Gene Expression

During a conversation with Nick Matzke he asserted that Creationists weren’t less intelligent necessarily. I contended that they were less intelligent. I based on this on snooping through the GSS when I was posting about the association between lower educational attainment and intelligence and religious fundamentalist & Biblical literalism. There are several variables in the GSS which ask respondents about their views on evolution, and the more intelligent and educated a person is the more likely they are to accept evolution. But this prompts a question: is this association simply due to confounds of variables? In other words, the association between intelligence and acceptance of evolution is simply a byproduct of other correlations. If it was not a byproduct of other correlations, one would expect that even among fundamentalists higher intelligence would predict acceptance of evolution.

To probe these questions I looked at some very specific variables in the GSS. Of the evolution related questions, SCITEST4 has the largest sample size. It’s also rather straightforward, as it offers the assertion that “Human beings developed from earlier species of animals.” This is a human evolution question, which I think tends to set a somewhat higher bar than a more generic evolution query. The responses are:

Definitely True, coded as 1
Probably True, coded as 2
Probably Not True, coded as 3
Definitely Not True, coded as 4

The numeric equivalents are important. In short, lower values are more “pro-evolution” and higher values more “anti-evolution.” It is a convenient ordinal spectrum. So we’ll think of it as a dependent variable where higher values indicate greater anti-evolutionism.

The predictors or independent variables I’ll be looking at are:

WORDSUM = number 1 out of 10 correct on a vocabulary test
DEGREE = highest educational attainment, 0 = less than secondary education, 4 = graduate (so again, an ordinal interpretation is valid)
POLVIEWS = 1-7 from most liberal to most conservative
FUND = 1-3, Fundamentalist, Moderate and Liberal
BIBLE = 1-3, Bible as Word of God, Inspired or Fable

(and yes, some of these independent variables aren’t that independent from each other, for example the last two and the first two….)

First, let’s look at all respondents in the sample…

i-226cc9d91c45614c05f358848ef1eb61-allrespondentevolution.jpg

The Y axis here is the mean value on a 1-4 scale in terms of attitudes toward evolution, with higher values indicating rejection and lower values acceptance. The X axis is straightforward; the number correct on the vocabulary test on a 0-10 scale. If in a particular class everyone accepted that evolution was definitely true the mean value would obviously be 1. If half believed it was probably true, and half definitely true, the mean value would be 1.5. And so on. The trend here is rather clear: the less intelligent are generally skeptical of human evolution (their mean value is closest to “Probably Not True”) and the most intelligent are most accepting of human evolution (mean value approaching “Probably True”). This is what I expected. But of course even if there is a correlation that doesn’t mean that intelligence is driving this relationship.

OK, let’s switch to educational attainment:
i-8eb77844988ba21799fc2fd4701419b4-highcollegeevolution.jpg

Because I dropped all classes where N < 50 the bottom of the distribution is cut off for those who have some college education or more. I lumped those with only secondary education or less, and those with some college or more. This increased the sample sizes for any particular class. Note here that the effect of intelligence is much more noticeable in the more educated group; even when you separate those with graduate school educations the effect is there. On the other hand the less educated sample shows little effect of intelligence on skepticism or lack of in regards to human evolution. I suspect that this tells us that the effect of intelligence, if it exists, is very conditional on the population segment.

Next, let’s move to politics:
i-d6fa63b8e5888cf26fe7e5a8a9470f75-politicsevolution.jpg

Two major results from this. First, greater intelligence seems to correlate with higher rates of acceptance of evolution across the political spectrum. But, the effect is much more powerful among liberals than among moderates than among conservatives. This relates to the second point: at lower levels of intelligence there is relative convergence at some skepticism toward evolution. The difference between the groups is due in large part to the huge variance at the very highest levels of intelligence. The most intelligent conservatives have similar views to the less intelligent liberals.

