Yes, I think it is conceivable. But at the same time, I have a hard time explaining a 10%/90% distribution just with biological differences. It seems to me that the biological differences are too small to explain such a huge skew. In particular, because the gender distribution for programmers was very different in the 1940s/1950s. At the same time, the cultural biases seem so obvious to me, that I think acknowledging them too would be a better basis for further discussion.