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.
These are the three biggest countries (40% of the world population), and also the ones most represented in tech here in the USA, at least (because they are the biggest).
And yes, there were proportionally quite many female engineers and computer scientists. I might guess that this perhaps was facilitated by prejudices/discrimination in the "more macho" fields of construction engineering and industrial engineering, so all the high school girls with good STEM results would go to math and computer science fields instead of all the programs with a focus on heavy industry - but even those had a decent proportion of female engineers.
We are talking about whole population segments that dream of one day making $75k. They know that tech pays well but have no idea how outlandish the pay is compared to their experiences and expectations. They think their counterpart is tech is making $10k more than them, when in fact they are probably making $50-100k more.
