Women are also currently well-represented in computing in Eastern Europe, Iran, China, and somewhat in Russia.

None of these regions particularly care about "women in tech", and most of them are certifiably more patriarchal than any of the Western countries currently suffering from under-representation of women.

One can conclude from facts on the ground that neither gender equality nor pro-diversity movements correlate with desirable representation outcomes. Yet, those appear to be the only approaches attempted here in the West.

We're not serious about tackling the problem. We are, however, quite good at grifting off the problem with proposals we know don't work.

According to an article in the Atlantic, https://www.theatlantic.com/science/archive/2018/02/the-more..., "nations that have traditionally less gender equality [tend] to have more women in science and technology than their gender-progressive counterparts do".

There's a plot of % women STEM graduates vs. a measure of opportunities for women. Algeria has the highest percentage of women STEM graduates on the chart. The Nordic countries have some of the lowest.

> We're not serious about tackling the problem. We are, however, quite good at grifting off the problem with proposals we know don't work.

Could it be that women tend to care about stuff like a healthy work-life balance, and things like the 996.ICU working week can discourage them from applying? Whoops no, we've been told that's sexist wrongthink.

Could it be that computers are boring af and most of these people who genuinely care about computing are somewhere on the autism spectrum (which is mostly populated by males)

We’re all in the spectrum. That’s why it’s called a spectrum. :-)

If women were afraid of long, brutal hours then they wouldn’t make up 52% of all new doctors.

Your comment is great and more observant than most everything in this thread. The answer is freedom, when women have lots of choices they don't choose to play with things as often as men.

>Eastern Europe >None of these regions particularly care about "women in tech"

there are initiatives of promoting IT among women in Poland, afaik.

Women's representation peaked at around 1/3rd in the mid 1980s. While higher than present days, it was never close to parity. The narrative that personal computers were marked to boys is criticized: a significant market segment was secretarial work, which was predominantly made up of women.

The drop in women's representation in computing corresponds to the elimination of overtly sexist policies in other fields. E.g. medical schools removed gender quotas (maximum not minimum quotas, sometimes as low as 10%). Mary Ann Wilkes initially wanted to be a lawyer, but changed to computing when she encountered explicit sexism in law firms. After the expansion of women's right, women in her position were able to pursue their original passions.

I doubt it boils down to any one cause, but I think it's crucial not to ignore women's agency in making career decisions.

I know quite a few female programmers. Almost all of them wanted to be something else, and only became programmers as a second choice.

Another thought: compared to other technical careers, programming careers are uncredentialed. No legal or even social prerequisites other than skill.

If you want to be a doctor, you must apply to college, be accepted to college, matriculate, graduate with a degree in sciences. Then you must take the MCAT, apply to medical schools, be accepted to a medical school, graduate, apply to a residency, complete the residency, and become board certified.

And then you are an entry level doctor, licensed to practice in a single state, with no job, no patients, no practice, and no insurance network contracts.

If any of those steps are sexist/racist, a woman or minority is less likely to make it past that step.

Programming doesn’t have as many gatekeepers. That’s good for people who have been kept out of things.

Any history of the industry needs to address what analysts did vs. what programmers did. “Programming” was clerical work, converting flowcharts or pseudocode into machine code on cards or tape. That role was obsoleted when computer time became cheap enough to spend running compilers and linkers, which enabled analysts to write code by themselves in a higher-level language.

It wasn’t until later that obsessed nerds could teach ourselves from scratch at home.