So far I’ve hit the variables which seem only indirectly related to why someone would reject human evolution. I’ve argued before based on GSS data that the conservative rejection of evolution is more about the representation of religious conservatives among political conservatives than conservatism entailing creationism as such. So let’s look at a religious variable, fundamentalism:
i-5e6f1d1728b35c2e1efd8f415579a380-fundevangevolutions.jpg

In this case “Liberal” does not refer to politics but to religion. This makes sense, since there have been religious liberals who are political conservatives or to the right-of-center. The Republican successor to Teddy Roosevelt, William Howard Taft, is an exemplar of this category. The first thing that jumps out is that there doesn’t seem to be an effect of higher intelligence on Fundamentalists. And there is a somewhat stronger effect of religious Liberalism than Moderation. Again, as with politics you see that the chasm between the groups gets really big the higher you go up the ladder in intelligence. In fact the less intelligent religious Liberals and Moderates actually lean a bit to the human evolution skeptic side.

Finally, we get to a variables that addresses Biblical literalism. This is the heart of the matter in many ways; a priori one would assume that those who would believe that the Bible was the literal word of God would be very skeptical of human evolution since a plain reading seems to refute that position. And that is what we see:
i-84e913a6ba6d8c5e983bad7be16cab66-bibleevolution.jpg

What jumps out at you? To me it is that more intelligent literalists are more skeptical of evolution than the less intelligent! This makes sense, doesn’t it? If you aver that the Bible is the literal word of God the brighter you are the more internally coherent you want to be because systematic relationships inferable from presuppositions are obvious. And again, note that even among those who believe that the Bible is a book of fables, and so reject its role as a descriptive model of the universe revealed from a divine being, the less intelligent exhibit more skepticism of evolution than the more intelligent.

From these data I conclude that it seems that ideology, group identification and social conformity probably has more to do with how plausible one finds human evolution than intelligence. Granted, as you can see, intelligence remains powerfully predictive for some subgroups, but it is not a significant predictor in other groups. All the variables above have interpretations in an ordinal sense. For example, Fundamentalist, Moderate and Liberal are on a spectrum. But, aside from WORDSUM, one doesn’t have a real natural quantitative interval on these variables. Nevertheless, I decided to get a qualitative feel for how the independent variables could predict attitudes toward evolution as the dependent variable. Since this isn’t totally kosher, I’m not going to give you the values, but just note that WORDSUM is not very predictive across the whole sample and is not statistically significant in terms of its deviation from 0 in its beta. DEGREE, educational attainment, does have a significant beta, which is non-trivially smaller than POLVIEWS, political orientation, and FUND, the fundamentalism spectrum. The largest effect is from BIBLE, which is one we might assume intuitively. Of course, if I constrain the sample to only a particular class (e.g., religious Liberals) then WORDSUM does turn into a statistically significant predictor.

And this last point is one I think which is worth keeping in mind. If you are a person who is liberal in orientation and you interact mainly with other liberals, then you will naturally see a strong relationship between intelligence and acceptance of evolution. If, on the other hand you are very conservative, that relationship will be harder to discern. Finally, if you are a fundamentalist Christian you won’t see that relationship in your day to day life. Daniel Dennett’s universal acid of Darwinism might not be so universal at all; among liberals it is rather acidic, especially as one goes up the ladder of intelligence and education. Among religious conservatives, far less so. I think it is plausible that the acid can turn one from a religious conservative into a secular liberal, and that is perhaps more to the point of Dennett’s thesis, but I think that that generalization elides the reality that social structure remains within our culture.

But why is it that the acid needs to burn its way through anything in the first place? Recall that even among those who are very liberal, whether politically or religiously, or who reject the Bible as anything more than fables, there was still a tendency to be more skeptical of human evolution with a lower intelligence. I think part of the issue here is what Paul Bloom addressed in Childhood Origins of Adult Resistance to Science:

The main reason why people resist certain scientific findings, then, is that many of these findings are unnatural and unintuitive. But this does not explain cultural differences in resistance to science. There are substantial differences, for example, in how quickly children from different countries come to learn that Earth is a sphere (10). There is also variation across countries in the extent of adult resistance to science, including the finding that Americans are more resistant to evolutionary theory than are citizens of most other countries (24).

Part of the explanation for such cultural differences lies in how children and adults process different types of information. Some culture specific information is not associated with any particular source; it is “common knowledge.”….

…So rather than evaluating the asserted claim itself, we instead evaluate the claim’s source. If the source is deemed trustworthy, people will believe the claim, often without really understanding it. Consider, for example, that many Americans who claim to believe in natural selection are unable to accurately describe how natural selection works (3). This suggests that their belief is not necessarily rooted in an appreciation of the evidence and arguments. Rather, this scientifically credulous subpopulation accepts this information because they trust the people who say it is true.

