What’s up with the new NSF GRFP priority areas?


The new solicitation for the NSF Graduate Research Fellowship Program arrived last night with NSF’s daily digest bulletin. There were eight items they brought to our attention as changes from last year, but when I was going through it late this morning, the soundtrack screeched to a halt:

4. Although NSF will continue to fund outstanding Graduate Research Fellowships in all areas of science and engineering supported by NSF, in FY2021, GRFP will emphasize three high priority research areas in alignment with NSF goals. These areas are Artificial Intelligence, Quantum Information Science, and Computationally Intensive Research. Applications are encouraged in all disciplines supported by NSF that incorporate these high priority research areas.


Ummm, what?

I am simply asking, what the heck is going on? I’ve got a lot of questions.

Why? What is the motivation for this change? What is it supposed to accomplish?

Have the folks running the program looked at the distribution of prior applications in this realm and thought about how this might impact the diversity, equity, institutional distribution, et cetera, of the applicant pool and funded applications? I know this is a priority for the folks running the program, but I have no idea whether they’ve done the assessment to forecast the impact on the demographics of the funded students.

How many of the funded applications will have to come under the umbrella of Artificial Quantum Computation Stuff?

Could it have been possible to give more heads up to the folks developing their applications? It’s unlike NSF to rush this kind of thing, so I’m confused.

Is there going to be specific guidance for applicants and the people and organizations supporting applicants, with respect to this change in program emphasis?

Meanwhile, if you’re working on your own GRFP application, or you’re mentoring someone who is, I do have a few thoughts that might put this in context. First, we don’t know how much of this is window dressing, and how much of this really will drive funding decisions for the bulk of applications. Second, nobody ever expects you to actually do the project that you propose, and it’s 100% okay if you do a different project once you’re funded. Third, nowadays, it’s a recurring theme in ecology/evolution/behavior to incorporate some genomic work. And nowadays, nearly all genomic work is computational. So there’s that. Fourth, these proposals don’t fund supplies for the research, and genomics and serious computational work can get pricey. But, hey, it’s not like you’re actually expected to do the proposed project.

I hope we’ll hear more from this from NSF, because dropping a few bombshell sentences about this in the new solicitation without additional context is not helpful to anybody.

9 thoughts on “What’s up with the new NSF GRFP priority areas?

  1. The grants don’t fund supplies but don’t they give access to computational resources? I thought they did at one point.

  2. NSF GRFP used to say they fund 2000 awards, then they went to 1500, and this RFP says 1600. Doesn’t matter, look at awards given each year they still have funded 2000 consistently. This could be a change or it could be motivated by Congressional pressures. This RFP gets 13000 to 16000 applications each year and reviewing them is a feat. If putting a statement like this deters applicants, then it solves problems: reduces number of applications to peer review and increases number awarded in priority areas.

    • While a large applicant pool is a challenge for NSF to review the applications, the smaller the pool, the more bias and inequity in its composition. I’ve been working my ass off trying to increase the pool by getting students from populations that rarely apply (low-income students from regional public universities), and measures that deter applications are likely to be a greater deterrence to this population.

    • I think your argument is exactly why this language is dangerous to include arbitrarily. It will discourage people from applying but it should not be a guess who. URMs, socioeconomically-disadvantaged, and first generation students will read this sentence and realize the goal is about the kind of science and not the future researchers goals. This should be an obviously terrible idea because many students from these backgrounds do not have access to these higher level topics until grad school because their paths are so different from pedigreed academics.

  3. One more point: NSF GRFP will return proposals without review on topics that seem too medical or health care related The same topic could be funded by NSF if a faculty PI submitted it. It’s returned without review to the senior undergraduate or early stage graduate student. I would prefer NSF to be forthright about its intentions. For the past couple of years, some students proposing basic science research on medically related topics have worked hard on applications that are returned without review. This has a significant adverse impact on early career scientists and counteracts efforts to retain students in STEM disciplines..

  4. Well many research groups have already been doing everything they can to put AI to work on their respective fields, so now NSF GRFP applicants will need to find some way to incorporate this technology into their proposal. It’s not a bad thing in my opinion, this tool will be increasingly important to research in the next ten years, and those who resist will likely be left in the dust.

  5. I don’t have any informed insights about this, but isn’t this just politically motivated buzzword-dropping, like “big data” was several years ago? My guess is that one of the reviewer rankings will be how well a proposal addresses the “priority areas.” In principle this shouldn’t make existing inequities worse, since it’s basically a directive to write your proposal in a way that name-drops this particular set of buzzwords.

  6. The language in the solicitation is reminiscent of this bill introduced in Congress:


    “The stakes couldn’t be higher, according to Schumer and Young. “The country that wins the race in key technologies—such as artificial intelligence, quantum computing, advanced communications, and advanced manufacturing—will be the superpower of the future,” they write.”

    What are the chances the changes to the GRFP reflect broader shifts, current or expected, at NSF?

  7. At the Botany 2020 conference there was a webinar by NSF program officers and obviously this topic came up. The presenters weren’t able to give too much information, only pointed us to the general statement that has already been released that “GRFP applicans will be and always have been selected based on their individual merit” but that its a move to “align with a coordinated federal strategy”


    The presenters said that we don’t know FY 20-21 funding levels, so its unclear how many GRFPs will be available this cycle, but stressed that this change does not mean that there won’t be GRFPs in other areas than the ones listed.

    DEB Program Officer Matthew Harrison said that as far as he is concerned there will be no change in how proposals will be evaluated, and that NSF is interested in maintaining a balanced portfolio of fellows, including a balance of subject matters, demographics, geography, and broadening participation in science.

Leave a Reply