The myth of the smarter Atheist

Ever since I published my previous article on the average IQ of students by college major, I’ve received several requests to analyze the correlation between IQ and religiosity. Below, I’ve written up an analysis of the existing published literature on the topic. I hope this serves as a springboard for future conversation on the topic — and hopefully puts some tired myths to rest.

Country-level evidence

One of the few peer reviewed scientific articles I could find on the topic had the straightforward name, “Average intelligence predicts atheism rates across 137 nations.” This article analyzes IQ and religiosity data from “137 countries that represent 95% of the world population” and claims to show that nations with more intelligent citizens tend to have more Atheists. I plotted the data from that article below.

iq-vs-religiosity-worldwide

(for the statistics nerds, the R^2 on a linear regression = 0.352)

What we see is a fairly weak relationship between national religiosity and average national IQ. Once we get up to about 20% of the population being Atheist, the IQ of the population flatlines at around 100 from then on. Even worse, in the ~0% Atheist range, there’s a wide range of national IQs from 64 to 100+ — with a cluster of low-IQ nations that appear to be driving the “trend.”

If we focus on the lowest IQ nations in the above chart, we notice that several of them are poor nations in e.g. Africa. That led me (and others who have reviewed the topic) to wonder whether the wealth of a nation better predicts the average intelligence of its citizens. After all, the more wealthy the average citizen is, the more time they have to dedicate to intellectual pursuits.

iq-vs-gdp-worldwide

(R^2 on a linear regression = 0.449; income per capita data from GapMinder)

Indeed, if we look at income per capita instead of religiosity, we already see a much better correlation with average IQ. The correlation between religiosity and IQ is too weak to suggest that religiosity predicts intelligence on the national level. Anyone who claims otherwise is grasping at straws.

Individual-level evidence

Looking at national data is fine and dandy, but what about individual-level data? In another controversial research paper on the topic, Satoshi Kanazawa claimed to explain “Why Liberals and Atheists Are More Intelligent.” In this article, Kanazawa analyzed data from a longitudinal study following students from middle school through adulthood. Conveniently, the study includes information on how religious the students are and the results of a standardized IQ test. Kanazawa mashed those variables together and claimed to show that Atheists tend to be smarter. I created a less misleading version of Kanazawa’s plot below; judge the data for yourself.

iq_vs_religiosity_ind

This is where we have to think about effect size vs. statistical significance. The most religious adults had an average IQ of 97.14, whereas the atheist adults had an average IQ of 103.09. That may seem like a wide gap — 6 whole IQ points — until we remember that anyone in the IQ range of 90-109 is classified as having “average intelligence.” Thinking about this in practical terms: Would you be able to tell the difference between someone with a 97 IQ and someone with a 103 IQ? It’s highly unlikely.

So really, all Kanazawa showed is that the average person has average intelligence regardless of how religious they are. I’ll leave the discussion of why this guy’s work was published in the first place for the comments.

Conclusion

The take-away message? To my knowledge, no amount of research has shown that Atheists are notably smarter than highly religious folks. It’s time we put this myth of the smarter Atheist to rest.

Dr. Randy Olson is a postdoctoral researcher at the University of Pennsylvania. As a member of Prof. Jason H. Moore's research lab, he studies biologically-inspired AI and its applications to biomedical problems.

Posted in data visualization, review Tagged with: , , , ,


  • http://pleiotropy.fieldofscience.com Bjørn Østman

    Previous post on the paper: http://epiphenom.fieldofscience.com/2008/10/atheists-are-more-intelligent-but-does.html

    I never had much patience with the straightforward conclusion. Not only for the weak correlation among the “atheist” countries, but because IQ is supposed to ideally measure general intelligence (g-factor), and the idea that there are brains that have a high value of this and low religiosity vs. log g and high religiosity never sat well. If would mean that the connection between upbringing and religiosity (which I understand to be solid) somehow connected on a larger scale with intelligence, and that just seem implausible to me.

    That is not to say that I don’t think that it is stupid to be religious, and that a higher intelligence makes it more likely a person will break free of religion. I do. Sue me.