So the rough model is like so:

1) Humans have a natural intuition about the way the universe works which is biased against evolutionary hypotheses. This is why it took so long for these models to ascend to accepted truths even among intellectual elites despite the fact that there were evolutionary ideas extant as far back as the pre-Socratics in ancient Greece. Paul Bloom has reported other research which shows that most young children start out with Creationist biases; but as those in families which accept evolution develop they shed their inchoate Creationist views and accept an evolutionary paradigm. In contrast, those who are raised in Creationist families have their intuitions reinforced through an ideological program where Creationism is offered up as the true and correct alternative to mainstream science.

2) I think it is important to acknowledge that most individuals who do accept evolution do not do so because of the natural and clear obviousness of its postulates and inferences. This is inferable from the fact that most people have only a rough and impressionistic idea of what evolution is. Very few people could elucidate Darwinian theory will any clarity or fidelity, let alone add the nuance of models such as Motoo Kimura’s Neutral Theory of molecular evolution. When I was in 7th grade I was going through a test for a gifted program and the psychologist asked what I knew about Darwin’s theory of evolution. When I explained my rather primitive ideas there was a long pause, and the psychologist told me that I was the first child he’d interviewed who had given a coherent response which exhibited some understanding of what Darwin had originally said. I don’t repeat this to show how well I understood Darwin’s theory; I will admit that I have only understood evolution with any great personal degree of confidence for the past 5 years or so. Rather, it illustrates the shallow knowledge base of even those children selected to go through the process of testing for a gifted program.

3) So why the relationship between intelligence and acceptance of evolution among some groups? I suspect that it has to do with identification with peer groups. The extremely intelligent are less likely to trust their own intuition, and receive more reinforcement from peers as to the power and veracity of specialized knowledge producers and providers. If one is, for example, a professional in an abstruse field then it stands to reason that there would be some rational self-interest in ceding to other professionals in abstruse fields the validity of their truth claims. Additionally, those with higher IQs see in their own lives the value which rational and systematic thought can produce, and, that the inferences of such thought are not always plain and obvious from the perspective of someone who relies on introspection and intuition.

4) But, among some subgroups the peer pressures are different. If an evangelical conservative industrial engineer has to pick between his religious identity and the opinions of scientific professionals in academia, it seems not totally implausible that he would select the former. Of course, if you changed it so that the evangelical was a biologist such as Francis Collins then the pressures crank up to a whole new level, and when it comes to evolution group identification with other religious conservatives is not nearly as easy. Conversely, when you have a situation where the peer group is not encouraging even high intelligence may not result in acceptance of evolution. Look at the lack of relationship between intelligence and acceptance of evolution for those who do not go to post-secondary education; my own suspicion here is that being embedded in a dull subculture effects those who are not not dull, just as being a dull person among brights also has an effect.

5) Acceptance of evolution, like science, is about trust for most people. Trust in elites, or trust in your peers, depending on who you are. Very few humans will ever read The Origin Of Species or The Selfish Gene. Similarly, during the 17th century very few educated Europeans personally replicated the observations of Galileo. Rather, Europeans trusted in a system where scholars were attempting to converge upon theoretical truths which mapped to a high degree of accuracy to the world around us. To a great extent the same holds today.

6) The fact that most people who accept evolution have little detailed grasp of its theoretical superstructure implies that they also don’t generate many inferences from the model.

Comments

  1. #1 Otheus
    October 1, 2008

    Very nice analysis, and all the more impressive that it rebutted your initial hypothesis. As a a former-literalist-turned-atheist and professional scientist, I am deeply entangled in both fundamentalist and very liberal social circles and affirm (non-quantitatively, of course) these results.

    However, since moving to Europe several years ago, I must say the climate here is much different. Not only do “Left” and “Right” in politics mean something quite different, but very few are skeptical of evolution at all. In Austria, at least, the minority of Protestants seem to be made up of more intelligent persons, and yet they are most likely to express skepticism of evolution.

  2. #2 razib
    October 1, 2008

    re: europe, see these data. the relatively low levels of creationism also are associated far less with political or religious differences.

