The right question to ask is: How big is the cumulative effect size of these genetic differences? The answer is — detectible but also negligible for most resear

The right question to ask is: How big is the cumulative effect size of these genetic differences? The answer is — detectible but also negligible for most research purposes (bias correction). Even negligible effects can be highly “significant” based on statistical criteria when sample size is large enough. The authors summarize the effect size in terms of standard deviation units. Here is the text, where pPGS mean “participation polygenic score” that is used to summarized the genetic scores influencing participation likelihood in UK Biobank. “The pPGS with weights derived from the GWASs using data of the first half (pPGS1) has average value for the second half of the sibling pairs that is 0.044 s.d. higher than that of the unrelateds (P=5.0×10−6). In reverse, the first half of the sibling pairs have average pPGS2 value, weights derived from the second half, that is 0.026 s.d. higher than the unrelateds (P=8.5×10−3).” To put this into context: From Wikipedia: “… the average height for adult men in the United States is about 70 inches, with a standard deviation of around 3 inches.” That means that 0.05 s.d. units is equivalent 0.15 inches or 0.38 cm.