May 18, 2005

How to settle the question of the racial IQ gap

New technologies for measuring genetic admixture have made it possible to measure directly the percentage of ancestry that an individual has from two or more parental groups.

If two groups differ in intelligence due to the possession of private alleles, or the frequency of others, then we would expect that individuals of mixed ancestry would have a lower intelligence than the higher-achieving group, and higher intelligence than the lower-achieving group.

If, on the other hand, differences in intelligence are due to cultural factors, then we would expect proportions of genetic ancestry to have little explanatory power.

Therefore, to determine whether or not there is a racial character to intelligence, we must sample a mixed population, determine the IQ scores and ancestral proportions of its members, and see whether the two are correlated.

Previous studies have used skin color as a proxy for admixture proportions. Skin color has been found to be related to genomic ancestry in some groups, but not in others. Therefore, we must sample genomic ancestry directly.

Moreover, some previous studies have sampled from groups which are mixed, but also differ in other respects, e.g., the offspring of American blacks and German women after WWII, or of Americans from different parts of the United States; e.g., it is true that Southern blacks are ancestrally more Negroid than Northern blacks (on average), but there are other socio-cultural factors which separate the two.

The best study would sample from a single localized population, e.g., the black inhabitants of a town or city neighborhood; moreover, this population must be variable enough, i.e. have substantial variation in ancestral proportions.

Finally, individuals resulting from admixture between a Caucasoid and Black parent should be excluded, to avoid any possible heterotic effects.

If such an experiment were to be conducted, then the discovery of a significant correlation between IQ and ancestral proportions would be clear and unambiguous evidence for racial differences in intelligence. If the correlation is very weak, or absent, then this would argue against this dependence, although the experiment could and should be repeated multiple times and with many different groups known to have differences in IQ and for which ancestral proportions can be estimated.

No comments: