I was spammed by the NSF multiple times to fill out some silly little survey on their graduate student fellowship program (GRFP), so I got annoyed and did it. I call it a silly little survey because I suspect no learning will occur from it, where by “learning” I mean “corrective action.” The cynic in me suspects these surveys are to prove whatever they want to prove — in this case, that the program is “working.” I, however, don’t believe the program is working at a core level.

A pair of questions that probed whether the NSF GRFP is both prestigious and desirable stood out. The NSF GRFP has always been a prestigious program, so perhaps it should be desirable on that ground alone for the recipient as well as the school. But in reality, its less than stellar financial terms make it less desirable compared to many other fellowships, especially industry sponsored ones. Let’s be frank, the primary purpose of any fellowship is to provide financial support to the student. The NSF GRFP is no different: “The program recognizes and supports outstanding graduate students in NSF-supported science, technology, engineering, and mathematics disciplines…” The weird thing about “support” is it’s either total or none. There shouldn’t be such doublespeak as “partial” support, when it would be better termed “aid.” This is especially true when you consider that a true fellowship is meant to support the student in the sense of allowing him or her to pursue a research agenda freely, without being encumbered by financial concerns. So support is not forthcoming when it turns out the money provided by this GRFP does not cover everything — tuition and stipend — and the NSF is unwilling to enforce its own stated terms on the schools to provide the remaining funds.

Over the 3 years during which a fellow is on tenure, the NSF gives $30K/year in “stipend” and $10.5K/year in “cost of education allowance,” also known as “tuition supplement.” The fellow doesn’t see a cent of this directly: it goes to the school. Graduate tuition at top institutions these days runs at $40K/year — outrageous, but that’s what it is. Then standard research and teaching assistantships are $25K/year. To the school, $65K/year is the standard price for a graduate student. NSF gives it $40.5K/year for free, so the price has been reduced to $24.5K/year. That’s a super deal for the school, but here’s the funny thing — some schools don’t see it that way.

To the school, the $10.5K/year in “cost of education allowance” given by NSF is just another source of general funds. Kaching! Now, being generous, the school passes on the $30K/year in “stipend” to the fellow instead of the $25K/year they’d normally pay, but then charges a “tuition shortfall” of $29.5K/year ($40K – $10.5K) to the fellow! What is the total value of the NSF fellowship now? $500/year is your answer.

Of course, the school, in a rare stroke of humanity, doesn’t expect you to actually live on $500/year, so instead of ponying up the remainder of the tuition like they are supposed to (per their tacit agreement with the NSF), they coerce the fellow to seek regular funding in a research or teaching assistantship and pay back all of the stipend from that. This quid pro quo service to the institution for purely financial reasons is about as blatant a violation of both the letter as well as the spirit of the fellowship terms as you can get.

Now I’ve brought up this ridiculous situation on several occasions when the NSF solicited commentary much like this survey and I’ve even called NSF, but NSF does not want anything to do with it, preferring to defer to the school’s coordinating officer — essentially letting the school’s own financial decisions override program policy. The student is caught in the middle. This is unconscionable bureaucratic buck-passing. For practical purposes, the true benefit of the fellowship, financial or otherwise, is not seen by the student, and this miserly arrangement does not befit the fellowship’s ostensible prestige.