Shop Mobile More Submit  Join Login
By Jonathan Kane

In our featured topic from October 2011, I asked what scientific discoveries people would most like to see during their lifetimes.  My own answer was that I'd like to see a complete understanding of the genetic influences on human intelligence, including whether or not genes play a role in group differences.  My prediction was that it would be decades before this was possible, but research about genetics and intelligence has advanced far more quickly than I could have hoped.  And while the wish that I made in October 2011 hasn't been granted yet, the possibility that this will be completely understood within my lifetime seems much more achievable now than it did only a few years ago.

At the time when I wrote that post, I wasn't aware that within the past month, the journal Molecular Psychiatry had published a study conclusively linking variance in human intelligence to specific genes.  This study has several authors, but the best-known is the psychologist Ian Deary.  This was a genome-wide association study (sometimes abbreviated GWAS), a type of study that scans the genomes of many individuals, and searches for alleles that are statistically associated with a particular trait.  Human intelligence is polygenic, influenced by vast number of genes that each have a tiny effect, so scanning the entire genome is typically the only way for studies about this to obtain a reliable result.

Genome-wide association studies establish that human intelligence is highly heritable and polygenic

A common concern with genome-wide association studies is that some of the alleles they identify might not actually be associated with the trait being studied, but instead be associated with a population group that happens to have above-average incidence of the trait.  For example, a GWAS looking for genes associated with Torsion Dystonia could potentially detect genes related to Ashkenazi Jewish ancestry, because Torsion Dystonia is especially common among Ashkenazi Jews, even though most of the genes associated with Ashkenazi ancestry have nothing to do with this disorder.  This sort of confounding factor is called "population stratification".  However, the risk of population stratification is well-enough known that in modern genome-wide association studies, it's standard to find ways of avoiding it.  The danger in Deary's study was that the study might detect genes associated with ethnic groups that have lower or higher average IQs, instead of associated with intelligence itself, so the authors avoided this risk by limiting the study to individuals of Caucasian descent.

In psychology, it has been well-established for a long time that IQ is highly heritable, and also that it is a very good predictor of certain real-life outcomes, especially scholastic achievement such as grades and SAT scores.  With this knowledge, one might expect that scholastic performance has a genetic component as well, and several studies have shown that this relationship also exists.  One of the most important recent studies on this topic was published by the behavioral geneticist Robert Plomin in the journal PLoS ONE.  While earlier studies had focused on the academic performance of children, Plomin's study is important because it examined the test scores of 16-year-old high school students, showing that the high heritability of academic performance persists into early adulthood.

Strong Genetic Influence on a UK Nationwide Test of Educational Achievement at the End of Compulsory Education at Age 16

Instead of linking the variance to genes directly, Plomin's study used the classic behavioral genetics method of comparing identical and fraternal twins.  Fraternal twins share the same prenatal environment, but are no more genetically similar to one another than ordinary siblings are.  On the other hand, identical twins share a prenatal environment as well as all of their genes, so the degree to which identical twins are more similar than fraternal twins for any trait shows how strongly the trait is influenced by genetics.  Plomin found that in science classes, genes account for around 58% of variance in test scores, while for humanities subjects such as art and music they account for around 42% of the variance.  Plomin has presented some of the data on this topic in his and Katheryn Asbury's book G is for Genes, which is a good introduction to this topic for anyone not already familiar with it.

While comparing identical twins to fraternal twins is useful as a way to estimate the heritability of a trait, the downside of this method has always been that it can't identify specific genes.  Some of these genes have been identified in a third study, published in the journal Science a few months before Plomin's study was published.  This study is sometimes referred to as "Rietvald et al", but its list of authors is in fact the longest I've ever seen in a scientific paper.

GWAS of 126,559 individuals identifies genetic variants associated with educational attainment

Like the Ian Deary study, Rietvald's study was a genome-wide association study, and avoided population stratification by examining only Caucasians.  It identified ten alleles associated with scholastic performance, accounting for around 1% of the trait's variance.  It's clear from Plomin's study that genes account for far more than 1% of the variance in scholastic ability, but more sophisticated methods will likely have to be used to find the rest of the genes involved.

