Some issues in statistical methodology grant writing

Note: Some items may apply only to NIH, or R01, or statistical methodology development.

Differences from writing a paper
  1. A lot more are at stake (lots of money, longer cycle, your future, etc.)
  2. You write multiple papers simultaneously.
  3. You donít include full evaluations of the proposed methods and extensions.
  4. You need to explain the rationale and methods well enough to convince the reviewers that the proposed direction is viable.
  5. You might cite more, because you donít know who will review your grant.
  6. Each proposed method is scattered in three parts of the grant. B ≈ Introduction; C ≈ Results (simulations and evaluations); D ≈ Methods + what you propose to do + potential extensions + limitations. Thus, cross-referencing is needed to avoid redundancy.

Items needed
  1. Key personnel and their biosketches
  2. Budget Justification (Why you and co-investigators qualify to carry out the propose work. If space permits, you may also repeat this in C.)
  3. Letters of support (You probably need to draft the letters.)
  4. Facilities & Other Resources / Equipments: Department resources, ACCRE, etc.
  5. Cover letter (Recommend institute(s) for funding and study section(s) for review)
  6. Abstract ! (about half a page)
  7. Narrative (for lay people; 3 lines)
  8. A. SPECIFIC AIMS ! (1-1.5 pages; 1 page is preferred)
  9. B. BACKGROUND AND SIGNIFICANCE (4-5 pages; may include real data descriptions, if any)
  10. C. PRELIMINARY RESULTS (4-5 pages; may include experiences of investigators)
  11. D. RESEARCH DESIGN AND METHODS (about 15 pages, including timeline and logistics)
  12. References
  13. Optional: Consortium/contractual agreement and consortium justification

Review process
  1. Three reviewers are assigned to review your grant. The primary reviewer probably is in the same field as yours, while the secondary and tertiary reviewers may be in relevant fields. These three almost determine the fate of your application.
  2. The reviewers review your grant before the study section meet. Each reviewer will review 10 or so grants. They have full time jobs. They may read your grant at the last minute (on the plane, the night before the study section meeting, after having beers with friends who came to the same section, etc).
  3. The study section has about 20-30 people, half permanent (4 years) and half invited. Everyone in the study section has received your grant even if he/she is not assigned to review yours; so in theory, everyone in the study section can evaluate your application (except for those with conflict of interest). For each grant, they (mainly the three reviewers) determine if it is worth discussing (about 50% grants are discussed and about 10% are funded). If they decide not to discuss your grant, it will not be scored; you still get the reviews.
  4. If your grant is discussed, the reviewers will talk about their impressions, good and bad parts, etc, and everyone can join the discussion. Except for the three assigned reviewers, the other people probably didnít read your grant; some may read your abstract or section A before or during discussion. [This is why the abstract and section A are important.] After about 10-15 minutes, they will score; your final score is the average of their scores (◊100). Everyone can score your grant even if he/she hasnít read it (except for those with conflict of interest). Their perception of the quality of your grant mainly depends on the reviewers and how the discussion goes.
  5. Because of all these, it is vital to write as clearly as possible and make the grant as organized as possible. Donít assume the reviewers have enough time to ponder over your grant or they are smart enough to fill the gaps. Write your grant as if you are writing for Scientific American.

What you need to do
  1. Plan early. You need to have more than one idea. The ideas should be good and can be convincingly demonstrated to be promising (without full simulation results).
  2. Write well. Think what a reader perceives while he reads. I benefit tremendously from The Science of Scientific Writing by George D. Gopen and Judith A. Swan (1990).
  3. Ask other people to read your grant. Leave enough head time so that other people will have time to read, and you will have time to address their suggestions.
  4. Know who will (likely) be in the study section (the roster). Cite their relevant papers.
  5. Emulate successful people (your advisors, senior colleagues) and take their suggestions seriously. The slides by Marie Davidian and the articles by Louise Ryan on this topic are helpful.

Journeys

Chun's journey (my R01 application was due on Monday, 01/29/2007, 5 pm)
  1. I had almost all the ideas in August, 2006, and decided to submit an R01 for the February 1 deadline. My collaborator had a new idea in November, and it became a sub-aim. I did some preliminary explorations on some of the ideas, and felt they were viable. It turned out that the one idea on which I didnít do any evaluations wouldnít work. But luckily I tweaked it to make it work one week before the deadline.
  2. We created a list of aims in October. About half of the aims were dropped or changed to other ones by late December.
  3. In November, we started to work on two of the ideas as separate papers. By early January, one paper was almost finished; the other was done with detailed method explanation and some simulation results to support the method. So, by January 10, almost half had been done in paper format.
  4. We decided to drop a sub-aim in mid January, because it wasnít strong enough and could be easily criticized.
  5. From January 14 to 28, I sat still writing whenever I was awake. (Ming helped a lot! I turned down my sonís many requests for play. Both Ming and my son sacrificed a lot!)
  6. I was debating whether to drop the last aim on 01/20/2007 because it wouldnít work. I ended up tweaking it into a different approach.
  7. By Friday (01/26/2007), I had finally gone through all the method aims (some were written by my collaborator), and modified some approaches as necessary. After the midnight, I sent the draft to a few people to read. Luckily, some of them read it through over the weekend and sent me feedbacks. Unfortunately, a key person didnít have time to read it at all. [My fault. I didnít plan early enough.]
  8. By Sunday afternoon (01/28/2007), I had finally gone through all the sections and written up the abstract and the cover letter. My collaborator and I had a final read-through that night. It took me 3-4 hours to read it through! I had the final version ready by around 3 am.
  9. In the morning of the final day (01/29/2007), I spent about three hours on writing budget justification, contractual agreement/justification, resources and equipment.
  10. During the final days, I also spent hours on the so-called non-scientific parts (non-ABCD), including budgeting, contacting people for support letters, drafting support letters, trying out various margins and spacing, and getting familiar with PureEdge.
  11. My collaborator helped a lot with writing and editing during the whole process.
  12. I wish I had started writing earlier.
  13. My final advices:
    • Writing makes you focus. Writing with a deadline is painful, but may be fruitful.
    • If you think the grant you write doesnít have good quality, donít submit. Your reputation will be damaged if you submit a bad grant.
Topic revision: r1 - 23 Feb 2007, ChunLi
 

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback