March 11, 2010

"Why do we believe", and are atheists really more intelligent? editor Dave Munger has written an article for SEED magazine entitled "Why do we believe". The article summarizes recent blog entries regarding studies on the origins of religiosity. It's really worth reading to get a good overview of the subject, and what do you know he links my entry on god's will and beliefs in it.

Among the studies that are mentioned is a controversial study entitled "Why Liberals and Atheists Are More Intelligent" (link at the end of this post).

Medical writer Tom Rees devotes his blog Epiphenom to the scientific study of religion. Last week he examined a study on the relationship between intelligence and religious belief. Published in Social Psychology Quarterly, this study by Satoshi Kanazawa replicated the results of several earlier studies in showing that strong religious belief was correlated with lower intelligence. In this case, adolescents who scored higher on intelligence tests were less likely to be religious as adults.

But Rees says Kanazawa’s study goes beyond those earlier studies to arrive at a potential explanation of why less-intelligent people are more religious: Intelligence evolved in order for people to adapt to novel situations.

You should go over to Epiphenom to read a summary of the study as well as my commentary on it, posted as a blog comment. In summary: I don't think it's very good. Kanazawa's evolutionary argument is completely based on some pretty wild conjectures and lacks any sort of empiric support. His argument is at most "sorta reasonable", but we must do better than that surely. For evolutionary researchers that have to spend considerable time and effort gathering a solid line of evidence, this sort of jumping to conclusions can be a tad annoying.

Also, let's remember what kind of forces we're dealing with here, how evolution through natural selection actually acts and what it acts on. Even if we assume first, that intelligence tests do measure some sort of good approximation of intelligence, and second, that results gathered today actually reflect a past situation; what difference do a "good few" average points make for survival? Any conclusions made from the correlation between higher intelligence, as measured by intelligence tests, and atheism are only significant within a larger evolutionary and functional neurobiological context. So take claims that atheists are more intelligent (on average) with a considerable pinch of salt.

Read more on the subject on Epiphenom, here and here.

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

Swedish blog tags: , , , ,
Technorati tags: , , , ,


  1. So more intelligent people are liberals and less promiscuous at the same time, extremely contradictory.

  2. Hi Daniel,

    funnily enough I just read Kanazawa's study today and then stumbled upon your blog post here (and read your critique on epiphenom). I am a bit surprised at your sweeping dismissal of the author's claims.

    The question "how does a 6 point IQ difference make any difference?" seems - to the untrained eye - to be a valid criticism of Kanazawa. But it's actually quite easy to answer such a question. We work with these kinds of questions all the time, deriving p-values and determining if a difference is statistically significant, or not. I'm not saying that Kanazawa's difference is statistically significant, I'm saying that there is a way to answer these kinds of question, and it's a well tested way used in all areas of science. If it is statistically significant, that fact is independent of the above question ;-)

    Secondly, I think it would be extremeley hard to link a complex neurological characteristic, even less a psychological trait, to any gene, genomic region or even chromosome. I find it is a bit unfair to ask that of Kanazawa when "real" biologists have such a hard time to link genes to much less complicated phenomena. If K. came up with an article "the atheist gene", I think many people would be extremely suspicious (as with the "gay" gene and other such genes resulting from oversimplifcation of study results).

    So I guess what I want to say is that, even though I haven't checked the data behind his paper, I think that is - as you put it - presenting an interesting statistical phenomenon, and providing an interesting theory for its explanation. As with all theories, it will have to be tested and tested again, but we shouldn't dismiss it either because it has not yet been tested enough, or because we think all research should be performed in the way a biologist would do it.

    Take care!


  3. Hi Florian!

    I might leave a longer comment later, I just wanted to note that of course I'm not referring to statistical significance - I wrote ''what difference do a "good few" average points make for survival?'' So I'm referring to the individual. How much of a difference do a few IQ points make for the fitness of an individual? To make such big causal conclusions from differences that are probably not biologically relevant, although they might be statistically significant, seems hasty and reductive to the extreme.