Of course, when most people get an education, it isn't something they do just for fun.  The reason people work hard to do well in school, and later in college, is because it increases their odds of eventually obtaining a high-paying job.  Earlier studies have also found that childhood IQ is a moderately good predictor of income as an adult, so one might expect there to be some genetic influence on socio-economic status as well.  Last month there was a new study about this relationship published in the journal Intelligence, again with Robert Plomin as the best-known author.

Genetic influence on family socioeconomic status and children's intelligence

Examining the relationship between genes and socio-economic status has never been possible with the traditional method of twin studies, because socio-economic status is something that varies between one family and another, whereas any pair of twins would both be part of the same family.  But with the advent of genome-wide association studies, directly examining this relationship is now possible, and Plomin's new study is the first to examine it.  This study found that at least 18% of children's socio-economic status can be accounted for by specific genes, although the actual number is probably much higher than 18%, because the limitations of the study did not allow it to examine all genes that might have an effect.  Even more significantly, it found that almost all of the same genes that influence socio-economic status are genes which also influence IQ.  This study avoided population stratification by statistically excluding alleles that were found to cluster together, since this would indicate that they represent populations, while alleles that influence traits such as intelligence are distributed randomly.

Although all of the aforementioned papers push forward our understanding of genetics, and the influence of genetics on human behavior, there's very little they reveal that's truly surprising.  Among psychologists, it's been widely-accepted since the 1980s that IQ is highly heritable, and that the characteristics measured by IQ tests strongly influence both academic achievement and income, even if the specific genes involved weren't known yet.  Identifying some of these genes does not fundamentally change how these topics are understood in psychology, but is more a matter of filling in the gaps in what was already known.  However, there is one more study worth mentioning that is truly groundbreaking, and which brings behavioral genetics a step closer to answering a question which could not be answered before now.

Factor Analysis of Population Allele Frequencies as a Simple, Novel Method of Detecting Signals of Recent Polygenic Selection: The Example of Educational Attainment and IQ

For the past 45 years, psychologists have debated what the reason is why some ethnic groups have higher or lower average IQs than others, and whether or not those differences include a genetic component.  As of today, it has been exactly 45 years--the paper which initiated the modern debate on this topic, by the psychologist Arthur Jensen, was published on February 28th of 1969.  Davide Piffer, the author of the above linked paper, has attempted to answer this question using the data from the Rietvald et al. study.  Piffer's study found is that the genes which influence academic performance are distributed unequally between ethnic groups, and that their distribution almost perfectly matches earlier research about how ethnic groups vary in average IQ, with a correlation of 0.9.  (That is, 90%.)

This correlation is quite a bit higher than I would have expected, considering the alleles identified by Rietvald et al. account for only about 1% of the total variation in scholastic ability.  Piffer's explanation of this is that if the alleles which affect cognitive function are distributed randomly in a uniform way, then examining only a small number of them can still provide an accurate sample of them all, just as radiometric dating can accurately measure how isotopes are distributed in a rock by testing only a tiny sample of it.  These results also are disconcerting because for every previous study that's attempted to identify the cause of group differences in IQ, the results have been potentially interpretable as consistent with either a genetic or environmental cause.  But in this study, the correlation between the distribution of IQ scores and the distribution of genes affecting ability was so high that there's less than a 1% chance it could be due to coincidence.

During Bill Nye's debate with Ken Ham earlier this month, Nye made the point that the only thing it takes to revolutionize our understanding of the world is a single piece of data that can't fit within the current paradigm.  This is the ideal of how science is supposed to operate, but I don't think Bill Nye was accounting for how long it often takes for the scientific community to accept new ideas.  No matter how impressive Piffer's results are, it probably isn't realistic to expect that this study will single-handedly resolve a debate that's existed for 45 years.  However, the study does show how with the genetic data now available, resolving this debate could be possible.

In order to resolve it, what must happen is for other researchers to test whether Piffer's results can be reproduced using other datasets.  Piffer's study has received inordinately little attention among psychologists, so it's my hope that mentioning it here will encourage someone to attempt this.  Now that genetic data has made questions like this one answerable, it's also my hope that researchers will not be afraid to answer them out of fear over what the social implications might be.  In both of his studies mentioned here, Robert Plomin has pointed out that data does not dictate any specific social policy, because policies depend in part on what our goals are.  Another point he makes is that pursuing any set of goals related to intelligence will be easier with knowledge of genetics than without it.

