[dtdynasty]

Why wouldn't stem PhDs follow some bell curve of quality? I'm sure many PhDs that are leaving don't contribute but some of them do. I personally don't see a reason for it to be skewed for only PhDs which don't contribute to leave.

[recursive]

These things are not in conflict. It's possible that PhD quality has a regular distribution, and that most of them aren't contributing much.

[dtdynasty]

I agree with this. I guess you already believe they follow a bell curve.Then from your former comment you also believe it's worth it to lose many PhDs that don't contribute to also lose the few that do.

I guess the conflict is my value judgement that it's good to keep PhDs that don't contribute if it allows US to keep the ones that do contribute. I believe so for 2 reasons.

- Distinguishing between contributors and non contributors at scale is difficult.

- the value of research can be large from a few contributors.