Fig 1. Gsens analysis of the effect of maternal educational attainment on child educational achievement.

Estimated standardized effect of maternal education on child educational achievement (Y axis) after accounting for genetic confounding using observed polygenic scores and heritability-based scenarios explaining an increasing percentage of variance (X axis). Point estimates and confidence intervals in black represent main estimates of interest, after accounting for (from left to right): 1: the best-fitting polygenic score; 2: SNP-heritability of educational achievement as assessed by GCSE scores in TEDS; 3: twin-heritability of educational achievement. A lower bound of 0 was imposed on the estimate, which is reached for the twin estimate of heritability (63%). The line “k = Observed” corresponds to heritability-based scenarios using values of model parameter k derived from observed polygenic scores (see Methods). “k = theoretical” corresponds to the value of k when the same trait is in parents and children and the heritability is the same in parents and children. In this case k = 0.5 (see Methods).