alexhunterlang

If you are wondering how to get started, I wrote up my advice. I have also collected quite a few examples you can all check out. Please let me know if you have more examples to share! 

I have compiled lots of example essays that hopefully everyone will find useful. I have also posted my thoughts here.

I have compiled lots of example essays that hopefully everyone will find useful. I have also posted my thoughts here.

Have you worked with any faculty on these projects? Or are there faculty that work on similar projects in your school?
I think at this point, you really need to find a faculty advisor that you can pitch these ideas to. They will have a sense of what projects actually gets funded in your research area and should be able to narrow the list of 4 down to 1 or 2.
I personally know very minimal about your field, but in general, I think funding agencies hate the words "solving hard problems" that "don't really have broad impacts outside their topic". You need to put together a research proposal that is broken out into steps, and the funders want a least a few of the steps to be reasonable to accomplish, while the last step might be a stretch. So without knowing anything specific, I would say those 3 tough ideas need some work. But this is where a faculty advisor in your field could help you shape those 3 ideas into manageable steps that would be acceptable to NSF.

Good luck everyone on the NSF GRFP applications! I have a few miscellaneous thoughts...
1. I want to make sure that first year grad students didn't miss this...
I don't envy anyone having to make the choice of whether to apply your first or second year. I hope all undergrads reading this use this as extra motivation to apply now!
2. While the gradcafe is of course awesome, also take advantage of the experienced resource list.
3. I just updated my collection of examples of successful applications to everything I know about, and I also updated my advice to the newest solicitation.

I also can't find a specific link, but I recall seeing multiple references that the percentage of winners is very closely proportional to the number of applicants for a given category. So since last year, 2000 fellowships were awarded out of around 16000 applicants, the success rate was about 12% per category. The number of applicants is definitely going up every year, so if I had to guess, there will be 17-18,000 applicants, so a success rate closer to 11% (unless funding increases).

Great tip! I never thought of that. So some general advice for finding useful examples:
Google I've put together a table with all the examples I found. I also have given my general advice on applying. @coffee_grad's email tip (past winners here) Local resources at your college. Most schools have some writing / application advisors. Or you can start with the experienced person list.  Good luck everyone

Great tip! I never thought of that. So some general advice for finding useful examples:
Google I've put together a table with all the examples I found. I also have given my general advice on applying. @coffee_grad's email tip (past winners here) Local resources at your college. Most schools have some writing / application advisors. Or you can start with the experienced person list.  Good luck everyone

So first, maybe I should describe more about how I did my application, and then I'll answer specific questions.

I applied to the NSF GRFP in November 2009, the same time I was applying to graduate programs. I was doing research as an undergrad, so I worked with my adviser at the time on my research proposal. The research proposal was an extension of what I was doing at the time. So it was original, but it was a logical extension of our work, so it wasn't too original. Below are the intellectual merit criteria and you can see that they focus on new research that is doable.