There is a lot of anecdotes in the comments about these interview processes.
The ones with the best data are the company themselves, especially those who interview thousands of candidates a year.
These companies (e.g., all FANG) have enough data to assess which interviewer, which interview questions and which interview process have positive correlation (and even causation) with future employee performance. And they can optimize the pipeline to interview most candidates with the interviewers that can actually assess future employee performance for instance.
Either these companies have already done these optimizations and analyze the trove of data they have, and the process is what it is for a reason. Or they haven't yet and there is still much room to optimize the process, and applicants have a chance to see those processes updated for the better (although here better means better for the company, not necessarily better for the applicant).
There is no data could collect that would tell them whether they are efficiently sorting good candidates from the herd.
Say that I wanted to sort men from women to only identify women. One thing I could do is have the candidate walk under a bar that was 5 foot 4.
Most of the people who would get through would be women. But this would not be an efficient screening method. I would still be losing many women to the screening method.
If I am only polling the women who get to the end of the screening to see if they are women, I can't evaluate the efficacy of the screening. Facebook/Google would need to hire some of the failed candidates to compare results.
There is literally no way right now for any company to assess the false negative rate, i.e. how many well-qualified engineers they turned away. They can only assess internally the false positive rate, i.e. how many engineers they thought were going to work out that did not. The fear of making a "bad" hire, is so pervasive that we optimize entirely to minimize the number of false positives, completely ignoring that making the interview more and more difficult is probably having a marginal effect on reducing the FPR but is drastically increasing the FNR.
The ones with the best data are the company themselves, especially those who interview thousands of candidates a year. These companies (e.g., all FANG) have enough data to assess which interviewer, which interview questions and which interview process have positive correlation (and even causation) with future employee performance. And they can optimize the pipeline to interview most candidates with the interviewers that can actually assess future employee performance for instance.
Either these companies have already done these optimizations and analyze the trove of data they have, and the process is what it is for a reason. Or they haven't yet and there is still much room to optimize the process, and applicants have a chance to see those processes updated for the better (although here better means better for the company, not necessarily better for the applicant).