How to Respond to Grant Panel Reviews (NSF, NIH, ERC)

Your proposal was not funded. The summary statement runs four pages, half of it reads like praise, and somewhere inside it is the actual reason. Before you touch the document you have one decision to make. Is this the same proposal with better execution, or is it a different proposal?

Grant panel reviews run on a different logic from journal peer review, and the mistake most PIs make is treating them as interchangeable. Journal reviewers evaluate whether your work is correct and significant. Grant panels evaluate whether your proposed work is worth funding, which introduces dimensions (feasibility, team qualifications, budget justification, broader impacts) that journal reviewers never consider. The grant proposal rebuttal, in turn, answers a different species of concern. Not "is this right?" but "will this succeed, and is it worth the money?"

What follows covers panel feedback from the three big funders (NSF, NIH, ERC), along with the strategic decisions that determine whether a resubmission succeeds or costs you six months.

How Grant Panel Reviews Differ from Journal Peer Review

The structural differences are worth stating explicitly, because they shape every part of the rebuttal strategy.

Panels evaluate proposals and not completed work. Journal reviewers can check your results, replicate your analysis, or verify your claims against the data. Grant panelists are evaluating a plan, and plans by nature involve uncertainty. The panel's job is to assess whether your plan is plausible and well designed, and whether it is likely to produce meaningful results. Your rebuttal therefore has to answer concerns about feasibility and design as much as concerns about scientific merit.

Panel reviews are consensus documents. Unlike journal reviews, where Reviewer 1 and Reviewer 2 may hold diametrically opposed views, a grant panel summary represents the group's negotiated assessment. Individual reviewer scores vary, and the summary reflects what the panel as a group decided matters. That means the summary carries more weight than any individual review, and contradicting it takes stronger evidence than contradicting a single journal reviewer.

Scores constrain the program officer's discretion. At NSF and NIH, the program officer can fund proposals outside the top-ranked group, and doing so requires justification. Proposals that scored poorly cannot be rescued by a brilliant rebuttal alone. The scores have to improve on resubmission. The rebuttal is a means to that end and not an end in itself.

For the general principles of responding to reviewer feedback that carry across both journals and grants, see the systematic process for responding to peer review comments.

Decoding Panel Summaries: What the Language Actually Means

Grant panel summaries use a distinctive vocabulary that can be opaque to early-career PIs. The language is chosen carefully, and the specific phrases carry more information than they appear to at the surface.

"The proposal is not competitive" means the panel found fundamental weaknesses, in the research question, the methodology, or the team's qualifications, that minor revisions cannot fix. This is not a signal to resubmit with cosmetic changes. It is a signal to rethink the project's core framing or scope.

"The proposal would benefit from a clearer articulation of…" is a precise diagnostic. Whatever follows that phrase is something the panel actively discussed and found wanting. Treat it as the highest-priority revision target.

"While the topic is timely, the proposed approach raises concerns about…" means the panel agrees the question is important and does not believe your plan for answering it is adequate. That is good news (the topic has support) wrapped around a specific critique (the design needs work).

"The investigator is well-qualified, but…" signals that the panel's concerns are not about you personally but about the proposal's design, scope, or feasibility. Do not read it as reassurance. Read it as a redirect toward the specific weaknesses they identified.

"The broader impacts are underdeveloped" is one of the most common NSF criticisms and one of the easiest to fix. It usually means the broader impacts section was generic, vague, or clearly an afterthought.

"Enthusiasm was dampened by…" is NIH study section language for the thing that actually lowered your score. Whatever follows is the sentence to build your entire revision around.

NSF-Specific Guidance: How to Read and Respond to Panel Summaries

NSF proposals are evaluated on two merit review criteria, Intellectual Merit and Broader Impacts. Panel summaries address both, and your resubmission has to show improvement on whichever criterion was weaker. For what panels systematically get wrong when they apply those two criteria, see the NSF merit review criteria breakdown.

Reading the summary: NSF panel summaries typically open with a brief characterization of the proposal, then strengths, then weaknesses, then a summary assessment. The weaknesses section is your revision roadmap. Every item listed there is something at least one panelist raised and the group agreed (or did not disagree) was a legitimate concern.

The resubmission strategy for NSF. NSF has no formal "rebuttal letter" mechanism for most programs. Instead you revise the proposal and resubmit it, sometimes to the same program and sometimes to a different one. The revised proposal should answer the panel's concerns inside the proposal itself instead of in a separate response document. Some PIs include a brief "Changes from Previous Submission" section at the beginning. NSF guidelines permit this, and it is useful for signalling to panelists (who may or may not be the same people) that you took prior feedback seriously.

