The Missing Methodology
There's an obvious additional methodology for studying nature vs. nurture in IQ besides twin, adoption, and PGS studies. Yet, few scientists have tried it.
Since Francis Galton in 1865, disentangling the effects of nature and nurture has been one of the big leagues of scientific and intellectual endeavour.
Despite the rapid improvements in DNA technology, this is currently a slow era for discourse on nature vs. nurture largely because the fear of career cancellation, such as the firing of tenured professor Bryan J. Pesta for carrying out a landmark study of the correlation between cognitive test scores and racial admixture as measured from DNA, has tended to cause employed academics who favor the Nature side to keep their heads down.
The Noah Carl and Nathan Cofnas affairs are similar examples of how wokeness isn’t dead yet in academia.
The basic problem is that the center-left Establishment has decided, in its infinite wisdom, that if if even more persuasive evidence were to emerge that whites and blacks differ in average intelligence in part for reasons that can’t be solved by more government spending and more censorship of crimethinkers, then that would be the worst…thing…ever.
It would just prove Hitler was right.
Or something.
So, many of the most outspoken scientists pointing out the influence of Nature are at present lacking in conventional academic credentials.
Conversely, most left of center academics with a strong awareness of the science are pretty discouraged by the piling up of data in this century against their fondest hopes. They are in a delicate situation. All their friends who are more ignorant than they are about the science think that they’ve won the Nature vs. Nurture debate overwhelmingly, with the hereditarian side permanently castigated as “pseudoscientists.”
I mean, that’s what it says in Wikipedia, right?
Trust The Science!
The top talent on the anti-hereditarian side may well fear that if they publicly intervene, they might in turn cause the real heavyweights, such as Harvard’s Steven Pinker and David Reich, to intervene on the other side more forcibly than they have chosen to do so far. This could then explode in the popular mind the current conventional wisdom that The Science is settled.
So, not much is going on at present in terms of outright arguing. Hence, the big names lately on the Nurturist side are people like Sasha Gusev and Eric Turkheimer, a big decline from the Stephen Jay Gould and Richard Lewontin era a half century ago.
Turkheimer sums up why he’s confident lately:
The long road to Jaishankar et al
Eric Turkheimer
May 20, 2026
If you had asked Francis Galton’s grandma why some people get more education than others, I think she would have said something like, “Well, people are different in their organic constitution, and they are born into different family circumstances, and then those two things mix together in very complicated ways as people grow up. People aren’t raised under controlled conditions, which makes the two ingredients basically impossible to separate.” Let’s call Grandma Galton’s account the default naive hypothesis [DNH]. …
When twin studies were pointed at the DNH, they suggested large if poorly specified genetic effects, which all in all seemed to vindicate Galton’s ideological hypothesis. Heritability coefficients for EA in particular were generally around 50%. Arthur Jensen thought the heritability of IQ was .8; The Bell Curve was written on the assumption that it was .6.
There are also adoption studies as well.
And the rare but famous twins-raised-apart studies, which are twin studies and adoption studies rolled into one.
But there are a lot of problems with twin studies,
For example, last year I looked at 87 examples of identical-twins-raised-apart.
But how apart? In 25 of the cases, the twins attended the same school at some point.
I would guess that when some sort of family tragedy occurs and the extended family steps in to take care of the orphaned twins, it’s rare for one twin to be raised in wealth and the other in poverty because in that case the rich adoptive parents would usually take both. So, it’s more likely for identical twins to be split up when nobody in the extended family has a lot of spare cash, so they wind up being both raised by adoptive parents who could only afford to take one. Thus, these case studies are seldom the Prince and the Pauper examples we’d love to see.
all of which come down to the fact that they are poorly specified genetically, and don’t really solve Galton’s methodological problem of “disentangling” nature and nurture. That’s an old story.
But then, just in the nick of time, a new tool came along: The Human Genome Project, which for the first time allowed behavior genetics to be based on real DNA. The field spent a decade pointing this new tool at the DNH, hoping to find the genes that cause educational differences, but once again it didn’t work. Genes obviously had something to do with all this complexity, but the attempt to boil it down to the main effects of individual alleles was a failure.
The Candidate Gene era of the first decade of this century.
Then, again just in the nick of time, GWAS [Genome-Wide Association Study] was invented, and things started to look up. A series of four papers, commonly called EA1,2,3 and 4, used ever more sophisticated genomics in samples crossing into the millions.
EA stands for Educational Attainment. It’s a much more commonly found datapoint than, say, IQ because it only requires patients to check a box indicating how far they got in their schooling rather than to take a test. (Of course, there is seldom much validation of whether they are exaggerating or not.)
However, Educational Attainment clearly has something to do not only with how naturally intelligent you are, but also with how much money your parents have to nurture you further along in school before you have to get a job.
Tiny, statistically significant associations with individual bits of DNA were identified, but the effects were tiny and lacked any obvious biological relevance. Grandma Galton’s DNH still held true.
But those tiny genetic associations could be aggregated, offering a new way to estimate heritability coefficients called SNP heritabilities, and better yet to provide concrete “polygenic scores” (PGS) that could be used as real-world genetic predictors of EA. And the results, while not exactly at Arthur Jensen levels, looked promising! By the time EA4 came along, the R2 for the PGS was up around 15%! …
But wait. It turns out that modern genomics offers us the possibility of studying genetic differences within families instead of just between them, in the process controlling for many of the confounds that worried Grandma Galton. Moreover, as Paige Harden pointed out, meiosis assigns genes more or less at random within families, producing something resembling a randomized clinical trial. Now, finally, we can crack Grandma Galton’s DNH and get some real genetic causation.
Do siblings who randomly got handed higher polygenic scores for educational attainment get more educational attainment?
The purported meiotic randomized trial instantiated by the PGS accounts for one tenth of one percent of the variance in EA.
Obviously, one thing that is likely going on is that parents tend to intervene to flatten out differences in educational attainment among siblings in order to lessen hard feelings. Modern Americans aren’t much like Imperial Chinese extended families where the whole clan gets together and determines who the one kid is who has the highest chance of passing the Emperor’s exam and becoming a mandarin.
But, still, it’s an awfully small difference.
How come?
Beats me.
It’s a pretty new methodology, and I’d suggest waiting awhile before drawing too many far-reaching conclusions.
I’d also point out that that the flip side of the Missing Heritability is the Missing Malleability. What progressives have been looking for without much luck for the last 60 years is not just too undermine hereditarianism but also to bolster the case for government spending on interventions to reduce The Gap.
Now, it well could be that there is not just Nature (in terms of the genetic blueprint) and Nurture (in terms of how much money is spent on your childhood care), but also a semi-random category we could call Development. For example, as your brain develops, how many of the countless steps happen exactly as planned in your DNA? Or do you get a random infection or a bump on the head? Or do you happen to watch a TV show about Hannibal’s war elephants that inspires you to read books about the Carthaginian War and become more of a bookworm than you would have been otherwise? Or whatever?
Now, it’s not at all clear how Great Society-like programs can fix problems with Random Development. Maybe they can, but it doesn’t seem obvious how they would.
So that leads me to recommend a fifth major methodology besides twin, adoption, GWAS, and within-family PGS. And they are something that has been around for a few generations now and hundreds of thousands of people have been born due to their existence, but scientists haven’t much studied the phenomenon of …
Paywall here.
