Implicit association tests probably don't work. Sources may be found at https://slatestarcodex.com/2016/08/25/devoodooifying-psychol...

Well, it's a little more complicated than that. They have extremely high state-specific variance, so it's difficult to show predictive validity. They have definitely been associated with particular sets of outcomes in non-gendered/racial contexts, where you would expect the studies to be easier to run.

Source: I did a bunch of them for my PhD, but my reading of the literature stopped in 2014.

The suggestion here is that even though companies can save a bunch of money by flooding their workforces with women, they are all too obtuse to do so.

77 cents on the dollar is a huge discount. Any company that exploited this would gain an enormous advantage. So even if nearly everyone is somehow blinded by their biases, surely at least one company somewhere is smart enough to discount their biases and collect 23% in savings. Where are they?

Is it possible that some company out there with unscrupulous management has already tried moneyballing the job market this way, and found it didn't work?

> Would you be more likely to hire a man than a women for an engineering role?

If they are equally good, there's a good chance I'll choose the cheaper one.

Indeed, that's what we like to tell ourselves. But is that actually what's happening? Did you take the test?

Happening what? 100% equal candidates that differ only by gender? No, it's never happening.

Can you do an implicit test when it's nature is made explicit?