Striking graph. Having been there at the time, I'd propose roughly the same explanation she does, though in less loaded terms. The decline starts about 5-6 years after microcomputers became widely available. Assuming 4 years to graduate, that means the decline in women CS majors begins almost exactly at the time kids started to be able to hack in HS.
Up till 1980 or so, CS was like other engineering disciplines in that it was something most students started to learn in college. Few students arrive in college already proficient in mechanical engineering. But starting in the early 80s, students started to arrive who were already proficient (or at least, experienced) in programming. These kids were self-selected, and they were almost exclusively boys.
I once had an enlightening discussion on a company mailing list about this. The common complaint among the female programmers was that their university classes were directed towards these self-selected male know-it-alls, who had been overclocking their PCs since middle school. Many of the women, who were just making their first foray into Comp Sci in university, got so frustrated they almost quit. And these women eventually became great programmers, so imagine what the women of average talent experienced.
The funny thing is, when prompted, quite a few of the men chimed in that they resented these male uber-nerds too.
So maybe society just forgot how to teach computer science or programming to adults. It might be that the schools have a reliable business, graduating the inveterate nerdz-4-life.
And maybe that's why practically every student I interview codes like they didn't learn anything in university. (But that's another story.)
When I was an undergrad at MIT, rumor had it that the biology department treated the high proportion of women biology majors as a topic of concern.
When I was an education major doing student teaching, I was told to my face that it was good for the kids to have another male teacher because I would be providing another positive male role model.
This trend of girls outpacing boys is apparent in the US even in elementary school. This is a big concern for a growing minority of educators. The gap widens as time goes on. We may be starting to see the effects of this current phenomenon as the first of these students enter higher-level education.
http://parentzone.babyzone.com/school/a2316 contains an article summarizing the situation. The general tone I hear is "we've saved our daughters, but are failing our sons."
Because the split there is about 55:45, which is close enough to even to be a wash. If men were 20-25% of the workforce in the life sciences (as women are in engineering and CS), then yes, it would be a problem. But they're not, so it isn't.
I'm always surprised at these claims of women getting over 25% of all CS bachelor degrees. I've given a couple of one day courses to CS students and there weren't any women. In my college days there was one girl (and she considered getting a masters in CS a hobby next to her main occupation, which was med school.)
At conferences with hundreds of attendees I see less than a dozen women, more than half of whom turn out to be the friends or sisters of the organisers.
Is that just my experience? I'm Dutch, but this includes my experiences at JavaOne, RailsConf and EuRuKo.
In many countries, such as India, there is actually a rise in women in Computer Science, probably more to do with the fact that in most cases they start off on an equal footing to boys in terms of exposure to computers pre-college and secondly due to the lucrative (in relative terms) job prospects and non-manual nature of the work.
The trends there seem to be following the early trends of the graph here (but in a more condensed timeframe), and it will be interesting to see how it continues.
I chalk this one up to the anti-intellectual barbie doll culture here in the U.S. There are plenty of girls studying computer science and engineering in Asia, for example, where academics are valued over sports and fashion attire.
How does this account for the continued growth of women in other technical fields? One of the biggest points in the article is that this trend differs strongly from many other intellectual areas.
Well, this is only an anecdote, but a close female friend of mine went to college to study CS. After two years in the program however, she switched majors. The vaguely hostile social atmosphere and the impression of a programming career as lonely nights sitting at a computer monitor contributed greatly to her decision. She was very competent in the field but those external factors were enough to push her away. It was probably the right decision for her, but it was definitely a loss for computer science.
Up till 1980 or so, CS was like other engineering disciplines in that it was something most students started to learn in college. Few students arrive in college already proficient in mechanical engineering. But starting in the early 80s, students started to arrive who were already proficient (or at least, experienced) in programming. These kids were self-selected, and they were almost exclusively boys.