Field of Science

Is this why atheists are, on average, more intelligent?

This post was chosen as an Editor's Selection for ResearchBlogging.orgThere's a new paper out by Satoshi Kanazawa which is causing a bit of a stir. You might have seen something about it already - I'm a little behind the curve on this one, but on the plus side I have actually read the paper, unlike many other pundits!

What's got people talking is the correlation between atheism and intelligence, although that isn't what the paper is actually about. It's already pretty well established that atheists tend, on average, to be more intelligent. This paper firms that finding up a bit more, but makes a bigger claim than that.

But before delving into the paper, I want to just cover a few commonly raised objections that tend to fly about whenever the intelligence-atheism link is raised.

Firstly, although there are many different aspects to intelligence (and intelligence tests), there really is such a thing as 'general intelligence'. People who score highly on one test will tend to score highly on others. That's statistically provable, and the only explanation is that there is some aspect of brain function that enables you to be generally good at tests of intelligence. Of course, there's a lot more to being a smart individual than general intelligence, but general intelligence is a real, measurable thing.

What is true is that the IQ test does not measure general intelligence (no test does, by definition). However, it does provide a good approximation. And in fact, Kanazawa's study does not use IQ tests, but other tests that are also related to general intelligence.

Perhaps more importantly, general intelligence has a genetic component and a substantial environmental one. And so any time anyone tries to use intelligence scores to make racial claims, you can probably disregard them.

So, with that in mind, what did Kanazawa find?

There were actually two studies, both using US data. The first looked at intelligence scores from a group of adolescents (junior high and high school), and compared it with their religious beliefs 8 years later.

The figure shows that atheists are smarter by a good few points on average. And the link remained even after Kanazawa corrected for age, sex, race, education, earnings, and even religion. It's not a trivial difference. In fact the effect is pretty strong - stronger than the effect of education, for example.

The second was from the general social survey - a survey of adults. Once again religion (belief in god and religious intensity) was strongly related to intelligence, even after correcting for a host of factors that you might think could explain the link.

So what? Well, Kanazawa believes that the explanation for the link lies in the Savannah hypothesis. This is the idea that general intelligence evolved as a way to deal with evolutionarily novel situations. It lets us transcend our evolved behaviour and do things that contravene our instincts.

In support of this, Kanazawa shows that intelligence is linked to liberal ideals in the same way. In particular, the link seems to be between intelligence and openness to support of people from other ethnic groups (i.e. whites supporting government intervention to help blacks).

What's more intelligence in those adolescents increases belief among men (but not women) in sexual exclusivity - i.e. that people should not sleep around.

If Kanazawa is right, then intelligence should not be linked to behaviour that is not evolutionarily novel. And, indeed, attitudes to children, marriage, family and friends are not linked to intelligence.

It's an interesting hypothesis, and an interesting analysis. To me, it seems intuitively reasonable that general intelligence should have evolved as a way of solving problems. But it will take more than this study (and his previous one) to convince me.

The problem is that intelligence and rational thinking are different beasts:

The idea that Bush is just one foolish smart person among many, and that intelligence is a poor predictor of "good thinking", comes from a series of recent experiments that compared the performances of people of a range of intellectual abilities on rational-thinking tasks. In a study published last year, Stanovich and Richard West of James Madison University in Harrisonburg, Virginia, found there was no correlation between intelligence and a person's ability to avoid some common traps of intuitive-thinking. (New Scientist)
You see, intelligence only helps overcome the cognitive biases that lead to poor judgement if you are the the kind of person who uses it. It's not so much about innate braininess, it's about how you approach the world.

But all this does is explain why some intelligent people hold irrational beliefs. Kanazawa's work does help explain why some people manage to shake them off!

ResearchBlogging.orgKanazawa, S. (2010). Why Liberals and Atheists Are More Intelligent Social Psychology Quarterly DOI: 10.1177/0190272510361602

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


  1. Does the study have a title? I'd like to read it as well, thanks.

  2. Did you see the hyperbole on Pharyngula?

    Paul, the reference is at the end of the post.

  3. I disagree with Satoshi's theory, and instead place his measure of "intelligence"--which is really verbal ability, same as in my study---as the dependent variable predicted as a negative function of religious fundamentalism and ties to sectarian religious groups. However, I think Myers is out of line. Satoshi is a strict evolutionary psychologist, and at least he's consistent! Myers long ago went out on a limb to argue (without access to either data or theory) that there is no relationship between religiosity and intelligence; and now he's defending his claim from empirical evidence which seems to falsify it. Baseless claims shouldn't be backed by ad hominem attacks.

  4. Bjørn: yes, I did. I thought it was fascinating. PZ dismisses the study without reading it and by simply launching insults at the author. Somehow this got converted in the minds of his fans as a repudiation of the study! A wonderful example of cognitive bias in action!

  5. Darren, for some reason I thought that the Add-Health study used a pictographic test, but now on re-reading I see you are right - it's a verbal test (the 'Picture Vocabulary Test'). So the results do dovetail with your own quite nicely. One difference is the marked effect even among young adults who report low levels religiosity.

  6. Paul, you can download the paper from Kanazwa's homepage here.

  7. Correlation does not prove causation, and I don't think that statistically correcting for a hodgepodge of other factors constitutes a convincing argument that a causal link exists.

  8. Tom, I would (also) classify PZ's post under 'moralistic fallacy'. It's fascinating, but sad, to see it in action.

  9. kanazawa has a tendency to really go off the deep end. that's a real issue. but, this isn't that novel of a finding, and that's the bigger issue IMO.

