Nsf Grfp Application Advice

Applying to the NSF GRFP

This is a collection of notes I took when preparing to apply for the NSF GRFP. Against all odds and to my enormous surprise, I won it! So I’m sharing these notes and resources here in the hopes that they will be helpful to someone else.

Links

General

• Most people don’t get the NSF, so make sure your time spent here is worth it. Have a higher goal: identify/define/articulate your research agenda. (I was certain I wouldn’t win, so I took this as an opportunity to force myself to define my research agenda as a second year).
• Tell the story - capture attention early on, see this as an opportunity for you to weave the story of your life together and see the thread that surely connects everything even if it’s not obvious.
• Use formatting to draw attention to important bits, and save whitespace. Bolding, subheadings, use structure to help people navigate. Make sure bold segments are coherent if that’s all you read. Makes a difference to the clarity of research proposal.
• Use the smallest possible citation format.
• It’s OK to propose research that you don’t end up doing. You’re not locked in to this proposal. They fund YOU, not the project.
• My app is definitely a bit of a mess, especially my research statement. In other words: you don’t have to be perfect.
• Consider your academic self, consider how the reviewers will understand who you are.
• Think about how your goals fit w/ NSF’s vision.
• I have a weird background, and it’s almost all in the app. See any deviation from the normal path as your superpowers.
• Apps are reviewed at the stage at which they are applying, so apply early. 2nd years are held to a higher standard. Reviewers compare apples with apples.
• Expectations for 2nd years? Someone who has had no graduate education will demonstrate competence and mastery of material. 2nd years: that’s not enough, need much more extensive understanding of literature, more sophisticated approach to research questions and design.
• Grades: not that important. Need good mastery of subject, but doesn’t correlate.

Process

• Read the prompt many many times and make sure you’re answering the questions!
• I had at least 11 people read my apps. Each looked at it at least 2 times, many 3 or more.
• Get an editing pipeline: I would send DraftA to person 1, get feedback, then edit to get DraftB and send to person 2, get feedback and edit to get DraftC which got sent to person 3, eventually creating DraftD which got edited and sent to person 1 again.
• Prioritize editing feedback by editing style: my Mom tends to focus on wordsmithing and we fundamentally disagree about commas, so I sent her later drafts that needed polishing. My friend Daniel absolutely rips things to shreds, so I sent him earlier drafts to get the overall shape right as early as possible.
• Get editors with varying backgrounds and levels of expertise. You are not guaranteed that the people reading your app will be experts in your subfield or area, so make sure things are understandable to an outsider. This also strengthens your claims about Broader Impact AND Intellectual Merit: if you can explain something complex such that it’s easy to understand, you are showing your skills.
• I wrote at least 15 full drafts, 8 of which were complete re-writes (I threw everything out and started over from scratch).
• For 3 weeks, I literally did no classwork, no research, only wrote my application.

Story

• Use verbs! “Measurements will be carried out” –> “We will measure this”
• It’s not hyperbole if it’s true.
• Get passion in there, reviewers look for that.
• Keep re-reading app info. Impression of what’s required may change with time.
• My first sentence: “I don’t recall feeling nervous about my decision to drop out of college and go to clown school.” My advisor really wanted me to take this out because it starts off w/ something “negative”. I kept it in because I kept getting feedback that it blew people away when they read it, and they wanted to keep reading. Plus, I knew I had enough “regular” accomplishments (publications, graduated from a good school, know how to write w/ language of academia, etc) to assuage any concerns that I’m just a kooky nut.

Research Proposal

• Part of what you’re showing off is your ability to scope research. The scope in this case is 3 years. Make it seem like it’s feasible in this time-frame. Longer vision can go into Broader Impacts.
• Reviewers look at proposal as evidence that you have research ability, can think clearly, can communiate clearly, shows your potential. Choose a topic that allows you to demonstrate that.
• Sub-goals of proposal: keep reviewer awake, be clear, be consistent, etc.
• Research Statement is about future, can mention current projects as sign that you’re already capable, and have access to resources (people, tools, skills, methods, etc.)
• Tell them what you want to study, why it’s important. Reviewers may know nothing about this topic. Tell people what’s known so far, in a way that makes your reviewers feel intelligent. They’ll like you and rate you higher.
• Don’t change your mind thru the proposal - write it down clearly and define it. Be consistent. Review what you’ve said you’re doing, to make sure it is consistent.
• At most 3 sentences of background before you get to what you’re doing in your research proposal.
• Include publications/references as relevant.
• Figures in proposal? Doyle: yes, do it. Flowsheet for methodology, pic of complex equipment. Not more than one. But they take up a lot of space, don’t use one unless it’s super worth it.
• “Write it like an academic” –> Don’t feel you have to write like research papers are written, because they’re often horrible. People who can write well feel that when they do tech writing they must adopt this new boring style. Don’t do that, keep reviewers awake and use an active voice.
• Technical communication is hard. Don’t underestimate this. Probably apocryphal tale of dean of eng here at Berkeley, speaking w alumni and donors about how well-qualified undergraduates are: Not only did eng. undergrads have better SAT scores in Math and other areas, but they came in w better qualifications in writing. One donor raised their hand and said “so what is it you do to them because by the time they get to me they can’t write at all.”
• NSF requests description of the resources, how do you do that as an undergrad when you don’t know what school you’ll need? Don’t get into crazy specifics, most universities will have access/can arrange visits.
• SHOW how the resources you’ll have access to will help you be successful by referencing already-completed successful projects.
• Share how you’ll assess success: details here are important.