As long as science educators such as Bill Nye are emphasizing how the scientific method relies on being open to new ideas, it's essential that scientists practice what they preach, and accept new data such as the results of these genetic studies even if it goes against their pre-existing beliefs.  Thomas Huxley put it best: "Sit down before fact as a little child, be prepared to give up every preconceived notion, follow humbly wherever and to whatever abysses nature leads, or you shall learn nothing."
Add a Comment:
Duxide Featured By Owner Mar 3, 2014
Piffer has recently published another paper that bolsters the contention of the cited one. Racial groups differ significantly, with East Asians significantly having higher allele frequencies than Africans. This paper is quite technical, and it presents new methods for detecting polygenic selection. If you are not intimidated by the technical jargon, it will be a very rewarding reading. It was published on IBC, an open acces Korean journal of biology:…
Agahnim Featured By Owner Mar 3, 2014
Thanks for telling me about that paper; I hadn't been aware of it before.

Are you a colleague of Piffer's?  I can see from your profile page that you live in Italy, and that you registered at DeviantArt just in order to comment here.  The reason I'm asking is because there were a few suggestions I wanted to give Piffer about his research.  I was thinking of e-mailing him about them, but if you know him, I also could mention them to you.
Sinande Featured By Owner Mar 1, 2014
Re: the first paper: Jesus H., is 1% of the total variance considered a good prediction in GWAS-land? :o

(So, basically... this is yet another heritability study except with more DNA, and its own results make it clear that it can't be generalised even across people of European descent. I'm not sure why it would make you so optimistic.)
Agahnim Featured By Owner Mar 1, 2014
Re-posted comment, because my first comment was explaining something that I'm pretty sure you already understood.  I must remember you're already quite knowledgeable about these topics.

I mentioned Deary's study because it's the first time anyone has been able to identify alleles accounting for more than half of the variance in human intelligence, so it's sort of a landmark in that way.  Before 2011, nobody was able to identify enough alleles affecting intelligence to account for anything close to that.  But at this point, the major hurdle to predicting more of the variance is just measuring effect sizes more precisely, which seems like something that'll inevitably refined over time.

I'm not sure this study would have been enough on its own to make me optimistic, but you can get an idea of how rapidly this field is advancing by comparing it to some of the more recent studies, especially the second Plomin study from January.  There have been some pretty significant advances in methodology in just the past two years.  I'm not accustomed to anything in psychology advancing this rapidly--in psychometrics, it's not abnormal for studies upwards of 15 years old to still be thought of as current.
Sinande Featured By Owner Mar 3, 2014
Fair enough. I was just somewhat puzzled that you'd be so excited about a study that (1) confirmed something we already knew (intelligence is reasonably heritable) (2) doesn't seem to have much applicability to humanity at large.

Also, I don't encounter many genome-wide association studies in my normal reading, and looking inside this one was a bit shocking. A little too much harping on "significant" p-values of pathetic-looking correlations for my naive and unspoilt statistical taste :)
theubbergeek2 Featured By Owner Feb 28, 2014
We must be VERY carefull in such topics, as we face the rise of a 'Neoracism' using (pseudo)'science' to sputter crap...

The IQ by example had been criticised for being not so neutral, culturally and academically biaised, unreliable.
Agahnim Featured By Owner Feb 28, 2014
I think this is a good example of the sort of reaction that will impede progress in genetics, if any scientists react the way you're reacting.  Caution is fine, but what you're saying seems to be that you don't trust any methodology if it reaches conclusions you find unacceptable, at least in this area.

In general, once the biological variables that underly a trait are being identified, it's no longer a good argument to say that that the trait is a meaningless cultural construct.  (This applies to any trait, not just IQ.)  Even twenty years ago, before any of the genes affecting IQ had been identified, it was already known to be influenced by the brain's nerve conduction velocity, and by the brain's efficiency of metabolizing glucose.  (I can't post a link for the second one, but there's a chapter about it titled "Cerebral glucose metabolism and intelligence" in this book.)  Would your perspective be that measurements of nerve conduction velocity, which is a physical measurement, are culturally biased also?

