Applying for U.S. grad fellowships (NSF GRFP, NDSEG)

As I was applying to U.S. Ph.D. programs, I also applied for the National Science Foundation Graduate Research Fellowship Program (NSF GRFP) and the National Defense Science and Engineering Graduate (NDSEG) Fellowship. I was rejected from the NSF GRFP, and put on the wait list, and then eventually rejected, for an NDSEG. In my first year of my Ph.D., I was no longer eligible for the NSF GRFP (because I had completed a master’s degree), but I was able to apply for the NDSEG again. In the intervening year, the NDSEG completely revamped their requirements (as they now seem to do quite frequently), so it was a very different process. This time around, I was awarded the NDSEG.


There are tons of great resources online with tips about applying for the NSF GRFP (like this site). That said, most of these resources have examples of winning essays from which people extrapolate general guidelines for how to put together a successful application. This can be very useful, but figuring out how to write a winning essay should also involve comparing winning applications to losing applications and figuring out how they differ. But there aren’t that many examples of losing essays online. “Fortunately,” I got pretty mediocre scores, so here are some tips I’ve come up with based on the negative feedback I received.

  1. Think carefully about the panel you submit to. This won’t apply to everyone, but if your discipline falls in between two categories (e.g. developmental psychology and cognitive psychology), think about which group of researchers is likely to find your proposal the clearest and most compelling. Ideally, you will write your application to appeal to everyone, but if there are any weaknesses, think about which group is likely to find them most glaring, and avoid that one. Decide what panel you are going to submit to ahead of time, and tailor the language in your application to the language of the field, which can vary even across very closely related subdisciplines. (But avoid jargon altogether!)

  2. Lack of clarity or sufficient detail is probably a bigger kiss of death than lack of scientific impact. I haven’t really seen people get criticized for their ideas not being exciting enough. But I have seen lots of people (including myself) get penalized for failing to mention clarifying details and/or control conditions that might be important for interpreting results. None of my reviewers wrote that my research topic not be interesting or innovative — they were just concerned that my methods wouldn’t adequately answer the question I proposed. My hunch is that the type of criticism I received is far more common than criticism along the lines of: “Your methods are really solid but the answer to your question is unimportant.”

  3. You don’t have to do the research you propose. If you’re torn between several ideas, write about the one you think you can describe most effectively in two pages, not the one you most want to pursue.

  4. This is obvious, but — get feedback from lots of people. The most helpful feedback will probably come from people who are in your discipline but who do not share your highly specific interests, since that likely reflects the composition of your reviewers. Is anything in your proposal unclear? Do they understand both what you aim to do and why it’s important?


Though this is the fellowship I am currently supported by, it remains a mystery to me. Here is how I think it works.

  1. A panel of researchers first reviews all the applications and scores them. This scoring is based on your transcript, GRE scores, research experience, prior awards, personal statement, research proposal, and recommendations. I’m not totally sure how the scoring works.

  2. The top scoring 40% of proposals then get sent to the various Department of Defense (DoD) agencies and they select the winners. I’m not sure the extent to which they consider the actual scores of the initial panels or whether everyone is starting fresh once they make it past the initial review round. I was named an alternate the first year I applied, and I am pretty sure that meant that I made it past the initial review, but then was not selected by a DoD agency. Based on this assumption, I decided to seriously ramp up references to relevant DoD initiatives in my research proposal the second time I applied.

    That said, on my second go around, I included specific references to research programs that were listed as interests of the Army and Navy. Ultimately, I was funded by the Air Force, which I did not mention at all in my application.

There are no guidelines on the website for how to structure the research proposal, and no examples of winning proposals online! How am I supposed to know how to write mine?

I applied in the first year that a 3-page research proposal was required, so I had “nothing” to base my application off of. However, if you are in grad school, or planning to go to grad school, you most likely work with a professor or mentor who has had experience applying for grants from lots of different organizations, all of which have their own idiosyncratic requirements, but also share commonalities. Ask them to help you come up with a general structure for your proposal.

Also, if you are applying for the NSF simultaneously, you can just add a few more details to your NSF proposal given the longer length limit and try to reframe some of your “broader impacts” sections to focus more on DoD relevance.

Here’s how I structured my research statement:

Section 1: Introduction

Introduction to the research question I aimed to address, very brief recap of previous studies that addressed similar questions and the gap I aimed to fill, real-world example of a scenario in which my research question is relevant.

Section 2: Relevance to the Department of Defense

How my research relates to the specific interests of the DoD. Like many people, I do not think of my research as being particularly relevant to defense-related activities, and describing how it could be used by the military makes me uneasy. In general, there is a way to spin research on most aspects of human behavior or cognition in a way that makes it relevant to the DoD’s aims, particularly if you spend a long time looking through their materials to see what types of research each branch is interested in. It is obviously up to you whether you feel comfortable “selling” your research in this way and accepting money from the DoD to carry out your work.

Sections 3 - 5: Aims 1 - 3

Hypotheses I aim to test, methodological details of experiments. I had 1 – 2 related experiments for each aim.

Section 6: Benefit to society

How will answers to my research questions help people in general?

Section 7: My relevant qualifications and institutional support

Very brief info about why I can carry out this work effectively and how I will be supported by my advisor and institution.

Section 8: References that I crammed in to take up as little space as possible

(I think they now let you put references on a separate page, so the cramming might be unnecessary.)

You can read my full research statement here.