  10. I think part of what motivated PZ's diatribe was that it really does come across as immature and almost patronizing when us nontheists run around saying, "See, this study proves we are so much smarter than you!" (e.g. The head of a local atheist group, who also happened to be one of the primary organizers of last year's AAI conference, posted the study to the mailing list he runs with the subject "Something to shove in the theists' face" -- I'm not one for a mindless adherence to politeness, but even I found that a bit unseemly)

    Of course it is fallacious to go from "nontheists shouldn't run around chanting nanananabooboo over this study" to "the study is wrong"... and as such I am really glad Tom looked at this paper. I don't have a huge amount of science background myself (I'm simply a lowly engineer) so I have to just trust what the "experts" say.

  11. Still wish I could read the study but damn it if I have to pay for it.

  12. More intelligent, but socially more inept? If you spend a good amount of your life thinking of spiritual things, of course you'll average a bit lower on an IQ test. Not a bad trade-off I'd say. But, then, I'm just a lowly research scientist. David Mc

  13. Anon said "If you spend a good amount of your life thinking of spiritual things, of course you'll average a bit lower on an IQ test"

    Depends on how you define spiritual. Atheists are not spiritists and are not superstitious (both concepts are often confused with spiritual), but in the US a lot of them have an interest in values and ethics and philosophy (which some would shorthand as "spirituality").

    It's amusing how when "polyamorists" try to circle around atheist groups they call themselves "ethical polyamorists" trying to appeal to our basically very ethical predispositions. Unfortunate for them, there is evidence that the highly intelligent (especially men) tend to be both atheistic and monogamous. I'd say to them: "Sorry, pal, try Craigslit."

  14. I think your criticism of PZ Myers is unjustified. Sure, his style leaves little option other than liking it or getting used to it, but his criticism is valid. If you look into the links he puts in the post you see why he dismisses it, as I imagine most evolutionary biologists do. And it's not because of moralistic bias...

    Here are some examples of the objections that an evolutionary biologist might have:

    It's not a problem of the validity of IQ measurements per se - I happen to agree that IQ tests are measuring something, we just don't quite know what - it's the fact that the evolutionary analysis is ghastly... no, I take that back - it's completely absent. For a real actual evolutionary researcher to publish something as conclusive as Kanazawa's evolutionary "analysis", putting such causal claims forward, even if they were just suggestions, it would require HARD WORK over a long time and a substantial body of evidence to back it up. Show me a study of... I dunno, a selective sweep over a chromosomal region, you don't even have to have a candidate gene, just show me some substantial evidence, not statistical finesse and wild conjectures.

    As it reads right now I can only qualify Kanazawa's main conclusion with your own words - intuitively reasonable - at best, but most of us need quite a bit more than that. That's what science is for, right!? So that we don't have to rely on intuition. Evolutionary research is hard! If you work in evolution and you know the amount of work and effort required to provide a strong line of evidence, I completely understand if the expletives come tumbling out when you read something like this.

    In general these evolutionary psychology arguments are incredibly reductionist and treat causality like putty. I think PZ's question is perfectly justified, how does a 6 point IQ difference make any difference? How would one go about proving it in the first place? Kanazawa's description of the different situations an early human might find him/herself in on the African savanna are giggle-worthy. How does he know this stuff? I could go on making examples of the wild conjectures. Like the argument that "less intelligent people" have more children because they haven't adapted enough to the evolutionarily novel invention of contraception. It's laughable!

    He also mangles behavioral genetics while he's at it. Traits don't evolve separately in a vacuum.

    The whole thing brings to mind Jacques Monod's famous quote: "Another curious aspect of the theory of evolution is that everybody thinks he understands it. I mean philosophers, social scientists, and so on. While in fact very few people understand it, actually as it stands, even as it stood when Darwin expressed it, and even less as we now may be able to understand it in biology."

    In my opinion Kanazawa's study could provide an interesting representation of an interesting statistical phenomenon, that's about it. I'm not sure enough of the statistics and the error margins to comment on the validity of the actual results, but the evolutionary discussion carries little weight.

    You do write that it will take more studies to convince you and I appreciate your collected analysis of the paper. I just wanted to offer an evolutionary biologist's perspective on why PZ Myer's, and others with him, might be mad. ;-)

  15. If PZ had made the argument you just did, I would have no problem! But in fact he wasn't criticizing Kanazawa's theory. He dismissed out of hand the link between IQ and atheism. It was a silly polemic.

    But the link is there and it has to be explained. You can reject Kanazawa's hypothesis for the link, but the rejection needs to be evidence-based!

    As for whether the link is meaningful, as well as significant, well I think it probably is (see Razib's post for why).

    So, given that there is a link, can Evo Psych explain it? Now, I agree with your complaints. Evo Psych does have a history of making bold claims with little evidence (and people, especially the media, latch on to them because they sound right - i.e. due to confirmation bias). There's a tonne of daft stuff churned out in the name of Evo Psych, and like you I love it when the gurus get cut down to size.

    All that is true.

    But that accusation can be levelled at social psychology in general. The problem is that it's fearsomely difficult to prove anything in social psychology - the number of uncontrolled variables is so great. Often all that can be taken is tiny steps at a time, rather than sweeping, dramatic proofs. Nevertheless, these are important issues and they are worth researching.

    To be fair to Kanazawa, he has at least done some research to support his hypothesis. OK it's only a tiny step, but if tiny steps are all that can be taken, then that's not a reason to attack it. What would be good is to come up with a way of testing the hypothesis more rigorously. And, even better, do it :)

    I hadn't spotted the line about contraception tho. Now that is funny!


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