Guidelines for Writing Grant Proposals

Ann M. Peters, University of Hawaii, and
Lise Menn, University of Colorado


Our purpose in this essay is to help linguists, especially younger scholars, produce higher quality (and therefore more fundable) proposals for grants from NSF. In this era of tight job markets many linguists are feeling pressure from their institutions to get federal grants to support their research. But many of these same scholars, especially the younger ones and those in institutions that do not have strong research traditions, have not had the opportunity to see what a good fundable proposal for doing linguistics research looks like, nor do they have successful grant-getters available for consultation just at the time when they may be struggling with proposal writing.


Consider which agency/agencies (e.g. NSF, NIH, NEH, NIE...) are most appropriate for your topic. It is perfectly acceptable to submit proposals concurrently to more than one agency, as long as you inform each one about the others with which you are working.

Ask the agency/agencies to send you their standard application instructions, and the latest list of grants awarded in your research area. This will give you an idea of topics, amounts and durations. Contact the Program Director about the feasibility of your idea, and/or with a draft of your proposal.


We have organized the rules of thumb which we have gathered into five "themes": (1) the basic idea being presented, (2) methodology, (3) relevance to the field, (4) competence of the PI(s), (5) editorial considerations. This division is for convenience of presentation of the guidelines; keep in mind that all five of these themes should be thought of as facets that are relevant throughout your proposal, rather than as applying to separate sections.

    It has been said before, but it can't be said too often: there is no substitute for a good idea. Give your idea the benefit of a clear presentation with an overview in the introduction, details in the body, and a summary in the conclusion. Bear in mind that one of the criteria that proposal reviewers are specifically asked to address is the following: "Intrinsic Merit of the Research. Likelihood that the research will lead to new discoveries or fundamental advances within its field of science..., or have substantial impact on progress in that field..."

    Do be careful not to try to propose to do too much at once, such as testing all relevant theories on your excellent data base. If you have several studies in mind, use coherence as a criterion: it may be better to submit two proposals, especially if the separate studies are not particularly closely related.

    Coherence counts! Make sure the whole thing ties together. While you should have an organizing framework that unites the parts of your proposal, it does not absolutely have to be a theory, and a tacked-on theory is worse than none. On the other hand, theoretical relevance IS a major positive factor in the evaluation of a proposal. If you are proposing to do several studies, consider whether they really belong together. Leave out those cute little extra studies that don't relate to the proposal's main thrust. (You can always manage to squeeze some in if you get funded.)
    Method is important; often as important as the basic idea! Descriptive, sociolinguistic, historical, and some theoretical projects really depend on method just as much as experimental studies do, but often fail to present adequate information about how the work will be carried out. Make sure you are explicit; what you intend to do is the heart of your proposal, after all. What are the data, how will they be obtained, and what will be done with them in analysis? Show how the data you expect to gather will help address your issue. Never just say that your "data will be analyzed" -- how will they be analyzed? What will you be looking for and in what ways? Consider alternative methods of obtaining data, analyzing, processing, etc. Show you are aware of existing approaches, and that you have anticipated and given some thought to addressing at least the current standard problems associated with the methods you have chosen.

    Justify your budget with some care. Explain why you need the equipment you are asking for, e.g. in terms of the productivity it will offer your project. You should also explain why you do not already have access to such equipment. Other details that need to be thought through include the use of participants. What role will the various personnel play? You don't need to get silly about details (the research assistant will sharpen pencils), but show that you have thought about it and that all those salaries are needed. For example, it helps to indicate how long it takes to run a subject or code and enter an interview.

    If you plan to use powerful statistics, pay a professional to help with the design, and make sure s/he fully understands your goals and the constraints you are working under.

    If you are claiming that the data in studies in your area need "ecological validity", make sure your procedures aren't full of peculiar stimuli and situations. Try your ideas of "natural" out on someone who knows the kind of people you'll be working with. If a special population is needed for your study (e.g. second language learners, language delayed children, emigres), present the criteria you will use to define this population, and consider carefully any potential problems in finding out if a subject actually belongs to this population (can you depend on self-report? can you obtain needed records? etc.) Be sure to include the appropriate forms for the use of human subjects if they are required for your type of research. It is good to supply assurances of cooperation from other institutions and various categories of potential collaborators. If you will be doing fieldwork, show that you have the concurrence of the communities in which you plan to work, and that you are aware of potential cultural problems.
    Your organizing framework should help your readers see the relevance of your work. As noted above, this framework does not have to be a theory, but it should provide a context for relevance. If, however, you are using an elaborate notational or theoretical framework, do spend a little time explaining why you have chosen that framework instead of another.

    Your literature review should show how you think your proposed work will fit into what has already been done. If you say that previous work has been inadequate, indicate what is wrong with it and how your approach will remedy the defects. (Be sure that you have included the major relevant sources in the literature, but see also the notes on bibliography in section (4).) Then show how the data you expect to gather will help to address the issue(s) you have raised. Your study will seem more versatile if you can show that the data (or at least some of the data) from your studies will be interpretable in models other than your own. You should also briefly show how your proposed work could lead to future research possibilities -- another aspect of relevance.
    One of the most important criteria for evaluating proposals is the perceived competence of the PI(s) to carry out the project in question. Reviewers are specifically asked to comment on "Research Performance Competence", including "capability of the investigator, technical soundness of the proposal, and the adequacy of institutional resources available". Comments on "the proposer's recent research performance" are also solicited. Therefore it is crucial that you convince your readers that you do indeed have the requisite capabilities. You don't have to oversell yourself, but you do have to be quietly convincing. It is well to bear in mind that an impression of competence is gleaned from little things, such as a solid common-sensical approach, general coherence, and attention to detail, as well as from lists of accomplishments or publications.

