[mrosethompson]

Not necessarily. I'm not presenting this as fact, but merely as a a possible situation and counterexample to what you describe:

Your reasoning would hold if you could assume identical skill distributions between men and women graduates. However, it seems somewhere between possible and likely that women graduates are of higher average skill than the men graduating. With so many barriers and cultural impedances for a woman pursuing tech, it takes at least the same skill (to pass the classes) as their male classmates, plus enough extra competency to ward off the detractors and skeptics that men don't have to deal with.

Again, I'm not saying this is reality, I certainly don't have the statistics to back it up. But it fits my experiences pretty well, and I think your analysis probably is too simplistic to be useful.

And, in regards to your second point, about starting much earlier: I absolutely agree. The more the culture changes, the less this will be an issue.

[Timeroot]

As a possible countereffect -- not one I necessarily believe to be stronger, but one to consider -- is that women might be weaker due to the culture they're coming from. Plenty of males have been encouraged to play with computers from the age of 8 or 10, and so will have large amounts of [at least some kind of] experience coming in to the workforce. Women are far less likely to have had this experience. Additionally, males (who will have more male friends) will quite possibly have plenty of coder friends whom they have learned from. Females statistically will have less, and have less exposure this way.

i.e. the culture not only makes it harder for women to get through learning compsci, but they'll also have less exposure to it in the process. This could potentially make them weaker candidates.

I don't know to what degree that actually happens.