I don’t think it assumes that the candidate is in any way not as qualified. It assumes the candidate is not as likely to get hired because of bias in the hiring process.

If the pool of candidates doesn't itself have the race or sex proportions that you're hoping to hire, then you intrinsically have to lower your standards at least for the underrepresented group.

Then you end up with a wider range of quality levels among the underrepresented group (because you had to scrape deeper in the barrel to fill your quota, and the overrepresented group ends up being in the higher quality band (because scooping the cream off the top filled the quota).

This has the really perverse effect of creating a true observation (in the context of your organization) that the average quality of the underrepresented group is lower than the average quality of the overrepresented group. e.g. the asian men you hired are all the cream of the crop and the black women you hired are everything from the cream to the dregs.

I suppose you could artificially hire lower quality overrepresented groups to correct for this, but then you're really setting yourself up to be outcompeted by companies that only hire the highest quality employees they can.

"If the pool of candidates doesn't itself have the race or sex proportions that you're hoping to hire, then you intrinsically have to lower your standards at least for the underrepresented group."

This is uninteresting thinking. If the pool doesn't have enough qualified candidates, you expand the pool. That requires more effort. That's good effort that should be taken. It's lazy to do otherwise.

But you can't unilaterally expand the pool. You can't force people to do jobs they don't want to do. At least not outside of forced labor (this is part of why Israel has above average women in STEM, since their conscription forces women to gain technical experience in the military). Some suggest that underrepresented demographics should be attracted with higher wages. But that's illegal, you can't pay people differently for the same job.

You expand the pool by sending your recruiting team to more colleges. This isn't hard unless you want to make it hard.

And how does that change the demographics of the applicant pool? If I get 80 men and 20 women applying, and I send my recruiters to twice as many colleges I'll probably get 160 men and 40 women. The proportional representation is unchanged. You can get more applicants, but you're drawing from the same pool that's 80% men and 20% women. I could direct my recruiters to exclusively advance women to the interview stage, but that's illegal.

There are a few universities where there are above average proportions in STEM fields like MIT and Harvey Mudd. But these are few and far between, and their graduates have heavy competition for offers. There aren't enough graduates to meaningfully make a dent in representation. Furthermore, these universities have big gender disparities in admissions rates - over two to one [1] - so directing recruitment at these universities is still achieving "better" representation through discrimination, just indirect discrimination.

1. https://www.collegetransitions.com/blog/can-your-gender-give...

> I don’t think it assumes that the candidate is in any way not as qualified. It assumes the candidate is not as likely to get hired because of bias in the hiring process.

That's the sneaky part. There's an assumed "bias," but it's rarely rooted in fact or backed by statistical evidence and more of a religious statement. At best, it's admission of a guilty conscience [1].

[1] https://www.youtube.com/watch?v=XASNM1XEQPs