Do broader impacts corrupt science?

NSF proposals are divided into an “intellectual merit” section on how the proposed work will advance science and a “broader impacts” section that details how funding will help the PI do public outreach, build infrastructure, train postdocs, involve underrepresented groups, and other such things.

One may think of “corruption” of a practice as when external standards are allowed to trump standards intrinsic to the practice.  For example, it would be a corruption of science for a man to be denied the Nobel prize because of his race, political beliefs, or status as a convicted child molester.  It would be an even bigger corruption for scientific theories themselves to be judged on their implications for race relations, politics, or child molestation.

The question is, do broader impacts corrupt science?

Not necessarily.  Research and teaching have gone together since universities were founded, so the connection is likely a natural one, and it is easy to argue why this is so.  The point of science is to bring undiscovered truths to the public, and this requires both making new discoveries and communicating them effectively.  It’s also easy to see how training new scientists benefits the enterprise as a whole.  And if one can go this far, why not reward scientists for making efforts to share their findings with the general public, whose taxes after all are paying for everything?  Making modern science accessible to the public, I would add, involves a nontrivial amount of time and effort.  It seems fair that an organization like NSF should take this sharing with the public as part of its mission.

There’s also a “broader impact” that PIs don’t write much about, but it’s actually the most important one from program directors’ points of view.  If just doing the science were the only criterion, maybe it would be more efficient to just give all the money to Caltech and MIT.  In fact, NSF tries to spread things around to keep smaller research groups running throughout the country.  I think this is a wise policy.  It’s best not to put all your eggs in too few baskets, and it’s good that students all over the country have a chance to participate in real scientific research.  Sometimes, really good students even come out of weaker undergraduate institutions.

But if we can justify explicitly favoring geographic diversity, what about “underrepresented groups” (women, NAMs)?  Surprisingly, I think special outreach targeted to certain groups could be justified in a non-corrupt way, but the justification does matter.  If the worry is that, for instance, hispanic children aren’t exposed to science as much as other children, I don’t have a problem with NSF rewarding a PI who wants to rectify this.  However, the effort should not be to achieve equal representation, a goal that strikes me as clearly corrupt.  Let all have accessible ways of learning about the results of science, but removing differences in interest or ability–even differences that are cultural in origin–should not be our job.

I actually could make my proposals look a little better if I gave demographic information on my research group.  It’s been half female at times; I’ve had two Middle Eastern and one Central American student.  Leftists would be jealous.  But I’m not going to point this out, because I have to look at myself when I trim my beard, you know.  And anyway, I’m not paying much for this stand on principle, because everybody knows that, except for CAREER proposals, the “Broader Impacts” section is something you can just blow off.

The danger here is that we all know that social justice tends to take over every institution, subtly transforming it until all the original goals are replaced with social justice goals.  If you love your practice, you must fight this process.  But if what I’ve written above is true, it’s very hard to draw the line where the legitimate internal standards of the scientific practice end.  How do you know when your real goal has switched from sharing the joy of discovery as widely as possible to rejiggering the ethnic makeup of the STEM disciplines?