It always mystifies me to see this sort of fear of scientific discoveries coming from the same people who make fun of the attitudes of creationists.  In Bill Nye's debate with Ken Ham, Ken Ham said he's resistant to discoveries in evolutionary biology because he's afraid that they'll undermine our morals, and Nye explained that we have to be willing to let science tell us things we don't like.  I don't believe there's any scientific discovery that could justify racial discrimination, but if there ever were, being afraid to accept the discovery for that reason would be no different from Ken Ham's attitude.  How can someone have a problem with that attitude from creationists, but think the same attitude is okay on the topics of neurology and genetics?
theubbergeek2 Featured By Owner Feb 28, 2014
I will say this is not just reaction fear (no polit pun intended) but knowing the world... Watching history - the history of social darwinism, eugenism, 'scientifical racism', and other related things like the supposed biological 'deep' brain differences between genders (where as feminists and others may say well and true, it is quite societal-cultural)...

It is going to happens. We have had Doc Mailloux sputtering crap around that Bell Curve thing, blacks supposedly being 'with a lower IQ'.. Bigots SHALL use such thingvs if that can push their agenda...

Biological determinism is dangerous.
doctormo Featured By Owner Feb 28, 2014  Professional Digital Artist
The careful part I think is three fold:

That if the western caucasian genetic bag is stupid, that doesn't mean that any one specific "white person" would be. Variability between groups vs variability within groups. And mixing, both modern and ancient.

That people can't be taught in different ways to suite their brain structures or supported by society so they can reach their potential without writing them off because of their perceived race or even their genetic profile.

That IQ and earnings equate to success. A subjected packed with more cultural class assumptions than an English boarding school.

I don't think we should resist the data or what the data certainly says; but I put my skepticism on high alert for any leaps of faith about what that means, implies or what we should do about it. It would be helpful if studies talked about "genetic sources" rather than populations, peoples or races. And not just to avoid reactionary politics, but to reinforce how the data shouldn't be misused to try and construct a narrative about racial predetermination.
Agahnim Featured By Owner Feb 28, 2014
The data has always shown that ethnic groups overlap in IQ more than they differ.  Anyone who claims that they don't, or who claims that you can accurately predict any individual's IQ from their race, is being just as unscientific as people who deny that this data exists at all.

Anyway, the point you're making is similar to the one that Plomin makes in both of his papers, and I agree with it.  Just as our fears about social implications can't dictate what science says, science also can't dictate any specific social policy, because social policy depends on what's important to us, which is a moral question rather than a scientific one.  I think your suggestion that everyone should be given the best education possible is a reasonable one, but what I think even more is that this sort of question isn't something for scientists to decide.

When I talk about this sort of research, the reason I usually stay off the topic of how it could be applied to social policy is because I think in general, those sorts of questions shouldn't be answered by the same people who are presenting the data.  I also don't have a strong opinion about how politicians should answer them, as long as they don't ignore the data altogether.  What matters to me has always just been that people accept the data itself, especially if they're people who criticize creationists for being unwilling to do that about evolution.
Paleo-King Featured By Owner May 17, 2014  Professional Traditional Artist
This is true, the data should be accepted for what it is, as long as it's proven to be robust and unbiased.

If one really accepts that the human genome is "boring" and there are not drastic differences in intelligence based on race, people should not be afraid to look at the data unbiasedly. Why be afraid of data if it hasn't been cherry-picked for political reasons? In the end the data will point the way to the truth after enough experiments. This goes beyond fear of racism, it it just plain ignorance and prejudice against the scientific method. Political correctness it seems has become a religion unto itself. People fear mere data, despite the fact that data can be challenged, re-tested, re-gathered, and on its own doesn't lend any support to political discrimination.

The scientific method can never be a legitimate basis for segregation or genocide anyway, and real scientists today know that. The Jim Crow laws in the USA and the Apartheid laws in South Africa had no scientific basis at the time they were written, yet they still were pushed through. Indeed for the politicians (NOT scientists) who instituted these laws, they had a far stronger basis in Genesis 9:20-27 than anywhere in science. Science was merely cherry-picked to "support" them well after the fact.

Of course in science no data is ever final. So if a study appears to show something distasteful, test again to see if you get a different result. Test, test, and test again.
Add a Comment:

:icondomain-of-darwin: More from Domain-of-Darwin

More from DeviantArt


Submitted on
February 28, 2014


1 (who?)