Advice for Grant-Writers
I am a biological anthropologist by training and yet many of the questions on which I work can easily be approached from the perspective of social anthropology. Indeed, most anthropological demographers and anthropologists working on questions of infectious disease are social or cultural anthropologists. Since coming to Stanford, I have found that most of my students are more social than biological anthropologists. So it goes.
I have been on numerous scientific review panels, including the NSF Cultural Anthropology Doctoral Dissertation Improvement Grant (DDIG) committee. While committee work of any kind can be a hassle, there are definite up-sides to working on scientific review panels. Top among these is the incredible amount that you can learn about grantsmanship. There's nothing like looking at 20 grant applications -- and talking about 100 in committee -- to learn what works and what doesn't.
My goal here is to translate some of this experience into practical advice for students. Ostensibly, this is advice for students applying for DDIGs in cultural anthropology. However, much of it transcends the particulars of this panel and applies to grant-writing more generally. Some of the stylistic advice I have copied from my Guide to Research Papers.
Read the Guide to Grants . Pay particular attention to the Proposal Preparation Instructions. If you are applying for the Cultural Anthropology DDIG, there is a page of program-specific instructions as well. If your proposal is too long, or you use the wrong font, it will be sent back unreviewed. That is a pretty bad reason for not getting funded.
Your proposal should not be too long. NSF has very strict guidelines for formatting your proposal to which you must conform. Even so, don't push the limits of the letter of the law; follow the spirit. No one ever complained that a colloquium talk was too short (particularly if there are refreshments waiting after). Similarly, no reviewer is likely to complain if a coherent, well-edited proposal doesn't fill out the 10th page or brush up against the 1-inch margins on every page. The trick is to write enough to show that you have fluency in the field, have an interesting and scientifically important project that you can actually do in 12 months within the ten pages that you are allotted. This takes discipline and repeated editing. Get a friend to help you edit. Only include the information that is absolutely necessary.
On a related note, don't be overly ambitious. This is a dissertation proposal. Keep it simple and, above all, doable. You have the rest of your career to do everything else you want to do and getting a DDIG is a great start toward insuring that you have a career in which to do the rest of the work!
Particularly if you have an ambitious proposal, you must show some results from preliminary research. Many scientifically excellent proposals get dinged because the panel has no reason to think that the applicant will actually be able to carry out the research. Reassure the panel that you can get safely to and live safely in your field site, collect the types of data you plan to collect (in the numbers you say you need), and do the analysis you say you are going to do. Pre-test your survey instrument or experimental game. Oh, and pre-testing on a bunch of undergraduates at your home institution doesn't count...
Think about your audience. For the DDIG Cultural Anthropology panel, you are writing to the full spectrum of anthropologists. For other panels, you are equally likely to get a broad range of panelists. Write for a general audience. Avoid excessive jargon.Define your terms! Things that might be obvious to you may be utterly baffling to a specialist from another sub-field. Given my background, the terms with which I frequently have difficulty are those like "governmentality," "imaginaries," or "predicaments of modernity" (though I am equally befuddled by terms bandied about in evolutionary-psychology applications). I don't really know what those terms mean. They may encode very important concepts, but for them to communicate effectively, they need to be defined. I would not expect a non-specialist to necessarily know what I was talking about when I used terms like "eigenvalue", "centroid," or "betweenness," so I would define them.
Don't believe the folklore about the panel. The Cultural Anthropology DDIG panel recommends all sorts of proposals for funding. If your research does not fit the mold of natural science, don't try to make it sound gratuitously scientific. You'll probably just end up sounding scientistic and that's never a good thing. Similarly, if you are more of a natural scientist in your approach, don't try to sound like a literary critic. Describe your research in coherent, well-written, carefully-edited prose.
If you want real information on the panel, talk to the program officer. This is always a good idea before embarking on a time-consuming grant application. On a related note, do your homework. Read the abstracts online of funded proposals from previous years. But be careful. This is one way that the folklore about "what the panel is looking for" can get distorted and perpetuated. From the successful projects, you know only about the numerator. You know nothing about the denominator. Say there are only two proposals that were funded that resemble your research. How many proposals were submitted in that area? You don't know and there is no practical way to find out. NSF can not fund research in your area if there are no proposals!
Just because NSF funds a short course on some method doesn't mean that you need to use that method in order to get funded. Social network analysis and cultural consensus modeling are powerful tools. That doesn't mean that they should be used for every proposal.
If your proposal comes from the tradition of a cognate field (e.g., evolutionary psychology, science studies), make it very clear why a room full of anthropologists should give you funding. Come up with something better than "I couldn't think of any other panel." Talk to the program officer about appropriateness.
When you are asked for an NSF-style biosketch, don't provide an NIH-style biosketch. Nor should you write a discursive biographical essay. This is a specific instance of both reading the GPG and editing the proposal.
