Boom: the debate over ancestry and IQ just shifted. New within-family data show that genetic ancestry predicts IQ, even when siblings are compared. Previously, hard-nosed skeptics like Sasha Gusev cheered at null within-family results for education as “proof” that ancestry differences must be environmental. But using the same within-family design - this time with IQ and a large dataset - Cremieux demonstrated that the African-White IQ gap in the UK is explained by ancestry itself, meaning direct genetic effects are at play. Looks like Gusev just shot himself on the foot. Crucially, this effect did not hold for education or skin colour, for different reasons. Obviously skin colour is “genetic” but its sparse genetic architecture makes it difficult to detect within-family effects. For education, there are other barriers, which Cremieux discusses in his X thread. Combined with my work on polygenic score differences between populations, this is unequivocal evidence for a direct effect of ancestry on the B-W gap.