Is there any published research-level evidence that FAANG interviews are high-signal?

I'm no expert, but the only published research I've read on this is "Thinking, Fast and Slow", and the original researcher has recently cast doubt on his own results on that study. Regardless, the evidence there was that all that mattered was "any objectively measurable metric" -- which would mean that FAANGs are wasting a lot of time and money on useless interviews. (Of course they have money to throw away, but still...)

I worked at one of those companies. They were very self-certain that their hiring was high signal. They were likewise completely disinterested in doing a control group test where we randomly hire people from the fail pool and see how they compare once hired.

You mean, beyond the companies being successful and effective, and famous for the skill of their workforce? Beyond 9 of the top 10 most valuable companies in America being tech firms that all interview in similar ways?

What kind of evidence that tech interviews work well are you looking for, exactly? Isn't the obvious evidence of your own eyes sufficient?

Not the above poster, but if they only hire 10% of their interviewees, that's a sign that not every interview is a "close call', which is the context of an interview being high signal here.

>the original researcher has recently cast doubt on his own results on that study

Do you have a reference for this? I'm interested in reading more.

This HN post/discussion from 2017 [0] may be of some help.

[0] https://news.ycombinator.com/item?id=15228712

The phone screens are terrible signal. They intentionally give everyone who can spell Java or C an on-site to get multiple opinions. No one wants to be the single person who commits to a No Hire.