I’ve been wanting a LW article on polygenic screening for some time now, so thank you so much for writing it. I’ll be going through it in detail later. For now, I’d just suggest rewording this:
“The average gene has a tiny, tiny impact on how long you spend in school. The predictor used 2,925 genes to explain just 15% of the variance in how many years of school a person completed.”
Instead of talking about the average gene, consider pointing out that “the most important gene” has only a tiny tiny impact on how long you spend in school. Of course, with 20-25,000 genes, the average gene was never going to be very impactful even if a single gene controlled 100% of educational attainment.
The point about 2925 genes controlling 15% of variance in years of schooling doesn’t support a candidate gene hypothesis—it supports a limited role for genes in controlling educational attainment. Even if just one gene was responsible for all of the genetic component in educational attainment, adding in the other 2924 genes would result in the same 15% of variance explained.
I’ve been wanting a LW article on polygenic screening for some time now, so thank you so much for writing it. I’ll be going through it in detail later. For now, I’d just suggest rewording this:
“The average gene has a tiny, tiny impact on how long you spend in school. The predictor used 2,925 genes to explain just 15% of the variance in how many years of school a person completed.”
Instead of talking about the average gene, consider pointing out that “the most important gene” has only a tiny tiny impact on how long you spend in school. Of course, with 20-25,000 genes, the average gene was never going to be very impactful even if a single gene controlled 100% of educational attainment.
The point about 2925 genes controlling 15% of variance in years of schooling doesn’t support a candidate gene hypothesis—it supports a limited role for genes in controlling educational attainment. Even if just one gene was responsible for all of the genetic component in educational attainment, adding in the other 2924 genes would result in the same 15% of variance explained.
Thanks for the suggestion. I’ve edited the post.
It may not have saved, it still reads the same way to me (on a different device, so it’s not just cached or anything like that).