Interesting observations, Cathy.

In my (quick) read of the tables in chapter 5 (e.g. page 105) one other thing stood out: a larger fraction of women than men are in academic positions, at least through 1995. The data aggregate everyone who received a PhD from 1970 or so through 1995, and a lot has changed over that time period (though not necessarily the ratios in question). I'd be interested to see some individual cohort data for more recent graduates.

I tried to find further information on the two summary points you cite from page 221. It looks like the main supporting data for the first point (more women than men being "lost") has to do with the 1970 and 1973 PhD cohorts (e.g. pages 78 and 86). I'd be surprised if the differences were still that big for recent cohorts. Interestingly (at least to me) is that the differences in the percentages out of science are not actually all that big in some fields (e.g. math, life sciences, p. 78). The real story is differences in those working part-time or not at all. Judging from p. 90 and onwards, those differences have mostly to do with young children and cultural expectations for who cares for them.

Looks like I need to dig into this report at some point.