Write for Your Advocate: How to Draft an NIH Application Reviewers Can Defend in the Room

What Actually Happens When Your Application Is Discussed

The mechanics vary by study section and chair, so treat what follows as the common pattern rather than a rule. The scientific review officer calls your application. The assigned reviewers state their preliminary scores. One of them, usually the first reviewer, summarizes the project and walks through strengths and weaknesses. The others add what they agree with and what they'd push back on. The chair opens the floor, and then everyone votes, with the assigned reviewers' range anchoring the scores.

Notice what that means. Panelists who weren't assigned your application may have skimmed it, or may not have opened it at all. For them, your application is whatever the assigned reviewers say it is, plus whatever gets asked in the room. Nobody reads your Approach aloud. They paraphrase it, and they paraphrase the version they can hold in their heads.

I've sat on panels where a strong application lost points because the lead reviewer couldn't explain in a sentence why the second aim was necessary. The aim was fine. The explanation just wasn't on the page in a form anyone could repeat. That's a fixable problem, and it's on the writer.

Build a One-Sentence Takeaway Reviewers Can Repeat

Before you write anything else, draft the sentence you'd want your lead reviewer to say at the start of the discussion. Something like: "This team proposes to test whether X drives Y in Z patients, using a model they built, and if it works, clinicians will have a measurable marker they don't have today." If you can't write that sentence, the reviewer can't either, and they'll fill the gap with their own framing, which may not be yours.

Then make sure that sentence, or something close to it, appears in three places: the project summary, the end of the first paragraph of Specific Aims, and the opening of Significance. Repetition feels redundant to you because you know the project. To a reviewer assembling a critique at 11 p.m. from notes on six applications, it's a landmark. For a deeper look at the page where this sentence lives, see our guide to structuring the Specific Aims page.

Write Quotable Sentences, Not Just Correct Ones

A correct sentence is accurate. A quotable sentence is accurate and self-contained, so a reviewer can lift it into their critique without needing the surrounding paragraph. Compare these two versions of the same claim.

Practical habits help. Lead each paragraph with its claim instead of ending on it. Put the number next to the noun. Name the specific cohort, model, or dataset instead of "our established system." And watch pronouns that point back across a page break; "this" and "these" are fine in a draft and risky in a sentence someone has to read out of context.

One caution. Quotable doesn't mean punchy for its own sake. Reviewers have heard every superlative there is, and a paragraph of confident adjectives reads as a lack of data. Concrete beats clever.

Pre-Answer the Three Criticisms You Already Know Are Coming

Every application has weak points, and you usually know where yours are. Maybe the sample size is modest. Maybe Aim 3 depends on a reagent you haven't validated yet. Maybe the team has no one with experience in the clinical side. Reviewers will find these. The question is whether the answer is already on the page when they do.

List your three most likely criticisms, honestly, then give each a short, direct answer where the reviewer will be looking: in the Approach, under a heading like "Potential problems and alternatives," or in a sentence right after the claim that raises the concern. The tone matters. A defensive paragraph sounds like worry. A calm one ("If recruitment lags, we'll add the second site, which has agreed in writing") sounds like a plan.

I can't promise this changes a score by any particular amount, and nobody publishes that number. What I can say is that the discussion tends to go better when the concern and the answer are in the same place, because the reviewer can read both at once.

Make Every Claim Easy to Find Again

During discussion, someone will ask, "Did they show this works in a second model?" The reviewer then flips through a PDF, fast, under social pressure. If the answer takes forty seconds to find, they'll often say they don't recall seeing it. Structure is a form of advocacy.

Use informative subheadings that state the point, such as "Preliminary data show the assay separates responders from non-responders," rather than "Preliminary Data." Number your figures consistently and reference them by number in the aim they support. Keep a repeating internal structure across aims (rationale, approach, expected outcomes, pitfalls) so a reviewer who has learned where to look in Aim 1 can find the same thing in Aim 3. Our notes on figures and formatting in the Research Strategy cover the layout side.

Figure legends deserve special attention. Many reviewers read figures and legends first and the prose second. A legend that ends with a plain-language conclusion ("Together, these data show that the marker is detectable before symptom onset") doubles as a quotable sentence.

A Practical Test: The Two-Minute Read-Aloud

Ask a colleague in an adjacent field to read your Specific Aims and Significance, close the file, and then do what your lead reviewer will do: summarize the project out loud in about two minutes, then name the two biggest weaknesses. Record it, or take notes without interrupting.

Compare what you hear to what you intended. Anything they skip is probably buried. Anything they get wrong is probably ambiguous. Any weakness they name that you didn't pre-answer goes straight onto your revision list. This is a lighter-weight cousin of a full mock study section, and it's something you can do in an afternoon when a full mock isn't feasible.

Do it twice, with different people, because one reader's blind spot is often your own.

Frequently Asked Questions