Personal Statement

• Outline the trajectory as person/student. Answer the question: what led you here?
• NSF recognizes that the playing field is not equal. Unusual background, hardship, etc. Don’t need to sound whiny. Matter-of-fact way, can talk about what you learned by overcoming it. People have a tendency to be humble/self-deprecating. This is not the time to be humble. Think of how someone else would describe you. (Could write a draft in the 3rd person).
• Go back “as far as you need to do to project a compelling story.” In general, focus on what you’ve been doing in grad school, but go back and revisit. 3 good places to start story: when you got into STEM (quickly transition), when you got into research, crazy experience that leads to current place.
• Bring up volunteering, etc. even if it’s not directly related to your research: they fund people not projects.
• Your background is crucially important. You are reviewed holistically. Include whatever you’ve done that reflects your individuality and gets their attention. They’re looking for your capabilities as a researcher. A few showstoppers though: GPA needs to be high. If your GPA may raise doubts, you need to address it in personal statement (definitely still apply). Don’t be whiny. Overall GPA is less important than the ones related to what you want to do.
• Example of a compelling opening paragraph: “I’ve always had a strong affinity for the performing arts, creative literature, and power tools. There’s something blissful about building something by hand: a short story, a song, or a birdhouse.” Beautifully written, plays with expectations. Signals that they could make connections beyond the obvious ways things are related. Tie this to your research ability. Demonstrate how the rich background that is you sets you up to make unique contributions in your field.
• Convey excitement above all. Use the first person. Don’t be passive, be active.
• If you have research experience in a field that’s not completely relevant, stress what you learned. Should be in personal essay. Research proposal is 100% for what you’re going to do.
• There are examples of BI on the NSF GRFP site, so you can choose one to help flesh out the potential impacts of your idea if you’re lost.

Letters of Recommendation

• Get letters from someone who knows you well.
• Absolutely get 3 letters, even if only 2 are “required”.
• One ref from someone who knows your ability to organize, your passion, teaching, etc. The rest should be from academic experts.
• Industry rec letters? Manager who observed you in research work might fit, esp if you worked in industry (it would be suspicious if you don’t have one). They look for academic, but one is important.
• If you’re in a program where your first year is really all coursework and you haven’t had any contact w a prof who could write you a compelling letter, wait until your 2nd year. Or, you can address the fact that your profs don’t know you yet, and that’s why you’ve chosen certain refs. Your goal is to reassure your reviewers that you haven’t come to Berkeley and fallen flat on your face (if nothing is said, they might be concerned about that).
• Should you get a letter from your undergrad year? Yes. Go as far back as you need to tell your story.
• Letters from postdoc weighed differently from PI? “One likes balance”, postdoc saying you’re the brightest student they’ve ever worked with carries less weight than prof emeritus. Diversity of opinions.
• Someone who can talk about your curiosity, and your passion for learning new things, and your devotion to science communication, etc. This speaks to your underlying passion/desire for learning.
• A variety of different people. People who write your letters should know you well. More important than status of writer.
• Letter writers want you to win this. They’ll put time and effort into helping you win this.
• Letter writers are scouring mind for nice things they can say. Exam answers, class behavior, etc can all contribute to painting a picture of you.
• Get people engaged in your research proposal, so they can write substantively about your expertise and skills.
• People who know you in different context to provide a broad spectrum so reviewers feel they know you.
• Letters: broad picture of who you are and how you’re going to be able to do research. Refs particularly helpful.

My letters of recommendation (all asked for drafts of my personal and research statements)

• My PI (I knew him as an undergrad, and was a 2nd year, so he was very familiar with my work)
• A prof from a class where I did very well in (knew me the least well of everyone, but had specific criteria for who she would write letters for that I fulfilled, and she is head of the 2nd field of my interdisciplinary work).
• The executive director of a group I had worked w/ on broader impact stuff from my previous job. (chosen because I knew her well, we had worked together on a broader impacts project, we were friends kind of, and she could speak to me in a different context related to my broader impact project).

General

• Proposals are ranked separately depending on your year/level of experience, so apply as early as you can.
• Reviewers read between the lines of what is not stated. Get people to read it and point out what you haven’t said.
• For undergrads, recommend against specificity of what you’ll learn with who. Reviewers aren’t experts anyway, they’re looking for well thought-out statement and familiarity. Can talk about techniques you might need.

Logistics

• Grad students can only apply once in their graduate career, unless you had a break of at least 2 years.
• You can apply in the first semester of your second year.
• May not hold two federal fellowships concurrently, but consecutively is ok.
• Major field of study determines pool of reviewers. Consider who you want it to be reviewed by. Review panel is not likely to have specific expertise in what you’re doing. There’s also an “other” category if you don’t fit.
• Cannot change the primary field of study in your first year on NSF.
• Proposed research study is not binding. Your commitment is to pursue A course of study in this field, not THIS course of study.
• NSF reviewers are faculty members.

Reviewers perspective

What reviewers think about:

• What proposer wants to do
• Why they want to do it
• How they plan to do it
• How they will know if they succeed
• What benefits accrue if project succeeds.

Student presenter advice

Student 2:

• Took a year off
• OK to propose research that you don’t end up doing. (Physics -> BioE)
• Writing essays was easier if working w/ older grad student directly to bounce ideas off of.
• Research Statement: Wrote about science behind WHY you would want stable lasers.

Student 3:

• Put 40-50 hours on it.
• Use bold to call out a few particular phrases. Impact, problem you’re proposing to solve (one-line summary).
• Write refs in smallest possible way, half a line / ref.
• Have people outside your field read/edit.
• Read a bunch of winning ones. Find the formula.
• Write a complete picture - proposal should be complete.
• Rec letters: diversify

Feedback

• Feedback from reviewers: I was not specific enough (this is a known clear weakness of mine - my editors also caught it).