walking through an NIH grant:

Grant$ for Lunch Handouts
by Janet S. Rasey, Ph.D.
Director, Research Funding Service
University of Washington
Seattle, WA

Specific Aims

Important Points to Remember

The Abstract and the Specific Aims are the only parts of your grant that many reviewers will read. Let each give a 'snapshot' of the whole proposal.

Grant reviewers look for answers to several questions: Who will do the work? What work will be done? How will the work be done? How much will it cost? Why are you doing the work? Why is it worth doing? Where is the work going?

The Specific Aims must answer the question: What are you going to do? A good set of aims also will give some very brief information about why you are doing the work and why it is worth doing.

It may seem obvious, but it bears repeating: Specific Aims must be specific (not glowing generalities about an area you want to investigate) and they must aim in a particular direction (not just "run in place," collecting data).

An aim needs a hypothesis, which may be explicitly stated or implied (less preferable). A good hypothesis has elements of an "if-then" statement: If I do "X" to my biological system, then I expect "Y" to occur. A well stated hypothesis causes the reader to visualize one or more explicit experiments.

Rarely, a specific aim may involve collecting baseline information in a new area or have as its goal the development of a new technique. Never try to build a whole proposal on such aims. Never propose to develop a new technique as your first aim, then have subsequent aims depend on the new method. If Aim 1 fails, there goes the whole grant.

Handout from November 13, 1996

Budgets and Budget Justifications

A grant budget states how much money you need for your research. The budget justification makes your case for the specific items and the amount you request. These parts of the grant also reveal something about:

How carefully you plan

Does the size of the budget match the number of experiments?

How much attention you pay to details

Is the addition correct?
Is every listed item needed?
Is each essential item listed?

How thoroughly you document and explain

Is every item in the budget justified equally well?
Have you made a believable case?

How realistic you are

Is the budget in line with that for similar projects being done by others?
In line with budgets typically awarded by NIH?

A reviewer may assume that you manage time and personnel resources the same way you manage money.

Basic Budget "DOs"

DO ask for what you need to do the work.
DO be aware of a typical budget for applications to a particular grant maker.
DO check prices and arithmetic carefully.
DO itemize with sufficient detail. Be a "splitter" rather than a "lumper".
DO write a detailed justification for all budget categories.

Basic Budget "DON'Ts"

DON'T pad the budget. It makes you look greedy.
DON'T try to do research on a shoestring. It makes you look naive.
DON'T budget for more experiments than can be performed with the requested personnel in the requested grant period.
DON'T use the phrase "based on recent experience in running our laboratory" as a justification for your supplies budget.
DON'T assume the supplies budget is obvious. This often is the biggest item after salaries and the easiest place for reviewers to cut. This section is often very poorly justified.

Budget Justification

Many NIH grant applicants do an uneven job of justifying different parts of a grant's budget.

Personnel -- usually justified very well
Equipment -- usually justified well
Supplies -- often justified very poorly
Consultant Costs -- variable justification
Travel -- often justified poorly
Other Expenses -- variable justification
An Experienced Reviewer's View on R01 Budgets, 1997
Assistant Professors $120,000/yr. Direct Costs
Associate Professors $170,000/yr. Direct Costs
Full Professors $220,000/yr. Direct Costs
(All numbers above have error limits of plus or minus 20%. Add $50,000/yr. for clinical trials)How to attempt R01 suicide: Come in with a budget in excess of $300,000/yr.
For young investigators, an alternative to an R29 is a three year R01 proposal for approximately $90,000/yr.

Handout from January 15, 1997

Problems with Preliminary Data

Problems of Content

Question: How good are the data? Are they convincing? Appropriate controls done? Large or small error bars? Any statistics done? Reproducibility between replicate experiment?
Solution: Data from replicate experiments with good reproducibility, small error bars, carefully analyzed, properly controlled

Question: Do the data conflict with data elsewhere in the grant?
Solution: Check all data carefully to make sure you do not contradict yourself

Question: If new, unproved methods or systems are proposed, is there any comparison to established techniques? (Far better to have this in preliminary data than to make in Specific Aim 1 of the grant)
Solution: as soon as you develop a new method, system, or reagent, compare it to the most commonly used or standard method in your field

Question: Have you put in so much data that the reviewers will wonder why you're asking for more money?
Solution: Don't feel you have to put in everything you have done on the subject of your proposal; save it for the first progress report. (The advice of an experienced grant writer is useful here.)

Question: Do you have too few preliminary data to make a case for you proposal?
Solution: Get an opinion from one or two experienced grant writers; this is sometimes difficult to judge, particularly for junior investigators

Problems of Presentations

Question: Are related graphs labeled similarly? Have similar units been used throughout?
Solution: Have someone not familiar with your work look over the graphs to check for inconsistencies and clarity

Question: Do all figures have legends? Are axes labeled in an intelligible manner? (Graphs taken directly from the lab notebook may have very cryptic, short-hand labels)
Solution: Same as for B1, above

Question: Are figures and tables adjacent to the text that describes them?
Solution: Finish the final copy of the grant early enough to make sure that figures, graphs, and tables are near the text that relates to them. Text-wrap features of word processing programs now make this possible. (Don't make the reviewer work to find the data being described in the text!)

Question: Are all the relevant figures and tables in the grant, rather than the appendix? Are all the figures and graphs in the proposal necessary to make your point?
Solution: Avoid the temptation to escape page limits by relegating essential figures to the appendix. Make sure the grant is clear to someone who does not have the appendix. Use only those figures and tables that make a necessary point

Question: Are the figures and tables legible? Are the letters and data points too small? Are there too many curves on the same graph? Can symbols for different curves on the same graph be distinguished from each other? Are different lanes on a gel identified with words?
Solution: Early on, prepare and size the figures you want to use to make sure they are legible. Make sure that you have a final copy in hand early enough to check for legibility. Have an intelligent non-expert read the application for you

Problems of Relevance

Question: are the data recent, or are they taken from the applicant's publications? ( A little bit of published data are OK, but not much.)
Solution: Be sure you do not weight the preliminary data too heavily with previously published material; have fresh data generated for this application.

Question: Have the preliminary experiments used the same species, cell line, or system as that proposed for the future experiments?
Solution: This one is obvious; do the relevant experiment.

Question: Are there some experiments illustrating the applicant's use of an established technique that s/he has not previously reported using?
Solution: Be sure to do some experiments to show that you can do the techniques you propose. No matter how expert you are with other methods and no matter how well the new methods are described later in the grant, there is no substitute for data.

Handout from February 19, 1997