Advice for the NSF GRFP?

October 6, 2006 11:32 AM

Any advice on the National Science Foundation Graduate Research Fellowship Program application? I'm applying in Social Science.
posted by k8t to Education (12 answers total)

Be as specific as you possibly can. This especially applies to your research proposal. They want someone who has a definite plan, support from their institution, and has already done a good bit of the preparation to carry out this research. So make sure you craft really clear, coherent, and well-researched statement about the project you want to work on. (You're free to do something different once you take the fellowship, but they're looking for proof that you're prepared to do top-quality research.) You want to be extra careful to address the "Review Criteria of Intellectual Merit and Broader Impacts," they want a serious proposal that will have some impact beyond your immediate interest.

You should also be as specific as possible when you write about previous research experience. Hopefully you've written an undergraduate thesis and have some papers published or have given a talk at conference or something. If you have, make sure that your advisor from that project writes one of your letters of reference, and make sure to mention these in your application. The personal statement needs to be something that will grab the reviewers' attention, but remember that this is not a college application. Unbridled enthusiasm without a clear plan is not a good thing.

Aside from things that you likely can't control at this point -- GRE scores, college transcripts, reference letter--the specificity thing in the proposal matters most. My advisor served as a reviewer for these and insists that top scores went to those with really thorough proposals.

The NSF has mandates to distribute these awards geographically, and this is done by place of birth. If you were born somewhere like Wyoming you have a better chance of getting a fellowship than if you were born in the Northeast or California.
posted by dseaton at 12:27 PM on October 6, 2006

Have publications. At least in the hard sciences, all the people I know who have received a fellowship have had published journal articles.
posted by betterton at 1:35 PM on October 6, 2006

The biggest thing you can do is fulfill their "Broader Impacts" criterion. In other words, be a woman, minority, or person with a disability (a cinch with more than one). Not that I'm bitter or anything. It most likely also helps if you can write about volunteer work or whatever that also fits into Broader Impacts. The NSF has made this way more significant over the last 5 years or so.

Applications with strong Intellectual Merit are a dime-a-dozen. Of course this means you'll need the GREs, transcripts, undergrad research, etc. But focus on addressing the Broader Impacts well.
posted by Durin's Bane at 1:50 PM on October 6, 2006

Since you are applying in social sciences, make extra sure to highlight the 'science' part of your research proposal.

I concur with "be as specific as possible", except make sure you still know what you're talking about. Better to be a little bit too general than to be blatantly wrong. (Of course, vetting things with a prof or other older mentor is always a good idea).

You do have a little control over your reference letters- make sure to pick wisely and get as many as NSF will take.

Next, most importantly, if you don't get it this year- try, try again. You can apply up through your second year in grad school- I did, and am now one very happy student (well, that way anyways). (Oh, and I didn't have any publications then either, so they *aren't* a requirement).

Good luck!
posted by nat at 1:52 PM on October 6, 2006

I wanted to clarify a few points Durin's Bane tried to make.

First, Broader Impacts is separate from the Diversity issue.

My understanding is that the broader impacts criterion asks how your research will help society. In other words, why should the government pay you to do your research?

The diversity issue is real, but I think much more important than *being* a woman/minority/disabled person is the volunteer work. If you haven't done anything formal, think of informal things- have you TAed a class and helped out a student with a disability? Have you mentored minority students? (and so on).

Out of curiosity, dseaton, do you have any references regarding NSF's geographic mandate? I've heard the same thing, but never seen any data nor official policy about it.
posted by nat at 2:03 PM on October 6, 2006

Unfortunately, Durin's Bane is correct about the broader impacts. The NSF has recently taken a turn from affirmative action-type leveling of the playing field (which I agree with) to full-on quotas (which I don't).

This is one of the underlying facets of Broader Impacts, and it does include geographic 'minorities'. A friend of mine just spent several months trying to set up a science conference with NSF funding, and the NSF functionary who she had to deal with (the program manager) demanded that the speaker list not only include a specific percentage of woman and minorities, each panel had to have a representative from every geographic area of the country. gShe ended up having to beg some people who really weren't doing research in that area to come just to fulfill all of the various quota requirements, which was a waste of the NSF’s money and the conferees time.

