Outcome: The women working as programmers at FooSoft make on average 10% less than the men working there.
Possible reasons:
1. Lack of opportunity (e.g. people in power at FooSoft have bias against them makes it harder for them to have their contributions valued)
2. They just aren't as good as the men
3. They...have different interests in some way that apparently doesn't make them worse but does make them deserve to be paid less?
A quick search suggests that only 20% of programmers in general are female, but 30% of Google programmers are female. I'm not certain these numbers are accurate, but if they are, female programmers are over represented at Google.
That would mean Google hired a greater percentage of female programmers than male programmers. That is, if Google hired the top 0.1% of male programmers, they hired the top 0.15% of female programmers.
If we assume two groups have identical distribution of qualifications, the female programmers hired would have lower qualifications, as an obvious consequence of a hiring policy that favors women. That's not saying women are less qualified, it's saying that Google hires women who are less qualified because they want to hire more women than the 20% that sex-blind hiring would result in.
Going by the "logic" of outcomes being the "only holistic measure" this would indicate that Google is biased against male programmers within the context of the pool of available candidates. Yet if we change the scope of our consideration to that of the general population, Google would be biased against females.
Seems like the "only holistic measure" has some kind of issue with logical consistency. One has to be careful with how the population sample relates to the specific pipeline of available candidates. If one isn't careful, then the interpretation of the numbers might as well be fiction. You can't just stop at, "Does it match the general population? No? Then BIAS!"
That would be expecting Google to hire untrained people as programmers, which would be absurd.
If we compare with the general population, any bias found wouldn't be at Google, but in the school system, children's entertainment, parenting, or some other group that influences people long before they apply to Google.
The pool of programmers "In general" is not relevant for Google. They don't hire programmers in general, but are focused on a specific subset (usually from some big universities).
I think my hypothesis is simpler:
- Women are discriminated against in the industry as a whole
- Google actively discriminates in favor of women
It's discrimination both times.
When Google (and other top tech companies) actively discriminate in favor of women, that causes the appearance of discrimination elsewhere in the industry, because the top tech companies have taken more of the best female programmers and those who weren't hired by Google et al are, on average, worse than the men who weren't hired by Google et al.
And when I spoke about interests, I was discussing why few women choose to study computer science.
I would expect in 2019, that the industry as a whole discriminates in favor of women, but I'm just a layperson when it comes to that. Do you have citations?
- Google actively discriminates in favor of women
It's discrimination both times.
I agree.
Outcome: Women working as programmers at FooSoft comprise 10% less of the workforce than men working there.
Possible reasons
1. Hiring bias
2. Not as qualified
3. Less applicants due to disinterest in programming as a vocation