  • Josiah

    You mention effect size and statistical significance… but don’t calculate either, and then dismiss the findings. It really doesn’t matter whether the group means of both groups fall within the band that someone arbitrarily decided is the “average” (defined as what? 1 SD from the grand mean of IQ)?

    What matters is whether there is a significant relationship between religiosity and IQ and, as you mentioned, what the effect size is. But you haven’t provided those numbers.

    • http://www.randalolson.com Randy Olson

      Let’s just compare the means. Would you be able to tell the difference between someone with a 97 IQ and a 103 IQ?

      • Victor

        Sure I will by asking them to check IQ test.
        And if you’re asking about purely observational difference, than you’re assumption that nobody want is also incorrect.
        Statistically, we will. Definitely you should be able to realise difference between two crowds made of 103 and 97 IQ individuals.

        Imho, this post is non-scientific bullshit

        • http://www.randalolson.com Randy Olson

          Imho, this post is non-scientific bullshit

          Reviewing published scientific research is non-scientific? This is news to me… :-)

          • Preston Oneal

            Yes, just like when 9/11 conspiracy theorists try to talk about physics.

      • sup

        Can you tell a difference between the median teacher and the median mechanic? Some Googling suggests that the IQ differences between the two professions is ~6 IQ points.

        Further, although Ashkenazi Jews account for 3% of the US population, they represent some 27% of Nobel Prize winners — and what’s their median IQ? ~107.5.

        • http://www.randalolson.com Randy Olson

          Some Googling suggests that the IQ differences between the two professions is ~6 IQ points.

          Let’s stick to published data on this topic, please. “Some Googling” is not passable.

      • Josiah

        Without knowing what the distribution of scores for IQ are – I have no idea. i.e. how many SD is that difference? The scores themselves are inherently meaningless.

        Regardless, I think that calculating the actual effect size would be helpful, and should be pretty simple with the raw data. Do you want to tackle it, or do you know if the raw data is available somewhere?

        • http://www.randalolson.com Randy Olson

          I believe the study data is here. Please link me/comment with any findings you make.

  • Eric Wilson

    Thanks for analyzing this.

    As a Christian, there is a temptation to fight the myth, and to retweet these results. See! Christians aren’t dumb!

    But it is much more fitting to accept that I am nothing special, and to keep in mind that God prefers to call those that are not “wise according to the flesh” so that He can be more glorified.

    For you see your calling, brethren, that not many wise according to the flesh, not many mighty, not many noble, are called. But God has chosen the foolish things of the world to put to shame the wise, and God has chosen the weak things of the world to put to shame the things which are mighty; and the base things of the world and the things which are despised God has chosen, and the things which are not, to bring to nothing the things that are, that no flesh should glory in His presence. But of Him you are in Christ Jesus, who became for us wisdom from God—and righteousness and sanctification and redemption—that, as it is written, “He who glories, let him glory in the LORD.”

    –1 Corinthians 26-31

    So I actually have some agreement with Bjorn, but I would express it differently. I believe, based on God’s revealed word, that intelligent people are more likely to rebel against their Creator to the point of denying His existence.

    I will pray for you, Bjorn, knowing that as long as you live, you have opportunity to turn to Jesus in repentance.

  • nobody

    >That led me to wonder whether the wealth of a nation better predicts the average intelligence of its citizens.

    You’ve got your cause and effect backwards here, my friend

    • http://benjamindavidsteele.wordpress.com/ Marmalade

      The two contribute to each other. Better environmental conditions have been proven to increase IQ, and better environmental conditions tend to correlate with greater wealth (especially in terms of lower rates of economic inequality and higher rates of economic mobility). But as you argue, the opposite direction of causation is also likely to be true. A more intelligent population relates to all other kinds of positive attributes: higher education rates, stronger secularism, more technological advances and innovation, etc.

