraging the belief that people are inherently wicked and will take advantage of you if you are decent encourages a kind of race to the bottom.

    Of course, as has been popinted out above, there are lots of reasons for people to be honest. Evolution favours co-operators and straight dealers. People behave according to how they are trained as children. People can be helped to understand that trust inspires trust and helps you in the long run - so it's in everybody's interest to be more trusting.

  11. I have to say I love Epiphenom. Keep up the good work!

  12. Tom, fascinating reply. I think I found that even more stimulating than the original article.

    Great points about the race to the bottom and self-perception as cheaters. When I saw the topic of evolution coming up I was afraid it was going to de-evolve into the traditional non-starter over inclusive fitness and trivial reciprocity being the basis
    for altruism. I appreciate your richer perspective here.

    Thanks very much!

  13. Thanx, Tom, that was helpful.

    I was puzzled when you said,

    "But I think that the belief that people are inherently bad and will only behave when they are being watched is corrosive. "

    If it is TRUE but corrosive, what should we do with that?

  14. Sabio, I think it's one of those things that is true if you believe it is true. What I mean is that we are genetically enabled to both be honest and duplicitous, but which tendency wins out depends on culture and environment. There's a reason Germans pay their taxes but Greeks don't - and it's not genetics!

  15. As an atheist I am my own conscience. My morality is innate. I know I have to treat people the way I want to be treated if I'm to be able to live with myself without self-disgust. I've seen Christians throughout my life, steal on a regular basis, and treat people cruelly. A prime example that Christians don't care if their god is watching them is the Catholic Priest sex-abuse scandal. The Vatican who is god's spokesperson covers up these unforgivable crimes of child-abuse. Sorry, but being religious is the opposite of being good from what I'm seeing in society. I believe in the comment that, "A good person, will do good things, and an evil person will do evil things, but for a good person to do evil things that takes religion."

  16. I like to think I'm an intelligent chap. In the first example, I instictively went for b).

    But ask an atheist:
    Jim is a parent who teaches his children that homosexuality is wrong. Is it more likely that Jim is from Texas, or that Jim is from Texas and is Christian?

    I'm an atheist. It cuts both ways. This is fascinating stuff, and I suppose that's why the formulator is a Nobel prize-winning Psychologist.

  17. If you believe in the depravity of man then you naturally believe people who fear a final accounting will behave better than those who do not. On the other hand if you don't believe in a depraved humanity you have to extend a presumption of goodness to people you don't know.

  18. As someone who has been both, I've had the experience of being "lobbied" to change views by both sides.

    I would say the distaste that I felt in both circumstances did influence the trust I would place in the other (the fact that I happen to adhere or not adhere is not an invitation to a discussion in which you militantly try to convert me to the "other team".)

    I think atheists and religious believers would probably feel a great deal more at ease with the other if they weren't nervous about being "converted." Knowing someone has a point of view on the matter doesn't mean you're welcome to make the challenge of converting them your next personal project!

  19. For kicks, you should add "Objectivst" to the choices and see what happens. Objectivists are a type of atheist but are moral absolutists. Also, add "pagan." Then to top it all off, add "Hollywood celebrity." Now that would be a mighty interesting survey.

  20. Is it possible that atheists ARE untrustworthy? Has that been studied? Maybe a large proportion of these subjects' personal encounters with atheists have resulted in them being mistreated by the atheists, resulting in feelings that atheists in general are untrustworthy.

  21. This person is saying that Christians aren't as distrustful of atheists, and to conclude that from this study is wrong:

  22. Well, he says a couple of things. Firstly that you can't say atheists are less trusted than atheists. That's quite correct.

    The other is about an alleged ceiling effect. Well, such a thing could happen, but in fact this is just special pleading.

    For a start there's no evidence of a ceiling effect - the SD (a measure of the range of the results) is the same for all groups.

    Second, even if there was, there's no particular reason to assume that the rapist target would score higher, but not the atheist target. Perhaps it would be the other way around.

    In other words, he's seen two bars that are the same length, and not wanted to believe it could be true. So he has searched for a reason to reject the data.

    It's what often happens when people see data they don't like.

  23. I don't know if anyone will even read this, but I somehow just stumbled upon it after seeing that the comments section linked to my post.

    To respond to your comments Tom: I am not special pleading or responding to data I don't like. The data simply doesn't support the conclusion.

    Let's begin by acknowledging that the idea that atheists are just as distrusted as rapists is an outrageous, extraordinary, counterintuitive claim, and thus should require more evidence than usual to draw us to accept it. So given that, accepting such a claim based on such a bizarre measure should seem strange.

    Second, I am not positing a ceiling effect. The data doesn't control for one, so the data can be accounted for much easily by a ceiling effect. It's not up to me to show a ceiling effect exists. It's up to people making the claim from the data to show why this much more plausible explanation doesn't apply, or why atheists being similarly mistrusted as rapists should be the more plausible interpretation.

    Second, where do you see SD stats? There can't be a standard deviation for an yes-no question. Are you confusing it with the confidence interval?

  24. Hi Vlad, I get an email for each comment, so I read them all!

    First off, I agree that you can't conclude that atheists are as distrusted as rapists on the basis of this one study. It does have flaws, and anyway the result could easily just be a fluke. Having said that, I don't think this study is any worse than the others I cover (and is better than a lot of them).

    A perennial problem in the psychology of religion is that there are few people working on it, and results are almost never tested by other researchers.

    You're right, the error bars are the 95% CI. But the general point still holds - these reflect the within-subject variability. If there was a ceiling effect, then you would expect variability to be lower for groups with high means.

    Doesn't rule out the possibility, of course, but I see no particular reason to raise this concern for this study - over and above all the other studies I cover. Hence my point that to reject this one on the basis of a supposed ceiling effect is special pleading.

    Personally, I don't know whether this study shows that atheists are as mistrusted as rapists. (although, it would not occur to me that rapists are particularly untrustworthy - aggressive, callous, etc, yes).

    But I don't think that's the important finding of this study. Rather, what's important is that it helps elucidate why atheists are distrusted. They're distrusted for the same reason that criminals are distrusted - i.e. atheists are potential criminals, at least in the minds of US Christians.

    And that in turn fits in with a series of other studies done by Gervais - and others.


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS