My only real problem with this is that I do not really care about what is achieved in children's mathematics classes. Surely, the better test of true ability in this field would be to look at researchers and research accomplishments.
Of course it is subjective, it is is me that does not care about children's mathematics. I make no claim for the wider world. That's why I said "My only problem" and not "the problem".
Its not entirely arbitrary though. I care about research mathematics as to do well at it(or to do it at all) requires very strong mathematical abilities. To do well in classroom mathematics is as much, and probably more, to do with being a good student.
So it seems to me that by looking at children's mathematics rather than research mathematics your results are going to be confounded by the issue that your measuring mathematics ability via a proxy(schoolroom mathematics) rather than the real thing(research mathematics). Your results might still be correct, but I find it hard to know if that is the case.
I suppose, if your just interested in day to day average numeracy skills then looking at children is fine, except even then I would prefer the study was performed on adults rather than regularly drilled students. That way you will find out who is more capable of retaining numeracy skills and who has the higher natural levels of numeracy, rather than just finding out which students revise the hardest and listen to the teacher most attentively.
It'd be interesting to see how these disparities change (or not) with age and through into university study. I've always heard that girls mature faster, which may mean that differences in ability by sex that appear in childhood are different in adulthood.
Completely aside from that, it does show how broad societal trends can affect maths learning for females.