Common NSF failure modes:

• Overpromising relative to timeline and budget. Panels are skilled at spotting proposals where the proposed work cannot realistically be completed with the resources requested.
• Weak preliminary data. For established programs, the expectation is that you have enough preliminary results to demonstrate feasibility. Not enough to answer the research question, since that is what the grant is for, and enough to show the approach works.
• Generic broader impacts. "We will train graduate students" is not a broader impacts plan. It describes what every funded project does by default. Broader impacts that are specific and measurable, ideally with a named partner organization, score substantially better.

NIH-Specific Guidance: Study Section Critiques

NIH still evaluates research grant applications against five review criteria: Significance, Innovation, Approach, Investigator(s), and Environment. What changed for applications with due dates on or after January 25, 2025 is how those criteria are scored. Under the simplified review framework, reviewers score two factors on the 1–9 scale (1 is best). Factor 1, Importance of the Research, covers Significance and Innovation. Factor 2, Rigor and Feasibility, covers Approach. Factor 3, Expertise and Resources (Investigator and Environment), is no longer scored. Reviewers rate it as appropriate or flag additional considerations. The overall impact score remains the panel's judgment of the application as a whole and not an average of the factor scores. Fellowship and training mechanisms and career development awards use their own criteria, so check the funding opportunity before you map your revision onto any of this vocabulary. What the change means at the study section level is covered in the NIH study section post.

Reading the critique: NIH critiques are more structured than NSF summaries. Each reviewer supplies individual scores and narrative comments per criterion, followed by discussion in the study section. The "Resume and Summary of Discussion" captures the key points from the group conversation, and it is the most important section of the entire critique, because it reflects what the panel agreed on and not just what one reviewer thought.

Prioritizing your revision: Look at the factor scores and the narrative under each. If Factor 2 (Approach) scored a 5 but Factor 1 (Significance and Innovation) scored a 2, the Approach section is where your revision effort belongs. The common mistake is distributing revision effort evenly across all criteria. Your scores are telling you exactly where to focus.

The NIH resubmission mechanism. NIH allows one resubmission (the "A1"). The resubmission includes a one-page "Introduction to the Revised Application" that responds directly to the prior critique. That page is the closest thing in grant funding to a formal rebuttal letter, and it deserves the same strategic attention as a journal response document. Address every major concern. Be specific about what changed. Do not be defensive.

Common NIH failure modes:

• Underpowered study design. Study sections scrutinize sample size justifications, and a power analysis that does not align with the proposed aims will be flagged.
• Insufficient innovation. "We will apply existing method X to new population Y" is rarely innovative enough for R01 funding. The panel wants to see something that advances the field's methods or conceptual framework and not just its data.
• Approach lacks rigor. This is the most common weakness across NIH proposals. Rigor means pre-specified analyses, consideration of biological variables (sex, age), appropriate controls, and a clear plan for handling the full range of possible results, including null findings.

A Worked Example: Turning a Summary Statement Into a Revision Plan

Numbers make this concrete. Suppose an R01 comes back discussed, with an overall impact score of 42, Factor 1 (Importance of the Research) scored 4, Factor 2 (Rigor and Feasibility) scored 6, and Factor 3 (Expertise and Resources) rated appropriate with no additional considerations. The Factor 1 narrative praises the importance of the question and doubts the innovation claim.

Read the shape before you read the words. Factor 3 raised no concerns, which means the panel believes you can do the work in the place you propose to do it. Factor 1 is split, with Significance carrying it and Innovation dragging it. Factor 2, at 6, is the score that most heavily influences the overall impact score at most study sections. That distribution is the best kind of bad news, because the problem is local.

Now read the narrative under both scored factors twice, and extract the specific sentences. Suppose Approach draws three recurring objections: Aim 2 depends on Aim 1 succeeding, the primary outcome measure has not been validated in this population, and the analysis plan does not say what happens if the effect is null. Suppose Innovation draws one, which is that the method is an established technique applied to a new sample.

That is a revision plan, and it writes itself.

The dependency between aims is fixed structurally. Rewrite Aim 2 so it produces an interpretable result under either outcome of Aim 1, and say so in one sentence in the Approach overview. Panels flag aim dependency constantly, and it is the single easiest Approach criticism to remove.