  • http://mchasewalker.wix.com/michael-chase-walker Michael Chase Walker

    I’m sure in other models, the findings might even determine some religious cultures e.g India and Israel, have a disproportionate number of Nobel laureates. Jewish Nobel Prize Winners http://www.algemeiner.com/?p=143785 . The study then might be refined to those religious cultures that promote science over faith, scholarship and critical thinking over rote learning. Does this mean that some traditionally religious cultures promote higher critical thinking as opposed to those that stress faith over knowledge, e.g. Spain, Italy, Saudi Arabia? Faced with problems like medicine, space travel, rocket science, nuclear fission, etc. would you select your team of scientists from Liberty College, or M.I.T. ? If your child was fatally ill would you turn to Oral Roberts or Johns Hopkins? You might find candidates of equivalent I.Q. in either, but how individuals apply that intelligent advantage becomes a qualifier in and of itself.

  • http://www.4simpsons.wordpress.com eMatters

    I realize this is a very small sample size, but my two conservative, Bible-believing Christian daughters just graduated from college. One had a 3.97 from the Honors College of a major state institution (in only 3 yrs). The other got a 4.0 from a private school. They have excelled at every standardized test they every took and got A’s in every science class.

    Also, with respect to the “intelligent atheists,” I love what this quote from J. Budziszewski, atheist-turned-Christian philosophy professor from the University of Texas: “Though it always comes as a surprise to intellectuals, there are some forms of stupidity that you must be highly intelligent and educated to commit. God keeps them in His arsenal to pull down mulish pride, and I discovered them all.”

    • http://www.cesletter.com Filmmaker30

      Hahahahaha. LMAO! The irony of this post is incredible. “We ain’t dumb, look at my bible thumping daughters in da college programs.”

      • Nathan

        WOW. And you atheists espouse “tolerance” and “acceptance”. What an absurd reply.

        • StupidShouldCausePain

          What do you call this, asshat?

          Though it always comes as a surprise to intellectuals, there are some forms of stupidity that you must be highly intelligent and educated to commit. God keeps them in His arsenal to pull down mulish pride, and I discovered them all.”

          • Ronky

            I call it self-awareness.

  • http://about.me/paul.a.rubin Paul Rubin

    Two questions:

    1. Why are there countries with average IQ as low as 70? Might someone be conflating education level with IQ?

    2. I wonder if government might be a meditating variable? Suppose that the intelligence/education level of the populace of a country correlates negatively with its propensity to have an absolutist/dictatorial government. (I have no data on this, but it seems plausible that less intelligent citizens would be more gullible, more susceptible to propaganda, or perhaps just happier to be told what to do.) Dictatorial governments frequently appeal to religion for a mandate for their actions, which can make the country uncomfortable for atheists, who may then either leave or decline to admit their atheism.

    This is all speculation on my part, but it might make for an interesting follow-up analysis.

    • http://www.randalolson.com Randy Olson

      1. Why are there countries with average IQ as low as 70? Might someone be conflating education level with IQ?

      I wonder if there’s sampling issues. One criticism of that data set is that it’s based on a fairly small sample in some countries, sometimes as few as 10-20 people.

      2. I wonder if government might be a meditating variable? Suppose that the intelligence/education level of the populace of a country correlates negatively with its propensity to have an absolutist/dictatorial government. (I have no data on this, but it seems plausible that less intelligent citizens would be more gullible, more susceptible to propaganda, or perhaps just happier to be told what to do.) Dictatorial governments frequently appeal to religion for a mandate for their actions, which can make the country uncomfortable for atheists, who may then either leave or decline to admit their atheism.

      A counterexample to this would be China, which is quite dictatorial but still Atheist. Regardless, I agree: this would be an interesting trend to investigate! The big question is how one would quantify “dictatorialness” of a government.

      • http://about.me/paul.a.rubin Paul Rubin

        With sample sizes as low as 10-20, you have to wonder if there’s some “convenience sampling” going on (institutionalized individuals, prisoners, political office holders) that might bias the IQ value downward.

        I agree that measuring “dictatorialness” might be tricky. Absence of an elected body of representatives would be one indicator, but not the only one. Propensity of critics to be put on trial or to simply “disappear” would be another, but would be hard to get data on. A measure of human rights (or human rights violations) might make a plausible surrogate.