  3. #3 Bob S
    October 1, 2008

    Evolution per se needs to be distinguished from Darwin’s natural/sexual/kin selection mechanisms. Virtually all university faculty (generally smart people) believe in “evolution.” But, go over to the humanities and social sciences, and you will find that nearly everyone there rejects the relevance of selection to human capacities, especially intellectual capacities. The best example is David Stove’s “Darwinian Fairytales.”

  4. #4 steve martin
    October 1, 2008

    Nice work!

    On your quote that:
    Acceptance of evolution, like science, is about trust for most people. Trust in elites, or trust in your peers, depending on who you are.

    This is true, but for most of the North American evangelical community, the ID marketing machine has functioned as the “scientific elite”. It takes a lot of work for those of us that question the anti-evolution orthodoxy of our community; books by Denton, Behe, and Dembski get trotted out very quickly when you start asking questions. And at first blush (for those without an indepth background) they look impressive. This might explain (or at least affect the results) of your last two graphs.

    Interesting, I believe there is quite a recent phenonemon that may affect your data; Evangelical biologists are speaking out. From my own experience (very much anecdotal and I have no hard facts) evangelical biologists mostly support evolution. But it is only in the last 5 years that they have been popularizing their message. Collins is only 1 of (from my count) 10 books from Evangelical scientists supporting evolution published in the last 5 years. Before that, there were none, at least on this side of the pond. As the previous commenters noted, things are different in Europe. Anyways, all of a sudden, Evangelicals in the pew are now being forced to think since we have guys like Collins, Falk, Colling, Lamoureux, Haarsma, Giberson, and Gingerich and others disagreeing quite strongly with Behe etc.

    You might also be interested in Marlowe Embree?s series on the “Social Psychology of the Origins Debate” over at my blog An Evangelical Dialogue on Evolution (The 6th article in the series was posted this AM). Embree (a psychologist from University of Wisconsin Colleges) is doing active research on the subject now (he’ll be providing readers access to his research intrument, a questionnaire, in his next post). Here is a brief snippet from his introduction to the series :

    This first article in the series will consider the question of how attitudes and beliefs are formed in the normal process of intellectual growth, as seen through the lens of Jean Piaget’s classic formulation of cognitive development. The next two articles will ponder the formation of biases and prejudices, with a view to understanding psychological forces that generate stereotypic understandings of social groups. The following two articles will examine the potential role of differing thinking styles and related personality factors on the question of why some people become evolutionists and others become creationists, or why some people change their minds about such matters while others do not.

  5. #5 RPM
    October 1, 2008

    Looks like there aren’t enough “smart” biblical literalists or enough “stupid” fable-ists.

  6. #6 Charles
    October 1, 2008

    Ludwik Fleck, in his book “Genesis and Development of a Scientific Fact,” talked about the “harmony of illusions” of thought collectives, which has some correspondence with your identity/peer pressure perspective on belief.

  7. #7 Noah Gray
    October 1, 2008

    Interesting work. Your correct description that many, even the well-educated, are still accepting scientific theories on “faith” is indeed an important point. Therefore, the mindset required to believe in the science is identical to the mindset required to believe in the creationist rhetoric, regardless of education received. So in retrospect (at least for me, since I was probably in the same generalizing boat as you were regarding the role of intelligence in this debate), it seems intuitive that all of the additional variables you list BEYOND intelligence are actually influencing the decision of which “faith” to follow.

  8. #8 derek
    October 1, 2008

    I don’t think the adoption of scientific thought is due to intelligence. I think that intelligence is due to the adoption of scientific thought.

    This is why all the talk about how our ancestors were no less intelligent than we are is missing the point. Sure, it takes a smart guy to come up with St Anselm’s Ontological Proof. It takes a smart guy plus science to be smarter and say of the proof, “You know, that’s bullshit.”

    If “natural genetic intelligence” were the key to adopting science, we could just test all the kids, teach the “smart” ones science and not waste our time on the “stupid” ones. We teach, or we should, science to all the kids, to make them all a little bit smarter than they would otherwise be.

  9. #9 abb3w
    October 1, 2008

    One quibble in two parts on the methodology: the use of a vocabulary test to measure intelligence. First, vocabulary is in part a measure of education, rather that raw ability. Second, verbal ability is semi-independent of mathematical ability, and I suspect math may be more important.

