Right now a lot of us are waiting to hear if we won the NSF lottery. Someone recently on a panel who did not want to be identified shared some thoughts about what makes a good proposal. Some of these suggestions may seem obvious and yet many make these mistakes. The funding rate might be as low as 7%, so even if you make no mistakes, you may not get funded. But I like to think that a careful proposal, even not funded will strengthen your research and your writing.
To my confederate’s list I would add a couple of things, which I’ll just put here up front. First, be sure you read and cite the literature. The person you did not cite is likely to be a reviewer. Also this puts your research in context. Second, make it really clear what your big goals are and why your research is important.
Advice for all Proposals
1.Write very clearly. So many people are submitting proposals, which makes them ineligible for panels. This means your proposal is likely to be judged by someone distant from your immediate field. These people will have 10-12 proposals to review and they will probably be trying to do those reviews in just a few days, one after the other. In particular, do not use abbreviations except for things everyone really knows like DNA and RNA. You will just make the reviewers feel cross and confused if you try to teach them your secret code.
2. Make your proposal beautiful. Have white space, figures, and the legal fonts. Some older panelists might actually print your proposal and these are the very ones whose eyes are more likely to be failing. Make sure fonts and resolution on figures are readable – somewhere between 33 and 50% of the proposals my correspondent reviewed failed to do this. The figure must be readable if someone prints the proposal out and it also needs to be of good enough resolution that if people are reading it on screen and blow the proposal up to fill the whole screen, they can read your figure. Be sure that it does not turn into Monet’s rendition of your figure.
3. Remember you are fallible. Have a section on limitations, pitfalls and alternative hypotheses for each goal. You need to have shown that you have thought about these things and how you will deal with problems that may arise. This is really important. The strong proposals did this, the weak ones did not.
4. Show your preliminary data. Try not to leave the reviewers with any gaps in knowing you can do what you propose to do. So few proposals are funded, reviewers can grasp at details to deny you. This is an easy way to do so, no matter how amazing your ideas are. Yes of course this is frustrating, especially when your techniques are nothing special or very different from things you have done before.
5. Pay attention to methods. Tie them to cool ideas. Panelists often sink proposals not on the ideas, but on the methods. My confederate says this panel was better than another one served on, but sometimes you get people serving on a panel who hone in on the details of the methods, the sampling, the analysis and sink an awesome proposal because of it. It is really hard (essentially impossible) as a sympathetic panelist to rescue a proposal that has been sunk by details about the methods, because you have to ask someone to trust the PI/team know what they are doing, which of course is the thing that is under question by the panelist who is focused on the methods! Trying to rescue a proposal in this trap doesn’t go so well. Do not say that you will use an alternative that is harder and less likely to succeed than the approach you propose to use.
6. Report on prior funding. Report on BOTH the intellectual merit and broader impacts of the proposal and that you include the grant number as it asks you to do in the RFA. In general follow the instructions very carefully. Ideally, get a recent successful proposal as a model.
7. Make sure you include a project timeline, says my friend. I personally don’t set much store by timelines, but if they want it, better do it.
8. Think carefully about data management. Data management plans are looked at and depending on who reviews the proposal, may be evaluated. There are no official guidelines right now, but there may be. Right now, it really seemed that it was just a matter of the data management plan being there or not being there.
9. Have a good postdoc mentoring plan. Postdoc mentoring plans are more important, they were discussed with respect to some proposals and some critique may have been given on them for some proposals. Make sure you have one if you are proposing to hire a postdoc, and that it is thoughtful. Make sure you plan to pay your postdoc a salary that is consistent with the new federal guidelines.
10. Make your broader impacts strong. Without strong broader impacts, your proposal will probably not be funded. They don’t have to be insane, but they have to be strong. They also don’t have to be novel, if you have some good stuff established, you can keep going with that, maybe seeking to strengthen it in some way.
Extra advice for CAREER proposals
11. Read the instructions. Follow the instructions. CAREER proposals are not just another opportunity for junior investigators to submit a regular proposal. CAREER proposals are different. The panel my confederate was on reviewed a bunch and where in the past these have been some of the most outstanding proposals, this time, they were some of the worst – maybe the pre proposal system has contributed to this change.
12. Write an integrated education plan. CAREER proposals also need an INTEGRATED education plan – make sure you have one and one that is INTEGRATED
13. Write an integrated career plan. CAREER proposals also require a CAREER plan, but only one of the CAREER proposals my friend reviewed had one – again, read the instructions carefully.
14. Get a strong letter of support from the department chair. The letter of support from the chair of the department is really important and is evaluated and discussed by the panel. It is helpful if the chair indicates strong support by, for example, offering teaching relief should the proposal be funded. Chairs are busy people, so you might volunteer to write a draft for her, or at least a few paragraphs highlighting the importance of your proposal, scientifically, educationally, and broader impactly.
That is it! If this time didn’t work, check back for next year!
While NSF does not have data management plan rules, the biological sciences directorate has issued guidance for preparing data management plans: https://www.nsf.gov/bio/pubs/BIODMP_Guidance.pdf
Along with Joan’s advice about using the literature to put your research in context, it also lets you discuss the potential impact for the field (that the panel is attached to as well as others). I would also add that for full proposals there are written reviews from people directly in the area of research along with panelist reviews that may or may not be directly in the sub field of the discipline you work in, so in a sense you are writing for both audiences.
Michelle Elekonich, NSF/IOS
Thank you for this Joan! As one that submitted their first CAREER this summer, I was a little apprehensive/nervous about your point #13 the “integrated CAREER plan,” thinking that I must have missed something obvious. I swear I read and read and re-read the “PROGRAM SOLICITATION NSF 15-555” and never saw this section laid out – and I just went back and skimmed my saved PDF of this document again and could not find this point (while your points 14 and 15 were clearly itemized in this document). That said, I hope that I made clear my broader context within the document, but I worry that reviewers are looking for a section that may no longer be requested?
Not sure! This is from a confederate. I hope you did it perfectly!
Pingback: Friday links: NSF proposal tips, WHEEEEE!!!, and more | Dynamic Ecology
I really like that you point out that you should make sure and include a project timeline. This seems like a great idea when you are making a federal proposal. Hopefully, the people making these proposals do research and find the best companies to help them. http://atlas.team/services/Federal-Proposal-Development/