Greg Cochran has pointed me to this opinion piece by DiRienzo and Hudson (2005). From the abstract:

Unlike rare mendelian diseases, which are due to new mutations (i.e. derived alleles), several alleles that increase the risk to common diseases are ancestral. Moreover, population genetics studies suggest that some derived alleles that protect against common diseases became advantageous recently. These observations can be explained within an previous evolutionary framework term in which ancestral alleles reflect ancient adaptations to the lifestyle of ancient human populations, whereas the derived alleles were deleterious. However, with the shift in environment and lifestyle, the ancestral alleles now increase the risk of common diseases in modern populations.

This chord is in tune with findings from the chimpanzee genome project:

Starting from 12,164 catologued disease variants in 1,384 human genes, we identified 16 cases in which the altered sequence in a disease allele matched the chimpanzee sequence ... Six cases represent de novo human mutations associated with simple mendelian disorders ... The remaining ten cases represent common human polymorphisms that have been reported to be associated with complex traits ... In all of these cases we confirmed that the disease-associated allele in humans is indeed the ancestral allele by showing that it is carried not only by the chimpanzee but also by outgroups such as the macaque. These ancestral alleles may thus have become human-specific risk factors due to changes in human physiology or environment, and the polymorphisms may represent ongoing adaptations.

Recall that an allele of dysbindin-1, previously implicated in schizophrenia, was associated with lower IQ in a recent study (caveat: replication with better design absolutely necessary). In light of the above, the following facts are suggestive: (1) the low-IQ/schizophrenia-risk allele is ancestral; (2) the high-IQ/schizophrenia-protective allele was found by Pritchard's team to be under selection in Europeans; and (3) the frequency of the deleterious allele is ~0.05 in Europeans/East Asians and ~0.30 in Africans. It may be that one driver of human biodiversity is the staggered times at which dispersed populations have reached certain transitions.