    The last graph results looks STRONGLY suggestive of a catastrophe theory transition; specifically, a Cusp Catastrophe. (See link for picture.) Using the labels on the axis as depicted, variable x indicates acceptance versus doubt of evolution (positive = doubt evolution, negative = accept evolution), and variable u indicates verbal level (positive = weak verbal, negative = strong verbal; of potential interest for further investigations, children tend to move from positive to negative as they age). Variable v in this study indicates degree of acceptance of the Bible (negative = Inerrant, positive = Fables). I suspect the true underlying variable v may more accurately be a measure for relative degree of acceptance of the importance of the Bible as a tool for thinking about the universe, in contrast with Science and Evidence. (The math-versus-verbal balance may take part.)

    As v values become increasingly negative, the x versus u view will become dominated by belief of the biblical account. People will be on the “top” surface; as verbal sophistication increases, so does their skepticism of evolution. Contrariwise, highly positive v values lead to dominance of the evolutionary view, and location on the lower surface. As verbal sophistication increases, the x versus u view shows a decrease in skepticism about evolution results.

    Considering the x versus v view for strongly negative u (the verbally sophisticated), we have two stable surfaces and an unstable “double bind” intermediary. The lower surface represents acceptance of the evolutionary account; as acceptance of the Bible’s importance increases, it becomes harder to retain the evolutionary view. Contrariwise, the upper surface represents belief in a literalist biblical account; as science exposure increases, it becomes more and more difficult for the verbally sophisticated to remain with Biblical belief. The moderate “Bible Inspired” position may correspond to the unstable double bind surface.

    I would suggest that it is not an inborn human intuition that causes the initial bias, but rather a universal social bias (due to the evolutionary benefit of religion for organizing societies; cf. Wilson and Diamond), and conveyed to new members (children) very early in the socialization component of the education process. I further suspect that the other correlations of intelligence to social group membership are dominantly the result of the correlation of group membership to state on the v axis.

    I’m too lazy to prove that. =)

  10. #10 razib
    October 1, 2008

    First, vocabulary is in part a measure of education, rather that raw ability.

    well, using 0-4 for DEGREE the correlation is 0.45 between it and WORDSUM.

    I would suggest that it is not an inborn human intuition that causes the initial bias, but rather a universal social bias (due to the evolutionary benefit of religion for organizing societies; cf. Wilson and Diamond), and conveyed to new members (children) very early in the socialization component of the education process

    it doesn’t have to be EITHER/OR. it could be that a definite innate cognitive bias in conceptualization occurs. barring a shock or push away from that bias, cultures normally converge upon a model which “fits” with that bias. subsequent social pressures reinforce and exaggerate the innate cognitive predispositions on a society-wide scale.

    as i alluded to in the post, there’s a lot of evidence from bloom’s work that children in secular/non-religious milieus start with the bias, but shift away from it through socialization during childhood and adolescence. there isn’t an converse trend among creationists where a large number have an evolutionary bias that they have to be conditioned away from.

    thanks for the rest of the comment, it was illuminating and formalized some of my intuitions.

  11. #11 razib
    October 1, 2008

    If “natural genetic intelligence” were the key to adopting science, we could just test all the kids, teach the “smart” ones science and not waste our time on the “stupid” ones. We teach, or we should, science to all the kids, to make them all a little bit smarter than they would otherwise be

    1) most people are too stupid to understand science, because most people are too stupid to get math.

    2) that being said, it is probably important to expose most people to science because they need to have some cursory feel for the “magic” which underpins their day to day lives. additionally, it might serve to foster respect for science and scientists so that they don’t burn them for being witch-doctors.

