For years commenter Mugabe has complained that studies of identical twins reared apart are meaningless because the identical twins are typically separated within the same country (i.e. America) so when two separated twins raised apart both have a high IQ, it doesn’t necessarily mean they have “high IQ genes”, it could just mean that they have genes that reach their full potential in America and Americans who are dumber than them, may have been smarter than them, had they been reared in China.  In other words, the intellectual rank order of people with different genotypes might change depending on what country they are in, a concept known as norm crossing.

To solve this problem, Mugabe has suggested doing a study on identical twins raised apart in many different countries, but the problem with that is when you increase the amount of environmental variation in a sample, and hold the amount of genetic variation constant, you end up with a ridiculously high correlation with environment (and thus low heritability), even if no norm crossing occurs.  This is known as the extended range problem (the opposite of the range restriction problem frequently encountered in IQ studies).

An idea that would correct for this, and one Mugabe has suggested many times, would be to convert IQ to Z scores relative to the country one was raised in, but this would only make sense if you could control for the genetic variation that exists between countries allowing the Z score to reflect only how smart one is relative to environment, and not how smart one is relative to the gene pool of the particular country.

An idea I had for doing this is instead of calculating Z scores relative to the population of the entire country, calculate Z scores relative to only the non-immigrant “pure” West African population of the country.  West Africans are an ideal international genetic reference group because not only do they live in many Third World countries, but they involuntarily migrated to other kinds of countries so you don’t have anywhere near the selective migration problem that occurs when other races live on multiple continents, as long as you exclude recent voluntary West African migrants.

So all that’s needed to do this study is a campaign to recruit 50+ pairs of identical twins (of any race) each raised from birth to age 40+ on different continents (or islands) in countries with a large enough involuntary “pure” West African population to use as a reference group for each participant’s Z score.

Of course the Z score correlation between these identical twins raised apart would still overestimate heritability because even identical twins raised on different continents shared the same prenatal environment. How do we subtract the effects of shared prenatal environment?

An idea I had would be to do a separate study on the Z score correlation between half-siblings adopted and raised from birth on different continents [update March 30/15: with biological parents living on different continents] where some of the half-siblings shared the same unknown biological father and others shared the same unknown biological mother.  The difference in correlation between the two types of half-siblings would provide an estimate for the effect of prenatal environment.