    At the outset, when you are presenting your basic idea, you also need to explain why the problem is worth attacking, and show that you have the background and appropriate techniques to attack it. If you are not yet well known for work in a particular area, make it clear that your training and previous work qualify you to carry out the proposed work (but don't puff yourself -- let your factual description of your relevant previous work speak for you).

    One way to show common sense is to think your project through carefully and in detail -- enough to know what is really possible to do. The more details you have thought about before you start to write, the more convincing you will be when you do write. You won't want to include all the details in your proposal, but the fact that you have done your homework will show, nevertheless; it will help you show that you know what you are doing and how to do it.

    Don't bite off more than you can chew. If you are going to need cooperation or help from other people, say so, and provide assurances from them that they will be available for you.

    Don't write promissory notes: reviewers can tell when you haven't done the necessary groundwork. To be convincing, pilot! pilot! pilot! And summarize the pilot findings in the proposal. There is no better way to convince reviewers of the claims that what you want to do is indeed possible and that you know how to do it. The more innovative your approach, the more important this is.
    First of all, remember that you are writing this proposal to be read by colleagues who will be delighted to find themselves reading a good one. Remember also that these colleagues are, like you, well-intentioned but overworked. The 15-page text limit (bibliography and CV are not counted as part of the 15) is there for your benefit as well as theirs. And it's the spirit of the 15-page limit that counts! Don't use small type and tiny margins -- they're against the rules, and they only add to a reader's headache.

    Watch the proportion of space you allot to literature and general statements; don't spend 12 pages on the history of your field and the relevance of your proposal and have only 3 left to explain what you are going to do.

    Make sure your literature review and bibliography are complete: that you are not missing major sources of information. Make sure it is up to date, especially if you are recycling it from papers you have written earlier. You don't have to include everything ever written, but do make it clear that you know the relevant literature. Omissions of crucial relevant works can hurt your credibility. On the other hand, don't try to include works by all your potential reviewers with the idea of prejudicing them in your favor. Reviewers aren't egomaniacs -- don't cite things that are irrelevant, or just to pay lip service. And don't cite work you haven't read! It's better to omit a work than to cite it but reveal your ignorance of its contents. (Don't rely much on other people's bibliographies; sometimes they contain big errors or are misleading.)

    Do try to make the whole proposal easy to read and understand. Make it easy on the eyes: use dark type and good xerography. Perfect copy isn't required, nor slick typography. Some forms of right-justification actually make copy hard to read -- check how your word-processor does it. But of course, keep typos to a minimum -- use a spell-checker and have someone proofread for you. This is especially important if you have special symbols (Greek letters, sub/superscripts, whatever) whose presence is crucial to the intelligibility of your proposal.

    If your institution has a writing lab or other source of editorial services (many do), use it. Otherwise, ask a colleague who is not directly in your specialty to read and critique your proposal. Clarity is important. Your reviewers will try hard to figure out what you mean, but they might miss your points; why risk it?

    If you are re-using existing materials from papers or prior grants, in addition to checking that your references are up to date, make sure that the style and content really fit in with the rest of your proposal.

    If you happen to have a sense of humor, it is not necessary to suppress it entirely. But don't overindulge: it is a bad idea to be cutesy, to distract the reader from the real point, or to take up important space. And cute titles are not recommended.



If you are proposing a complex series of experiments, restate the point of each one after describing it, and show in summary how they fit together.

If you are going to conduct interviews, who are you going to interview in terms of the group under study? Make sure your elicitation procedures are adequate to obtain the variations you need.

Either show the intrinsic value of doing exploratory/ descriptive work, or clearly explain its theoretical relevance, showing what would count as evidence for or against your model or theory.

How will they be organized? Who are the potential users, and is the proposed presentation of material appropriate for their needs? Contrast yours with existing ones (if any): if you have a theoretical bias, don't hide it, but show that your product will be useable to people with other approaches as well.

Show that you understand the conditions in the field and are prepared to cope with them.

Why is the chosen population or data base the right one to work with to address your problem? Incidentally, it is fine to use an existing database (archives of videotpes, grammars, manuscripts, etc.), but show that it is capable of providing the kinds of data you need to address your issues. It is a good idea to provide samples of data contained in or coded from your sources.

In what way is the time ripe for such a conference? What are its goals, and/or what is to be gained? Is there a theme? Several themes? How will the structure of the conference further the stated goals? Why the particular division into sessions? If the conference is a "working" one, and data is to be presented and analyzed at the conference, show how it will be circulated ahead of time so that the conference time itself will be maximally efficient. Why have these particular people been invited? What will each of them contribute? Are differing viewpoints and/or theoretical approaches represented? Why have the participants been assigned their particular topics? A good way to justify invitations is to list a few relevant publications by each participant -- recent ones, if possible.

Make it clear what the special requirements of the project are. Evaluate reasonable alternative equipment, and make a case for the equipment you have chosen. Don't pad the price estimates; someone will notice and may become indignant about it.