Well this might be taking it a step too far. The first, obvious step would be simply to compare the actual hires (workers) with the available pool... almost nobody actually does that. People keep claiming that e.g. Google is sexist because they hire so few women, without regard to the fact that their worker pool has about the same distribution as their hiring pool (CS graduates). The UK passed a law requiring companies to disclose average salaries by sex etc. with no regard on the actual job / education level / level of employment.

We can't even begin to talk about solutions if all the media / researchers / activists keep pushing misleading statistics.

> The first, obvious step would be simply to compare the actual hires (workers) with the available pool... almost nobody actually does that.

I absolutely agree! (And that's also why I said in my post that the percentage should be determined by the applicant pool.)

There's a separate conversation to be had about why the applicant pool is skewed, but IMO that's well outside a given company's purview.

Why do we need to systematically counteract our biases? Why is that right thing to do? And who decides what is right?

I was biased when I married, I’m biased when I pick people I’ll spend time with, I’m biased when I apply for a job, and wouldn’t like it to be any other way... Of course, unless it’s me deciding for everyone else what they should optimize for their decision making process.

Because people in power have said you must or face financial ruin, freedom of association be damned.

There is an alternative explanation to the biases people are seeing here, unfortunately, and it would render gender and age confounding variables.

It's known that Valley tech firms of the type studied in this paper have a large left wing contingent amongst employees. We know that being female with a degree correlates quite well with being left wing, as does being young. In contrast being an older white male correlates somewhat with being conservative.

If the actual discrimination here is against "people likely to be conservative" - which is absolutely likely to be happening given what's going on in these firms - then setting quotas for gender and age wouldn't be addressing the underlying problems, only symptoms.