The unvalidated outcome measure needs preliminary data, which is why you start in month one. One small validation sub-study plus a figure and a paragraph will usually settle it.

The missing null-result plan costs half a page and buys back more score than anything else on the list, because it demonstrates exactly the kind of thinking study sections say they want.

Innovation is the harder one, and the honest question is whether the innovation claim was overstated or under-explained. If the conceptual advance genuinely exists and you buried it under methods description, restructure the Significance and Innovation sections to lead with it. If it does not exist, no amount of adjectives will create it, and that is where you consider whether the A1 should be a different proposal.

Notice what did not happen. Nobody rewrote the Environment section, which drew no additional considerations, and nobody added citations to the Significance paragraphs the narrative already praised. Effort followed the scores.

What Goes in the One-Page Introduction to the Revised Application

You get one page. Spend roughly two thirds of it on the concerns that moved the score, and use this shape:

Introduction to the Revised Application

We thank the reviewers and the study section for the thorough evaluation of [APPLICATION TITLE] (Application ID [ID]). The prior overall impact score was [SCORE], with the principal concerns centered on [CRITERION] and [CRITERION]. We have revised the application substantially in response, and we summarize the changes below.

Summary of changes

- [CHANGE 1: name the concern it answers, what is now different, and where it appears in the revised application]

- [CHANGE 2]

- [CHANGE 3]

Concern 1 ([CRITERION], Reviewers [N] and [N]). [ONE-LINE PARAPHRASE OF THE CONCERN]

[What changed, with the specific new element: revised aim, new analysis, power calculation, preliminary figure. Name the page and section of the revised application.]

Concern 2 ([CRITERION], Reviewer [N]). [ONE-LINE PARAPHRASE]

[Response.]

Concern 3 ([CRITERION], Reviewer [N]). [ONE-LINE PARAPHRASE]

[Response.]

Text changed in the revised application is marked [per the instructions in the funding opportunity].

Four rules keep that page effective. Paraphrase each concern in one line instead of quoting the critique at length, because quoting eats the page you need. Group concerns raised by more than one reviewer and say so, since a shared concern outranks a solo one. Never argue with the score itself. And where you disagree on substance, disagree once, with evidence, then describe what you changed anyway.

The formatting discipline is the same one that works for journal responses, and the copy-paste rebuttal letter templates translate directly, minus the length.

ERC-Specific Guidance: Panel Interview Preparation

The ERC process adds a dimension neither NSF nor NIH uses, the panel interview, for proposals that pass the written review stage. The interview typically runs 20–25 minutes and functions as a live rebuttal. Panelists ask about the weaknesses identified in the written reviews, and your answers determine the final ranking.

Preparing for the interview: Request the written evaluations as early as you can and identify the two or three concerns most likely to come up. Prepare concise, evidence-based answers for each. The interview is not a presentation. It is a defense.

Common ERC interview mistakes:

• Overlong opening statement. Some panels give you 5–10 minutes for an opening. Use it to address weaknesses instead of re-presenting a proposal the panelists have already read.
• Defensive responses to questions. The panelists are testing whether you can engage critically with your own proposal. Answering "I believe my approach is correct" without evidence is the fastest way to lose points.
• Inability to articulate the contribution in one sentence. If you cannot state, clearly and without jargon, what your project will achieve that has not been achieved before, the panel will question whether the contribution is as significant as you claim.
• Answering the question you rehearsed. Panelists notice when a prepared block gets deployed against a different question. Answer what was asked in two sentences, then offer the detail.

The Resubmission Timeline and What It Costs

Resubmission cycles are long, and the calendar decides more than most PIs expect.

At NIH, standard due dates run three cycles a year, and the resubmission deadline falls about a month after the corresponding new-application deadline. Summary statements arrive weeks after the study section meets, which means the gap between reading your critique and the next viable deadline is often shorter than it looks on paper. Check the funding opportunity itself, because special programs and RFAs run their own calendars.

At NSF, many programs have moved to annual target dates or to no-deadline submission windows. No-deadline programs are a genuine advantage for resubmission, since you can take the extra eight weeks that a preliminary experiment needs without losing a year.

Build the plan backwards from the deadline, and reserve the last three weeks for internal review. Proposals resubmitted without a fresh pair of eyes tend to answer the previous panel while introducing new problems for the next one. If a colleague who has served on the panel will read it, that reading is worth more than another week of your own edits.

