Interesting what you write, too, about the lower end of variance. I’m not sure about this. Again, I think boys’ socialization in relation to school, peer group dynamics, and other issues may drive this ratio into extreme disproportion. Even tendencies towards physical confrontation or aggression style might get more boys in trouble at school and lead to expulsion or other performance-deflating consequences. The difference might even be ‘biological’ but it would not be a variance in ‘math ability’ or another way of talking about overall academic performance. Part of the issue still seems to me to be that ‘math ability’ or IQ or ‘intellectual potential’ would be all be synthetic abilities, composed of a number of variables.

I think I’m less suspicious of the ‘biological variance’ part of the argument, that boys might be more variable than girls, than I am of the notion that math scores illuminate genetically predetermined ‘math ability.’ There are so many other, less speculative ways of talking about the test score variance. ‘Math ability’ as a yet-to-be-found genetic endowment seems like a convenient intellectual rug to sweep this question under, instead of taking the next step and starting to pull apart the notion of ‘math ability’ into something that might be more carefully and empirically examined.

Greg

Of course, if we are using IQ as some measure of something, we don’t *really* know the higher order statistics of it, right? And I’d imagine that most of the variance difference at the higher end on tests like this are due to cultural factors or are so plastic in response to cultural factors and work ethic that it might well end up being meaningless.

Oh, and I’d say the same thing about the reading gap between men and women. I never understand why I hear so much less about that. I’ve assumed it’s because those tests are partially a function of how much you read, and women tend to read more? But I don’t know.