One thing all these fields have in common is that they are more intellectually rigorous and harder to succeed in than the computer software industry.
Clearly, they have something else in common. We just need to figure out what it is.
This essay, which invokes the "Google Memo", is subtly attacking a straw man. Even those almost the entire rest of STEM is better than CS, it's true that it's not balanced; it remains deeply imperfect. Physics and mechanical engineering, clustered with CS, remain the province of men. There's a expanse of STEM fields with female participation between 25-40% that you'd want to explain or correct. Is it stereotype threat? Implicit bias? Who knows? Probably not?
But that has nothing to do with why Google has so few women engineers. The work that a commercial software engineer does --- even at the lofty heights in which the profession is practiced in such a cathedral of software design as the Alphabet Corporation --- is simply not that hard; most of it is just wiring form fields to databases in new and exciting ways.
Whatever is holding women's participation in our field at or below twenty percent is artificial, and a travesty.
Have you also considered a possibility that there is no explanation, and the gap is simply due to internal variability in the underlying dynamical system?
Although personally I favor the "different interests" explanation, let me expand on this a bit. Humans love explanations, even of random events, that's how superstition comes about.
There was an interesting biological experiment done with ants. If you put two identical piles of food the same distance from an anthill, you would intuitively expect that ants will eat from each pile in 50/50 ratio. However, it's not what happens; in reality, the ratio fluctuates over time, at points being 20/80 and can reverse to 80/20.
Why? Because behavior of each ant depends on strength of pheromone path to food that is produced by other ants. In other words, ants, to some extent, copy behavior of other ants. And this copying is enough to produce a large difference in ratio even in cases that are objectively absolutely equal.
So is it hard to conceive that humans also copy behavior of other fellow humans (actually, the above example is from Paul Ormerod's book Butterfly Economics), and that women might prefer a job where already are other women? Or where isn't an overwhelming majority of different people (i.e. men)?
The influence on individual decision can be very subtle, but yet can, statistically, lead to large differences in outcome.