  4. Hmmm, ok. But how would you determine that. In the end it always comes down to minor advantages that are then subject to selective pressure. We might as well ask "what good do a few percent of increased night-vision do a nocturnal predator", might we not? The accumulation of these small changes over time results in a shifting of the population mean (and median) towards the "better" value. Even though at any given point in time the variance within the population might seem small, the variance between the median "back then" and the median "today" is probably far greater, if the respective trait was under selection.

    You are right that it would be hasty to conclude a causal relationship from correlation alone. It might be hard or impossible to prove beyond a shadow of doubt that the conclusions suggested in the paper are correct. But I think K. does not assert as a fact that his conclusions are proven, he suggests them as a possibility, constructs a theory and is able to at least give a first supporting evidence from what statistical data he has available.

    What I see as the problem with such studies is rather that people oversimplify and say, "Aha! I am an atheist, THEREFORE I am more intelligent than you religious people." Which is not what K is saying and which is not what we should be apologizing for, because we are not making that argument.

    In his recent paper "Why night owls are more intelligent" (which is a pretty misleading title, I find), K. argues -using the same data by the way- that people with lower adolescence IQ tend to rise and go to bed earlier than "very bright" persons... So now I can sleep in and feel good about it ;-)

    I'm looking forward to alternative analyses of that data and alternative theories as to where these correlations come from. Maybe there are hidden variables that might explain the effects. (For instance that "less bright" people tend to be able to live in lower-rent suburbs and thus have a longer commute every morning, etc).

  5. There is NO dogma in science. Scientists always debate their theories rationally. They NEVER resort to violence. This seems to be a sock in trade of Religion.

    Christophe Luxenberg has written a scholarly book on the role that Syriac played in the devopment of the Qu'ran (and incidentally of the Arabic language itself). He describes it as "Dunkel und mehr deutig" Opaque and ambiguous.

    Is he right? I think he puts up a good case. Any scientist who disagreed with him would collar him at a conference and debate RATIONALLY. If you are religious you threaten to kill him.

    You might in fact say that the real conflict between Science and Religion was very physical. The technology of such things as drones

  6. I am sure that there are plenty of theists who are smarter than I am. What, though, of gullibility and the willful ignorance of those who use their intelligence to construct grand and useless apologia to defend that gullibility?

    There should be a test to derive a gullibility quotient. You can be extremely intelligent and very gullible simultaneously. Atheists can be very gullible, but, it's one thing to be taken in by a con-man or infomercial that one can see and that seemingly demonstrates the product and its usefulness; quite another thing to worship some other person's invisible god-concept, especially to the point of wasting time and money or even killing and dying for it.

  7. @Florian I'm reading a bunch of other articles and commentaries on the subject to sort of "close the folder" on this in my head... and I've thought about what you wrote above as well so I'm thinking of writing a separate post about it rather than a long comment here. But I still wanted to address something in your comment.

    I think that we're recognizing the exact same difficulties with the issue. Of course I completely agree with you about "the atheist gene" and so on, it's obvious that what you point out is true... and if you ever wanted to find a complex polygenic trait, this is probably one of the major ones.

    BUT! I'm not saying that Kanazawa, and others who make similar discussions, simply have neglected to do something, that they should use an alternative method or something like that. I'm saying that Kanazawa has not acknowledged the immense difficulty and the considerable amount of evidence that is required to be able to come to such a conclusion as he does. The fact that it's immensely hard, or even unlikely, is exactly my point.

    I think this is problematic since, like you say, it makes it more likely for it to be picked up in the wrong way by the media, like it certainly has in this case. For instance this article on National Geographic.

    It's also problematic because it gives the wrong idea to the public about which questions evolutionary biology can answer and which answers are more speculative.

    It's alright to be speculative... and Kanazawa has at least done some work to approximate the problem, but there's a place and a time to be wildly speculative and a research paper is not the best place. At least not in my opinion. Maybe a book or a talk at a conference is better. The fact is that Kanazawas article is incredibly sensationalistic and gets the evolution wrong... and now I'm discussing what I think is wrong with it, as is the custom in the scientific world, so I guess everything is in order.

  8. Daniel, thanks for your response. I'm looking forward to your next post on the subject. I agree that we are in agreement about the article :-). He could have been less definite in his assertions, for sure.
    I'll leave it there and wait for your next post.
    Take care!


Note: Only a member of this blog may post a comment.