[E.J.] Antoni made the remarks about intelligence in a discussion with summer interns at Heritage in 2024. He said that women’s IQs generally clustered around average scores, while men’s IQs varied more between “geniuses” and low-intelligence individuals, said the people, who spoke on the condition of anonymity for fear of retribution.
I understand that not all studies confirmed these differences but nothing here seems as shocking as some are claiming. I thought it was generally established that male IQ scores had basically the same mean as female IQ scores but that the standard deviation was greater and a higher kurtosis was particularly interesting.
@Dan_Eastwood, our local biostatistician, probably deals with kurtosis more than any of us.
I can’t do a deeper dive here, as my “spread thinness” is statistically indistinguishable from @Dan_Eastwood’s but an interesting biological fact to add here is that males have a single X chromosome while females have two. A simple prediction from that fact is that traits affected by the X will differ, statistically speaking, between females and males. It’s not really about dosage, due to X inactivation in females, but to the fact that females have two alleles to choose from while males (poor disadvantaged wretches that we are) have one.
Does that explain the IQ statistical observations? I’m not at all sure but it’s probably one of the first places to start. I recall that there were claims that genes implicated in IQish stuff were well represented on the X but I don’t know if that has held up.
I had understood that that was indeed the consensus—but the media reporting so far (which certainly can’t be trusted) is that the differing IQ bell curves are still up for grabs. Nevertheless, I do recall reading an article long ago explaining that a male’s single recessive gene on the X chromosome that negatively impacts cognitive function will most likely be expressed, as there is no second X chromosome to mask its effects. Makes sense to me.
Fun(?) Fact: The variance of kurtosis is on the order of \sigma^8, which makes it nearly indistinguishable from anything outside of ridiculously large samples. My masters thesis (now with a whopping 4 citations) dealt with an alternate method (L-moments) of fitting distributions to skewed/kurtotic data.
but an interesting biological fact to add here is that males have a single X chromosome while females have two. A simple prediction from that fact is that traits affected by the X will differ, statistically speaking, between females and males. It’s not really about dosage, due to X inactivation in females, but to the fact that females have two alleles to choose from while males (poor disadvantaged wretches that we are) have one.