That's the gender-equality paradox, STEM version: the more gender-equal a society, the more the genders self-segregate occupationally.
It's important that this is not about any absolute value, but about the sign of the correlation. The "blank slate" hypothesis would predict that the correlation is positive, i.e. the more gender equal a society, the more equal the distribution.
The opposite is actually the case, the correlation is negative.
This strongly suggests that the statistical differences in occupational preferences have an innate cause. That difference is moderated by societal influence, not caused by it. When societal pressure are lessened or removed, the innate differences manifest more strongly.
And of course, it needs to be stressed that these are statistical differences, not categorical ones, just like the outcomes are statistical and not categorical.
It's important that this is not about any absolute value, but about the sign of the correlation. The "blank slate" hypothesis would predict that the correlation is positive, i.e. the more gender equal a society, the more equal the distribution.
The opposite is actually the case, the correlation is negative.
This strongly suggests that the statistical differences in occupational preferences have an innate cause. That difference is moderated by societal influence, not caused by it. When societal pressure are lessened or removed, the innate differences manifest more strongly.
And of course, it needs to be stressed that these are statistical differences, not categorical ones, just like the outcomes are statistical and not categorical.