  12. #12 razib
    October 1, 2008

    Interesting work. Your correct description that many, even the well-educated, are still accepting scientific theories on “faith” is indeed an important point. Therefore, the mindset required to believe in the science is identical to the mindset required to believe in the creationist rhetoric, regardless of education received.

    there is a subtle to clarify here though. yes, those who accept evolution accept it because they have faith in science, not because they understand the specific detail of said science (this goes for an organic chemist in regards to evolutionary biology, and an evolutionary biologist in regards to cognitive psychology, and so forth). but, generally they are also often accepting claims at variance with “common sense” and “intuition” (think general relativity and quantum mechanics, or the “deep time” premise of evolutionary biology).

    creationists on the other hand don’t accept creation because it is a weird and exotic formula. i make the case above that to a great extent creationism is favored in a rigged game in our cognitive wiring. this is i think why creationists use common sense plausibility arguments so often; the whole game is to get a “hook” into intuitions and leverage that. ultimately the creationist rejection of evolution does have grounding in a religious faith and faith in a social system and elite who claim to be alternative arbiters of the scientific elite. but proximately they’re cognitive workings are actually less of a leap of faith because they are not accepting claims which push beyond their normal conceptualization powers (e.g., how can you conceive of what 4 billion years vs. 1 million years might mean?).

  13. #13 razib
    October 1, 2008

    Interesting, I believe there is quite a recent phenonemon that may affect your data; Evangelical biologists are speaking out.

    i wouldn’t be surprised if at some point in the future the “tide turns,” so to speak. skepticism of the reigning orthodoxies is always standardized to the specific time period.

  14. #14 razib
    October 1, 2008

    But, go over to the humanities and social sciences, and you will find that nearly everyone there rejects the relevance of selection to human capacities, especially intellectual capacities. The best example is David Stove’s “Darwinian Fairytales.”

    see #6 : -)

    also, the tide might be turning.

  15. #15 abb3w
    October 1, 2008

    see #6

    “I’m working on that.” (Amateurishly, I admit.)

  16. #16 outeast
    October 2, 2008

    If “natural genetic intelligence” were the key to adopting science, we could just test all the kids, teach the “smart” ones science and not waste our time on the “stupid” ones.

    My experience of streaming in schools suggests that we tend to do exactly that (though perhaps not in an explicit, intentional way); kids who do not have ‘natural’ mathematical ability tend to get sidelined in science education. But contra Razib, you can ‘get’ science even if you are relatively innumerate: it’s just that you need a different kind of model, models which are intuitive (Razib said it about religion when he wrote ‘the whole game is to get a “hook” into intuitions and leverage that’, but it can apply to science education as well).

    There are limitations to this, of course – if an intuitive model for QM is possible then it has yet to be found, and such an approach is not going to make an innumerate into a scientist – but I do not think it is true to suggest that innumeracy is an insurmountable barrier to grasping science.

  17. #17 Tony Jeremiah
    October 2, 2008

    Interesting analysis. I have a couple of comments:

    The Y axis here is the mean value on a 1-4 scale in terms of attitudes toward evolution, with higher values indicating rejection and lower values acceptance. The X axis is straightforward; the number correct on the vocabulary test on a 0-10 scale…The trend here is rather clear: the less intelligent are generally skeptical of human evolution (their mean value is closest to “Probably Not True”) and the most intelligent are most accepting of human evolution (mean value approaching “Probably True”).

    It’s important to note that the operational definition of intelligence for this analysis is WORDSUM, which is a vocabulary test. The particular form of intelligence being measured here is likely crystallized intelligence, which is known to be strongly influenced by experiences. Also, because of how intelligence is operationally defined here, the most conservative interpretation of the data is the presence of a negative correlation (statistically significant?) between vocab test score and evolutionary skepticism (i.e., higher vocab equates with lower skepticism about evolution). This might suggest that more time spent reading (in general), equates with reduced skepticism about evolution. And more time spent reading would probably also equate with more time in school (which likely explains the difference in evolutionary skepticism in the 2nd data set, broken down by educational attainment).

    From these data I conclude that it seems that ideology, group identification and social conformity probably has more to do with how plausible one finds human evolution than intelligence.

    I would agree, especially given the premise that WORDSUM is a measure of crystallized (experiential) intelligence. And further, that these experiences are likely stored in memory as described by the semantic network model concept.

    To assess whether “intelligence” is really responsible for comprehending evolutionary theory, fluid intelligence measurements (e.g., what Ed Yong reports on his blog) would be important.

  18. #18 Dutch
    October 3, 2008

    Anecdotal: I was surveyed once on this topic by a student who knew so little about the subject that he rated my understanding of natural selection 2 on a scale of 5. What I said was not in his manual. I had to take him to my bookshelf to convince him. We settled on 4 (-:

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.