As a math and science teacher, I've seen multiple studies on performance of different genders in math and science. There is a gap in North America, although it's closing rapidly. (In the past 40 years, men have gone from having 20% higher averages than women to having 2% higher averages than women. Evolution doesn't act that quickly; it's a purely social bias.) Men still perform slightly higher than women in this region because there are still teachers out there who expect more from male students and push them harder. In other words, if the teacher *expects* female students to get 60s and down and *expects* male students to get 70s and higher, then that teacher who sees a male and a female student with 68% averages, then the teacher will work with the male to improve his performance, but not put in the same effort with the female student. It's a horrible thought, but it's still happening out there. The same is true for race factors, for "learning disabilities" (which I would rather call "learning anomalies" but that's another story) and more.
Bottom line: there is a slight and closing gap between men and women in math and science in North America, not because there is any biological difference in this particular area, but because social biases that exist in the system are failing the female students more often than they are failing the male students.