        • http://www.randalolson.com Randy Olson

          Hmmm… perhaps the World Freedom Index could be used?

          • http://about.me/paul.a.rubin Paul Rubin

            World Freedom Index might work. The scale is a bit coarse, but probably good enough to tell if there’s a mediation effect.

      • Josiah

        China is not even mostly atheist, according to most of the polls I just looked at.

  • steve25

    As an atheist, I completely agree with your premise. Atheists are not atheists because we are inherently smarter people.

    We’re atheists because we were able to take a step back from our emotional attachment to myths, and think for ourselves, “Hang on a minute, does what I know about the world actually align at all with what I’m reading in this book?” That’s not something that can be measured with IQ or standardized testing.

    Also, some very, very smart people can believe some very silly things, if they are presented emotionally, rather than rationally. The smartest person I’ve ever met was a Mormon physics professor. Brilliant dude in the classroom, but he believed some unbelievably silly things at church.

    • CO

      so dark matter is a myth because its not proven yet?

      • steve25

        Dark matter is more of a mystery than a myth. We know something is there; it has a measurable gravitational effect. But nobody has any idea what’s causing it.

    • andy_the_bear

      There are silly Theists who have an overly simple view of God and who think science is nonsense, and silly Atheists who think that philosophy is nonsense and science replaces it. But that does not mean Theism or Atheism is “silly”. But perhaps you mean “silly” in a different way?…I am suspicious that there is an irrational emotional appeal to materialistic Atheism that arises from intuition related to our senses. We have a kind of “horse sense” that tells us that reality is like what we sense it to be, and it makes the idea of some huge transcendent super natural agent with aseity seem “silly”. On the other hand modern science has shown this “horse sense” assessment to only be accurate within a limited scope. As we move to small things at the quantum level and very large things at the cosmic level, it has proven to be less and less useful.

  • Gandalf Olorin

    ‘To my knowledge, no amount of research has shown that Atheists are notably smarter than highly religious folks.’

    http://www.anth.uconn.edu/degree_programs/ecolevo/divineintuition.pdf

    http://www.sciencedaily.com/releases/2012/04/120426143856.htm

  • Terry Chesnut

    Correlation is quite low in all the studies with a minimal difference shown. More likely if you look at other factors with higher correlations we see those with more money are less religious as expected as they trust the wealth and think they have no need of God. Those with higher IQ are shown to earn more as expected. If you want to do real science you would have to control for wealth and many other things as variables other wise you risk publishing spurious correlations as proof of nonsense. A good example is increases in the money spent on science correlate with increases in suicide rates at r=0.99789 almost perfect correlation but only a moron would believe spending on science caused suicides. Useless study very bad science by all measures.

  • StupidShouldCausePain
  • Naz

    I’m sorry, but all of this data points in favour of the argument that Atheist have a higher IQ on average. In fact, isn’t that exactly what the last graph is showing? However slight the difference may be, the graph does show a very clear trend. It’s so perfect I would be worried it had been made up. Just because each group is classed within average intelligence, doesn’t mean that there isn’t a difference within that range. A really bad analysis from my point of view with very little evaluation. I have genuinely seen a more in-depth conclusions written by my year 7 students.

    • Christopher VanHimbergen

      The point of the average IQ range is that anyone taking the IQ test could get anywhere within that range each time. There is a range because I eliminates luck and sheer guessing on the IQ tests. The idea here is that a person would be able to score a 96 the first time, a 100 the next time, and a 98 the third time and be classified as average intelligence. A person who gets a 64 on an IQ test one time, a 62 another time, but a 108 the third time is not considered to have average intelligence. As a matter of fact, I am willing to bet that if the experiment were to be re done with the same number of atheists as theists, the results would reverse with theists being 6 points higher than atheists. Consequently, if the experiment were to be done a third time, the results may be more equal such as theists and atheists both having an average of 100 IQ points. This article definitively proves that there is absolutely no correlation between religiosity and intelligence because there are too many other factors that make it impossible to isolate just these two.

