For years HBDers have been citing twin studies to argue that IQ is up to 80% heritable. The logic seemed straightforward enough: Take pairs of identical twins separated at birth (or shortly there after) raise them in different homes, and then give them an IQ test. When this was done in the famous Minnesota Twin study, the correlation was 0.75, suggesting 75% of the variance in IQ is caused by genes. Taking the square root of the correlation suggests an astonishing 0.87 correlation between IQ and genes, and that 87% of the individual differences in American IQs are caused by genes.
Critics argue that the twins were not raised sufficiently apart to warrant such strong conclusions. Some like commenter RR argue that the very idea of inferring causation is absurd because genes and IQ interact. For example good genes cause good environments and bad genes cause bad environments so even if identical twins are perfectly separated into random environments at birth, the fact that they both have the same genes means they’ll both end up in similar environments by age 40, so does it really make sense to credit the genes for their IQ similarity when the genetic effect was mediated by environment?
My answer has always been “yes”, because even if environment were what directly caused someone to have a high or low IQ, it was their genes that made them smart or dumb enough to end up in a good or bad environment in the first place, relative to the one they started in, so genes were the ultimate cause of the IQ.
But then I did a thought experiment: Imagine if in the future they develop really good gene editing technology and the richer your parents, the better the gene editing is. Now imagine if we did studies of identical twins raised apart in this society. We might find heritability drops to only 0% because when your identical twin is raised from birth in a much richer home, his adoptive parents can afford the best gene editing software money can buy, causing his intelligence to be twice yours.
Even if the IQ gaps between identical twins raised apart in such a society were 100% caused by genetic editing, twin studies would show them to be 100% caused by environment because the IQ correlation between identical twins raised apart in that society might be zero.
Even though it’s technically true that their environment caused their IQ differences since environment mediated access to gene editing, it would feel somehow wrong to say genes had no effect on their IQ, when literally 100% of the IQ differences were caused by gene editing.
So what I learned from this thought experiment is that heritability does not partition the variance into genes vs environment, but rather it partitions it into genes you were born with vs environment you were born with. Only in organisms that have very little ability to change their environment (like plants or to some degree people living in strict caste systems like 1940s India) , do twin studies capture the full spirit of heritability.