Be certain to explicitly describe both the intellectual merit and the broader significance of your proposal. In your project summary, set these out as separate sections.
It's the Theory of Mind, Stupid
This is so important that it is worth spelling out again. The advice in this section applies not only to DDIG applications but to any grant application, fellowship application, or research paper.
Each panelist has to read approximately 20 proposals in detail and many panelists will vote on your proposal having only read your project summary. This fact should condition everything you do in a proposal.
Make life easy on your reader. Don't make the font too small or the text too squished. Have logical subsections. Don't overdo the typography. Ensure that your summary is well-written and coherent. Define terms. Label tables and figures.
Clearly state your aims, hypotheses, and goals. Tell the panel why you are doing the research. This is never self-evident, although it may seem so to you.
Testable hypotheses are good. Don't overdo it. If you have fourteen hypotheses, no one will believe that you will actually be able to test them. And if you have this many and refer to them by number, you are sure to lose your reader.
Explain why the research is important. Will your research answer an scientifically significant question? To which debates does it contribute?
Sometimes a figure can help. It can also obfuscate. Avoid extremely complex flow charts. If you can come up with a simple figure that encapsulates your research, you have helped your cause greatly. A spaghetti diagram is just a waste of toner.
Yes, I'm sure you're very smart. Don't come across as a pompous jackass. Lots of very smart people get their proposals rejected. You don't want to run the risk of turning off your reader with your attitude.
Notes on Writing Style
- Do not be vague when you can be specific. Define your terms. If you suggest that "urban environments are bad for human health," what do you mean? What aspect of human health? What is "bad"? Increased mortality? Increased morbidity? When you write a scientific grant application, it is particularly important to define terms that have generally recognized folk meanings. Watch out particularly for words like "stress" and "fitness," which have both colloquial and technical meanings. Do not confuse these.
- Avoid extended quotes. Paraphrase (and cite appropriately) material that you are using.
- Avoid colloquialisms and trite neologisms. Do not use nouns as transitive verbs. Leave your soap box in the store room, your high horse in the stable, and squelch the sarcasm. Let the quality of your arguments, the coherence of your research design, and the elegance of your prose communicate whatever righteousness or criticism may be your just due. Your point will be made all the more powerfully.
- Don't fall into the trap of parochialism or ethnocentrism. Think about the generalizability of your statements. If you write, "people work to save for their retirement," does that really apply to people, broadly construed? What about arid-country hunter-gatherers? Horticulturalists in the Amazon basin? Nomadic pastoralists on the Tibetan Plateau? In statistical parlance, ask yourself the question, what is the universe of entities to which my arguments apply?
- That said, don't overly particularize your statements either. I am reasonably confident with the statement that people seek food when they are hungry and circumstances permit it. I don't feel the need to say something like, "As a Caucasian, middle-class man in my early forties speaking from this point in the early twenty-first century, I believe that hungry people seek food."
- Argue with logic and from first-principles, not by citation. Argument by citation is only slightly better than appeals to higher authority, which have absolutely no place in a scientific paper or grant application.
- That said, you need to demonstrate to your reader(s) that you have a command of the literature. Be thorough but be concise. With 20 researchers in the room evaluating your proposal, you have a lot of expertise. Don't try to snow anyone because you are likely to get caught.
- The passive voice is to be avoided. A proposal written with the author's voice removed does not sound more scientific, it just sounds old-fashioned and pompous. Some journals insist that all instances of "I" or "we" be excised, but this is becoming less common. That said, watch out for having your proposal sound solipsistic -- it should not read like a diary entry. Avoid, in particular, phrases like "I believe..." or "I think..." Clearly, you believe it if you are taking the trouble to write it.
- Organize your proposal with section headings and sub-headings if necessary. Such structure greatly facilitates readers' ability to follow the logical structure of a paper.
- That said, don't go crazy with headings and structure. In particular, keep fancy formatting to a minimum Bold face type is probably sufficient for most headings and sub-headings.
- Edit your proposal. You are trying to convince a room full of skeptical reviewers that you are a careful scholar with the tools to capacity to carry out the proposed research. Nothing turns a committee off faster than a sloppy proposal. Your goal should be to force readers to engage with your ideas, not to get distracted by sloppiness. Of course, this means that your ideas have to be good!
- Use page numbers.
- Avoid "sneer quotes." Yes, we get it. We're all terribly ironic. If a term is problematic, define it. Say what you mean and why it's problematic.
- Avoid excessive typographical acrobatics. Let your clear prose cue the reader to what the important point or the key hypothesis is rather than underlining and italicizing big blocks of text.
- Don't use as many exclamation points as I have in this document!
- Have a research design.
- Think hard about sampling. Yes, you need to have a sampling strategy. Even if your doing "pure" ethnographic research. How good will your ethnography be if you only talk to lames?
- More to come...