    • Xiao Dong

      Vegetarians who consume dairy also have an i.q. 6-7 points above the average while vegetarians who avoid dairy have an i.q. that is 5-6 points below the average. Does that mean eating animals is stupid? Minor differences cannot be used to ascertain truth.

      • Ray

        Wow, that’s one of the dumber arguments I’ve ever read. Where do you place on the IQ scale?

        • Xiao Dong

          I was in the gifted program in high school which is for people who score in the top 1 percent on i.q. tests.

          • Ray

            Gifted with stupidity more like

    • anmol13111993

      What if the sample sizes of Atheists and religious people were not the same ? At any rate, if the mean is shifted above the population mean, it means that there is a greater proportion of atheists that would be classified as ‘high intelligence’ (IQ 130+) assuming the variance doesn’t change

  • WreckTom

    If only based on averages, the correlation is crystal clear.

  • Zachary Bower

    Except the argument wasn’t that atheists are geniuses on average, just that they have higher IQs. And as others have pointed out, this has been found by a number of studies. I agree that “atheists are inherently smarter” or “atheism makes you smarter” is a myth, but there is a clear thing here, & it is worth understanding.

    Now, this is just my speculation, but I think this is because atheists are regularly expected to refute stereotypes about ourselves & justify our lack of beliefs. It just seems obvious to me, it’s like a student whose parents have really stringent standards about her grades & how often she studies, of course she’s going to score higher than the other kids. It doesn’t mean the other kids are “stupid.”

  • John

    Randy, you are using a range of data that is considered “normal” (90-109) and comparing it to a higher AVERAGE of another piece of data. If you want to compare apples with apples, it is easy to identify that the the AVERAGE of the highly religious would be on the LOW end of normal for atheists. In addition, atheists on the higher end of normal for atheists, would be considered ABOVE average intelligence of those who are religious. This relates to approximately 6 IQ points or 1/3 of atheists who are considered “NORMAL on the atheist scale, would be above average on the religious scale. This is not an insignificant statistic as I am sure you are aware!

  • Ky

    But… You just proved it. I’m so confused. You literally just proved your point. Soooooo… Argument done. Atheist have a higher intelligence. You can argue, “6 points,” but, yea, 6 points.

  • Shadowbalde

    This analysis appears to be not only spurious but also written from a religious apologist perspective.

    http://www.independent.co.uk/news/science/religious-people-are-less-intelligent-than-atheists-according-to-analysis-of-scores-of-scientific-8758046.html

    “Atheists are probably more intelligent than religious people because they benefit from many social conditions that happen to be correlated with loss of religious belief.”
    https://www.psychologytoday.com/blog/the-human-beast/201005/the-real-reason-atheists-have-higher-iqs

  • Arch Hold

    This is one of the shoddiest statistical analyses I’ve seen.

    You brought up the R^2 coefficient when a linear fit through that data is a ridiculous choice, as the trend looks logarithmic even just on inspection.

    You also say that the average IQ of groups in the last chart is not demonstrative because the confidence bound for an INDIVIDUAL is 90-109? The reason the confidence bound for an individual is so wide is to remove the obvious uncertainty of a test on a single person. When you take the average this obviously no longer applies, and a 6 point difference is highly indicative. The chart shows a blatant trend you are trying to ignore for some reason or other.

    Perhaps you’re religious, and have a reason to feel uncomfortable with the data, or perhaps your ability to analyse is evidence against the premise of your article.

  • Harry

    The fact is you have to be pretty stupid to disregard proveable scientific theories and instead believe everything was created by an unproven man in the sky. There will be intelligent religious people, and unintelligent athiests, but on the whole, athiests are going to smarter

  • Valdemar

    Normaly 6 points is nothing but sines so many People were in this Examination it actually meens that atheist are more intellegent

    I used to belive until the day i realisted how absurd it Was

  • vitalstatistic63

    This writeup seems like the work of someone trying to massage their findings to suit their beliefs.

    In the IQ vs Intelligence graph, I can see a clear pattern that supports higher intelligence athiests. Sure there are data points that jump away from the curve, but the overalll trend surely shows the link.