One number worth holding onto is that a resubmission that scores worse than the original almost always did the checklist version of the revision. The proposals that jump are the ones where something structural changed.

The Resubmission Decision: Revise or Submit Elsewhere?

Not every unfunded proposal should go back to the same program. The decision turns on the nature of the feedback.

Resubmit to the same program when: the panel's concerns are specific and addressable ("needs more preliminary data," "approach section lacks detail," "broader impacts are generic"), and the panel's language suggests the project has merit a stronger proposal could convey.

Submit elsewhere when: the panel's concerns are fundamental ("the research question is not significant," "this study has already been done"), the program fit was marginal to begin with, or you have reason to believe the panel composition was unfavorable and unlikely to change.

Rethink the project when multiple panels or programs have returned similar criticisms. That pattern points at substance and not presentation.

The One Thing Most PIs Get Wrong About Resubmission

The most common error in grant resubmission is treating the panel critique as a checklist instead of a diagnosis. PIs read each comment, make the corresponding change, then resubmit. They are surprised when the revised proposal scores worse or attracts new criticisms.

This happens because panel comments are symptoms and not causes. When a panel writes "the approach lacks rigor," checking a few boxes (adding a power analysis, mentioning blinding) answers the surface critique and not the underlying problem, which may be that the research design does not convincingly demonstrate how the specific aims will be achieved. Treating the symptom without treating the cause is a recipe for a second unfunded proposal.

The productive approach is to read the panel summary as a diagnostic narrative. Ask what overall story the panel is telling about why this proposal fell short, then revise accordingly. Sometimes that means restructuring the aims. Sometimes it means adding substantial preliminary data. Sometimes it means rethinking the approach from the beginning. For the parallel process in journal revision, see how to turn a major revision into an acceptance, and for the harder call about whether to keep going at all, see what to do after a rejection.

Responding to NSF, NIH, or ERC panel feedback? ReviewPanel.ai analyzes your proposal alongside panel critiques and generates a structured response plan with specific revision suggestions. $49/grant, no subscription.

Frequently asked questions

Can you appeal a grant panel decision?

Formal appeals exist at most agencies and almost never succeed on the science. They are designed for procedural problems: an undisclosed conflict of interest, a review that evaluated the wrong application, a factual error in the summary that changed the assessment. Disagreeing with the panel's judgment is not grounds for appeal. The productive move is a conversation with your program officer, who can tell you whether the concerns are fixable, whether the panel is likely to change, and whether resubmission is worth your time.

How many times can you resubmit an NIH grant?

NIH allows one resubmission of an application, the A1, which carries the one-page Introduction to the Revised Application. After an unsuccessful A1, the work can generally return only as a new application, which means a substantively different project instead of a lightly edited version of the same one. Because you get one shot, do not resubmit early to hit the next deadline. Resubmit when the weakest criterion has genuinely changed, even if that costs you a cycle.

Should I push back on a reviewer I think was wrong about my proposal?

Yes, once, with evidence, and then describe what you changed anyway. Panels dislike defensiveness, and they dislike the appearance that a factual error went uncorrected even more. Write one or two sentences that name the misunderstanding and point to the page that addresses it, then add a clarifying revision so the next reader cannot make the same mistake. Where a reviewer's error changed your score materially, mention it to your program officer as well as in the Introduction.

What does "not competitive" actually mean in a panel summary?

It means the panel saw weaknesses that a revision at the margin will not fix. The phrase is not a comment on your effort or your record. It is a statement about where the proposal sat in the ranked list. Treat it as a signal to change something structural: the aims, the scope, the team, or the program you are targeting. Resubmitting the same proposal with tighter prose after a "not competitive" summary is the most reliable way to spend six months for nothing.

How much of a proposal should change between the A0 and the A1?

Enough that a reviewer who read the first version notices immediately. In practice, a competitive A1 usually carries at least one structural change (a rewritten aim, a new preliminary experiment, a revised analysis plan) plus targeted fixes for every other named concern. Cosmetic A1s score worse than their A0s often enough that study sections comment on it. Where you cannot make a structural change in the time available, it is usually better to wait a cycle.

Do I have to resubmit to the same NSF program?

No, and sometimes you should not. Program fit is a real variable, and a proposal that reads as marginal in one program can read as central in another. Talk to both program officers before you decide, because they can tell you whether the panel's concerns are about your proposal or about its home. Where you do move programs, revise for the new audience instead of resubmitting the same document, since panelists in the new program will not have the prior summary and will judge it fresh.