From the paper itself: >it suggests that the GWAS effect-size estimates may be biased upward by correlation between educational attainment and

From the paper itself: >it suggests that the GWAS effect-size estimates may be biased upward by correlation between educational attainment and a rearing envi-ronment conducive to educational attainment. A full 60% of the already modest 'effect' of correlated genes can be immediately attributed to environment according to their sibling study.