They don’t seem as eager to develop these skills

From An 'Ether Of Sexism' Doesn't Explain Gender Disparities In Science And Tech by Isaac Cohen.

Cohen makes the same point I have been making for a while. Gender disparities at the highest levels of performance (awards, patents, tenure, competitive success, partnership, etc.) are generally between 15-30% (i.e. only 15-30% of the "winners" are female) and that that figure tends to be correlated with those that are willing to put in long hours of effort/practice over long durations.

His article is a pretty good source of the current state of knowledge. I wasn't aware of this information.

More recent evidence confirms that men and women of formidable talent don’t make the same educational choices. Last year, women were still seriously underrepresented among those enrolled in Harvard and MIT online courses, including computer science (19%), circuits and electronics (9%), and elements of structures, a physics course with a side of linear programming (5%). Women who do take these courses get the same grades, and they actually have higher completion rates than their male counterparts. But on average, even in the privacy of their own homes and without the pressures and publicity of the classroom, they don’t seem as eager to develop these skills.

This is compelling. One of the hard to empirically measure issues is the feminist argument that there is a social dynamic in classrooms that inhibit the performance of women in science. MOOCs offer a natural experiment in testing that assumption. If you remove the classroom dynamics, as MOOCs do, then you would expect to see an increase in female interest and participation. But you don't see such a change. Women enroll in the same small numbers as they do classes.