If you do suffer from being a white male on one of the coasts, nat is right, you need to have a good plan in place which will assure them that you will spend some time and effort communicating your field to underrepresented minority students at lower educational levels. If you're at an elite coastal institution (with almost nothing but rich white students), this means that you have to get involved in summer research experiences for minority college/high school students or lectures/demonstrations at minority-rich local educational institutions. (You should anyway, btw. These are Good Things.)

Also, look into the NIH training grants. If your Social Science project can be realistically presented as having to do with health behaviors and/or access to healthcare, you have a good shot. And their criteria are entirely merit based.

Good luck!
posted by overhauser at 2:47 PM on October 6, 2006

Response by poster: I love you all. :)
posted by k8t at 3:20 PM on October 6, 2006

It is true that demonstrating "Broader Impacts" is important in NSF proposals.

However, the claim that if you are a woman or a minority you have better chances at being funded is not true. Women and minorities receive grants from the NSF at lower rates than white men do; in other words, being a white man helps your chances.
posted by betterton at 3:50 PM on October 6, 2006

Also, note that the funding process is intrinsically random, for many reasons.

For example, the panel that reviews the proposals is made up of a finite number of people. If there is a professor on the panel with expertise in X, then several fellowships in X are likely to be funded that year. Some proposals might not be funded because the subject area doesn't fit well with the panel's interests.

In my experience, people are willing to say, "I didn't get funded because I'm a man," which completely ignores all the other (often random) factors that affect the selection process. Therefore, don't let whether or not you get funded go to your head.
posted by betterton at 3:55 PM on October 6, 2006

However, the claim that if you are a woman or a minority you have better chances at being funded is not true. Women and minorities receive grants from the NSF at lower rates than white men do; in other words, being a white man helps your chances.


In absolute numbers, sure of course.
In percentage of applicants, I'd love to see statistics.
In the case of everything else being equal, you've got to be joking.
posted by Durin's Bane at 4:30 PM on October 6, 2006

However, the claim that if you are a woman or a minority you have better chances at being funded is not true. Women and minorities receive grants from the NSF at lower rates than white men do; in other words, being a white man helps your chances.

Not true. The NSF funded this study, which found that (without making any judgement on the merits of proposals) women recieved the same percentage of grants and the same final grant sizes on average as men from the NSF and the USDA.

The interesting thing is that the NIH, which does not recognize minority status as a form of merit, does fund a lower percentage of grants with female PIs, and on average gives them a lower amount of money. Given that, for the most part, the same groups of scientists form the peer reviewer pool for both agencies (and which are very carefully composed of a mixture of genders and races), what does that tell you? I suspect it's that the NSF's playing field is pretty tilted. It's hard to say for sure.

One reason that it's hard to tell is that, while the NIH is upfront about giving out the actual numerical scores from peer review panels to PI's, the NSF is pretty evasive about it. The study even complained about that. If you were the suspicious type, you might think that is because they don't want a chorus of 'She got a worse score than me, and she got funded anyway' complaints.

I do agree that the people who whinge that they didn't get a grant because they were male are missing the point. Many white males do get grants, and the whingers lost in the competition with other men for that pot of grant money. They weren't competing with women, and women in the NSF are also competing only with each other.

One thing I'd really like to see is a double-blind form of grant review, like when people try out for the symphony from behind a curtain. That would make a lot of people of both genders more comfortable with the process.

I need to go back to work on my grant now :)
posted by overhauser at 6:54 PM on October 6, 2006

Betterton is right about the randomness of the grant review process, btw. Just cast your net wide. Look into a lot of agencies and foundations, there are probably several to which you can apply. And if you don't get any on the first try, either get the reviewer's comments or call the program manager and find out how you can make it better. Then reapply!
posted by overhauser at 6:59 PM on October 6, 2006