The fact that the quit rate is higher doesn't disqualify the pipeline problem as an explanation. It sounds like your assumption is that a hostile work environment must be the reason for the higher quit rate, and thus that fixing the pipeline wouldn't fix the issue. I'm not sure this is as ironclad as you're pretending it is.

I personally know people from underrepresented groups in tech (African-American, women etc) who attest that, even in work environments that made them feel 100% welcome/comfortable, they felt a mental cost from the very fact of being underrepresented. That is to say, even if you were to remove all the consciously and unconsciously biased behavior from workplaces, some people still have a negative reaction to being the only present member of a group that they identify with. The pipeline problem is the only solution that could possibly address that[1], by addressing the imbalance per se.

[1] I suppose short of having people stop feeling this way about being underrepresented. I've been underrepresented ethnically in almost every group I've ever been in, but it has never bothered me. Far be it from me to tell other to stop feeling a certain way though.

> It sounds like your assumption is that a hostile work environment must be the reason for the higher quit rate, and thus that fixing the pipeline wouldn't fix the issue. I'm not sure this is as ironclad as you're pretending it is.

I mean, the data does seem to support a tech environment that applies disproportionate pressure on women to perform certain activities. From the linked HBS study, for example (Figure 6.1), 71% of women in tech feel pressure to put in face time (vs. 19% in all sectors) and 36% feel pressure to be available 24/7 (vs. 26% in all sectors).

Both of these are time-consuming activities. Myriad studies show child care responsibilities, even today, still fall disproportionately onto women. Child care is time-consuming, and that comes into conflict with time-consuming professional work.

Of course, not all women want children. And this is one study. But it wouldn't surprise me if women were underrepresented in tech due to a combination of cultural issues and the pipeline problem. It's just extremely reductionist when people assign blame to the pipeline problem, since that shifts the focus from work environments to higher education. That's a convenient result for companies.

Thanks for the data. That's a pretty compelling perspective actually: policies that are less compatible with family care in a given industry have a disproportionate impact on women in just that industry, which increasing representation (through pipeline issues) won't necessarily fix.

I am not sure what that quit rate really means seen in that isolation.

What is the quit rate of women in other sectors ? What is the quit rate in high-tech in general ?

All those questions are very important. My wife left her job after marriage because she had to relocate and later she took a break for childbirth while I was at the same job.

"Pipeline problem" is not really a problem. It is not the case that women will suddenly come to CS if we paint classrooms pink. If the reason women are not attracted to CS is probably those young girls have done their research an decided that they will be better off being something else other than CS major. Hostility at workplace might be one of the reasons but I highly doubt if that is the only one. At least we need data to indicate that quit rate is not a useful metric.

How does it compare with the quit rate in highly-paid but non-high-tech occupations such as legal, middle/upper management, aviation, medicine or finance?

One would assume there's a natural attrition across all industries due to women taking time off to have/raise kids as well as some common base level of quitting in general due to reasons common across all industries (better pay elsewhere, bad managers). Is 2.5x coefficient very high, very low or just about same as other industries?

> Sorry, so you're suggesting the problem is that women don't study computer science? Hmm. [0]

The reference you cite says that

> about 214 women are majoring in computer science, accounting for about 30 percent of majors in that department

and

> only about 18 percent of all U.S. undergraduate computer science degrees go to women.

While the WSJ article states that

> Women made up 33% of its global workforce, up from 31% in 2014.

So, exactly a pipeline problem.

It's interesting that you have not addressed the high attrition rates for women who do enter tech. Is that also a pipeline problem? Should schools also teach women to tolerate hostile work environments?