4 Comments
User's avatar
Ferien's avatar

>and genetically correlated to the point where one can use one as a proxy for the other.

even if underlying genetics is same, phenotypically they are different. We already knew even before any genomic studies that rich parents could buy education for their dull children, but not IQ. EA is IQ passed thru non-genetic filters "add noise" and "average by sliding window".

Expand full comment
Sebastian Jensen's avatar

>(Why did they forget to remove one half-sibling on the left side?)

hehe

Expand full comment
Guy's avatar

"If siblings are dissimilar, say, one is tall and one is short (relative to age and sex norms), the product will become negative (2 * -2 = -4)."

But what if one is super tall and one is average? (4 * 0 = 0)? And yet the actual difference between siblings is just as large as in your example. I don't get this method.

Expand full comment
Emil O. W. Kirkegaard's avatar

Yes, it's definitely unusual but apparently works in their simulations.

Expand full comment