    Then the OP shows a graph of IQ vs Wealth which also seems to show a trend increasing in synch, but with a general fog of data points all over the place. But this data is somehow more acceptable than the earlier graph.

    Then the third chart once again shows a clear correlation between IQ and Atheism in individuals rather than nations, and this is dismissed due to the general belief that a range of scores are considered as average in a population sample. Well what a load of bollocks. The trend is still there, regardless of what is considered average. And let’s not forget that the data making up those averages is taken from a large population sample each with individual scores that are not shown but may vary by a large degree from the average result shown.

    These statistics cannot be dismissed on the whim of someone with a preconceived idea of what the results will show, as seems to be the case here.

    Further, why does Randal S. Olson require me to verify my email address before posting? I’ve already had to loging via my Disquss account or by Facebook. Is their account validation process not up to your standards Randall S. Olson? You sound like a bit of a wanker to be honest.

  • vitalstatistic63

    This writeup seems like the work of someone trying to massage their findings to suit their beliefs.

    In the IQ vs Intelligence graph, I can see a clear pattern that supports higher intelligence athiests. Sure there are data points that jump away from the curve, but the overalll trend surely shows the link.

    Then the OP shows a graph of IQ vs Wealth which also seems to show a trend increasing in synch, but with a general fog of data points all over the place. But this data is somehow more acceptable than the earlier graph.

    Then the third chart once again shows a clear correlation between IQ and Atheism in individuals rather than nations, and this is dismissed due to the general belief that a range of scores are considered as average in a population sample. Well what a load of bollocks. The trend is still there, regardless of what is considered average. And let’s not forget that the data making up those averages is taken from a large population sample each with individual scores that are not shown but may vary by a large degree from the average result shown.

    These statistics cannot be dismissed on the whim of someone with a preconceived idea of what the results will show, as seems to be the case here.

    Further, why does Randal S. Olson require me to verify my email address before posting? I’ve already had to loging via my Disquss account or by Facebook. Is their account validation process not up to your standards Randall S. Olson?

  • Christopher Smith

    Who exactly is grasping at straws here?

    You write:

    That led me (and others who have reviewed the topic) to wonder whether the wealth of a nation better predicts the average intelligence of its citizens.

    It might, however this is a red herring. There are multiple factors that correlate with intelligence. The point under debate is whether religious belief does. Your attempt to cancel out the less intelligent believers due to other unrelated possible correlations is entirely ad hoc for this very reason. In short, all you have done is wipe out the less intelligent believers from the list and then announced that this removed the significant difference. No kidding!

    You commit the same willful torture of statistics with your final point regarding a difference of 97 to 103 in points in another study: “Why Liberals and Atheists Are More Intelligent.” Here, your misdeeds are equally clear: you purposely mix together two different measures of significant difference. Yes, a 6 point deviation may not be significant when measuring overall IQ , but it IS significant when measuring for a correlation in a population!

    Others here see the same problems. Your bias is showing.

  • andy_the_bear

    I do not think dividing people into “religious” and “atheist” makes much sense. Not all religions are the same, nor have the same plausibility. As an overly simple example imagine there were three world religions, A, B, and C, and also Atheism. We therefore have four different sets of doctrines about reality. Moreover let us say that the doctrines of religion A are always chosen by the smartest and wisest among us. Let us say crude and selfish people always choose religion B and go out spreading it through conquest and blood shed. Let us say that people who are in a particular kind of culture strong in ancient traditions have huge social reasons for staying with religion C. And let us say that some people for various reasons when living in a libertarian society with freedom of thought created by religion A become Atheists. Its quite plausible you would end up with the kind of data these studies have.

About this blog

The data visualizations on this blog are the result of my "data tinkering" hobby, where I tackle a new data analysis problem every weekend. If I find something interesting, I report my findings here to share with the world.

If you would like to use one of my graphs on your website or in a publication, please feel free to do so with attribution, but I would appreciate it if you email me first to let me know.

Archives

Subscribe

Enter your email address to subscribe to this blog and receive notifications of new posts by email.