I have an anecdotal evidence related to this. In the secondary grammar school, I participated in the so-called math olympics, which consisted of three levels, school level, district level, and then country level. It was not related to "carrier choice", and both boys and girls participated in the school level . Once I managed to advance to the country level (the winners there go to the International olympics), and from around 100 participants, I hardly remember there was any girl... probably a few, but certainly less than 5%. And there wasn't any artificial discrimination, nobody cared if you are boy or girl.

Anecdotally, that high variance seems to show up for a whole lot of things. I'm not sure why that would be the case, but two unfounded guesses would be biological (females spend more resources birthing children thus males are more proactively competing, and high variance is a reasonable strategy) or cultural (it's more acceptable for males, particular young and adolescent ones, to be "obsessed" with some narrow skill or hobby).

With a discussion centered on ability, but no attempt to define this nebulous concept, it seems impossible for this argument to be "true" in any strong sense.

To begin with, this discussion is almost totally void of even the mention of education (the word not appearing once), acting like all children are objects in in the global school array, with a method called "getIntrinsicScore" that gets called a few times in their life before the are casted into adult objects.

However, a simple recollection of your time in school will tell you that not everyone was treated equally by the teacher and the other students, and also that school was a pretty big deal in your life at the time.

So to talk about the correlations in test scores of children and neglecting to mention the influence of the school is not very informed. But since the influence of a school on the test scores of a child is not particularly well understood (witness the debate on "how to fix the schools"), it is really hard to make any firm claims if you include it.

It is also remarkable how the text underlies that the gap in scores appears with age, but does not attempt to explain this observation while advocating a static theory of "ability". The author states outright that there a property fixed at birth that determines your "ability", and then always talks like "ability" and test-score are perfectly correlated.

(For effect I will here do the semantic simplification the "ability" and "test-score" are perfectly interchangeable,because if the implied perfect correlation.)

So, the thesis is that the test-scores are fixed at birth, but the author also accepts as empirical fact that the difference of test-scores appears first around puberty. This is a curious behavior for something that is "fixed". It is highly worrying that this is not addressed at all in the exposition of the theory.

It might be true, but how usefulness is it when dealing with society as a whole?. For the Puttman, it probably kicks in, but for a whole field like e.g. programing/computer science/IT, it almost certainly doesn't for the simple reason that the field isn't composed of anywhere near the top 0.001 percent of the population in terms of IQ.

> but for a whole field like e.g. programing/computer science/IT, it almost certainly doesn't for the simple reason that the field isn't composed of anywhere near the top 0.001 percent of the population in terms of IQ.

If programming requires significantly higher IQ than average, then if male and female IQ variance are disparate then males and females will be differently represented in the programming field.

It's not really an issue of 'kicking in'; in any normally-distributed function, differences in variance are always relevant, and can rapidly be more important than the mean the further one gets from the 50th percentile.