Another (Slightly Less) Innocent Question for AALS Interviews

So in my first installment in the gripping series "Innocent Questions in AALS Interviews," I suggested the question: "At [Your School], what percentage of the job are the different pillars of research, teaching, and service?"  But the commenters were right that it's probably not the best question to ask. I referred to the question as "moderately innocent," and the qualifier is because I did have something else in mind.

Which is this, my next, slightly-less innocent question: "How do you determine who to hire, and what information do you use to make judgments on the criteria you think relevant?"  OK, so in case it's not obvious, this is NOT a question one should actually ask. Because if one did, well, to quote the college football announcer Keith Jackson, "Whoa, Nellie."  Get ready for the awkward silences and people nervously looking at each other. Not so much because that's an unusual question to ask (though it is), but because the hiring committee may lack a consensus answer.

Here's my perhaps-overly-rationalist view on what the answer to the first half of the question should be that I offer more to spark discussion than with any kind of certainty that this is right. To simplify, assume you've decided a particular curricular need (as most schools do for most slots), just entry-level, some "bump" for diversity of whatever kind, and have narrowed candidates to 15-20 people through some combination of resume/publications.  Then what? (1) Take the faculty's view of what percentage of the job each pillar is (say 55% research, 35% education, 10% service), or should be for new hires. (2) After the AALS meat-market interviews and committee discussion, each committee member rates each candidate on a scale of 1-5, or 1-10, three times: for their potential for excellence in research, education and service.  Each member's ratings is shared with the rest of the committee. (3) The chair aggregates and appropriately weights the committee's ratings, and uses the results to decide who to invite back in.

Anyone know of committees that use this kind of approach? If not, what do they do? I realize there are political reasons why this might be resisted, but as an ideal, what's wrong with it? Separate and big issues on the second half of the question -- what information to use to predict future success -- particularly on education and service pillars. Likely a bit easier these days with research since you can just look at existing publications, and listen to their ideas for future work.