throwaway417164

Google CEO Sundar Pichai (note to employees):

> Our co-workers shouldn’t have to worry that each time they open their mouths to speak in a meeting, they have to prove that they are not like the memo states, being “agreeable” rather than “assertive,” showing a “lower stress tolerance,” or being “neurotic.”

Evolutionary biologist Suzanne Sadedin (quora answer linked here):

> It’s true that women and men, on average, have been found in some studies to differ in empathizing/agreeableness, systematizing, gregariousness versus assertiveness; and neuroticism.

Right: it's clear that there's no sharp divide between men's and women's aptitude for STEM jobs. There are many excellent women engineers, and many men who have no aptitude for engineering.

It's also true that there seem to be population-level differences that are connected to aptitude for engineering. If we're trying to answer the question "to what extent is sexism excluding women from engineering?" then these differences become important. If there are no population-level differences then any deviation from a 50/50 sex ratio among engineers is probably due to some kind of sex discrimination. But if there are real population-level differences then it would be a mistake to insist on a 50/50 sex ratio, and a mistake to assume that sexism is the problem if the ratio is not 50/50. (Of course, there might still be sexism, even if there are also population-level differences.)

The problem comes when people try to apply population-level sex differences at an individual level. Even if fewer women have an aptitude for engineering, that says nothing at all about any individual female engineer. The right way to assess an the ability of an engineer, whether male or female, is by looking at the information about what they've done -- they've achieved X qualification, won Y award, built Z product etc. -- not by assuming gender differences that can only be observed at the population level. Unfortunately, both negative and positive discrimination muddy the waters here, making these signals less useful.