@Blaspie @NWOBruh @eric_knowles @DialecticBio @eyeslasho 1: These tests are extremely sensitive because they aggregate many effects. Prior work used simulations to show high power in just ~17k families (https://t.co/6sAe6ob5Pc). I'm open to evidence that t
@phil49472744 @thebirdmaniac @L0m3z Coming back to this, I don't really buy the issue of power anymore for the Qx test. Robinson et al (https://t.co/VSlFVoNWNs) showed they have high sensitivity with GWAS data from just 17k sibs.
@SashaGusevPosts Don't you think that there are innate differences between groups (important: not arguing that they follow stupid stereotypes)? Visscher et al. showed that they exist for height/bmi. Why not for other traits? https://t.co/qmB5j5cza4
@StevePittelli But the magnitude of the bias in effect estimate for any individual site depends on the magnitude of differentiation within the sample (equation 2.3 in the supplement here: https://t.co/9Ejg7FnXsq, or here: https://t.co/uKauMJCHQt for a more