> Men outrepresent women about 2:1 among perfect SAT Math scores. Let’s assume for the sake of argument that this represents a real difference in mathematical ability. Even if programming ability were 100% correlated with mathematical ability, you’d expect a much higher ratio than you see in practice.
That doesn't follow for reasons explained in the article. Choice of whether to go into sciences over humanities isn't correlated with just science ability in school, it is correlated with relative ability between science and reading. The 2:1 ratio tells you precisely nothing about expected ratios of entrance into science programs until you compare it with the performance ratio on the English portion (assuming here, I don't know anything about SATs) of the SATs.
All this also assumes that SAT scores map well to interests and not aptness.
You might do well enough for a PhD in maths and in sociology and still decide try to become carpenter because you love working with wood more than sitting in an office, no matter if you're arguing math or social contexts.
For a woman that might be the hardest career of all these, too (2% of carpenters are women).
That doesn't follow for reasons explained in the article. Choice of whether to go into sciences over humanities isn't correlated with just science ability in school, it is correlated with relative ability between science and reading. The 2:1 ratio tells you precisely nothing about expected ratios of entrance into science programs until you compare it with the performance ratio on the English portion (assuming here, I don't know anything about SATs) of the SATs.