The summary statement you get back is not a complete list of what's wrong with your grant.

It arrives after the decision, reflects only what reviewers chose to write down, and under current procedures may contain less than ever. 

Meanwhile, competition has tightened: NIH funded 13.0% of R01-equivalent applications in fiscal year 2025, down from 21.6% two fiscal years earlier, while applications climbed past 42,000. Every weakness that survives to submission now costs more than it used to.

The most rigorous applicants have responded by running grant proposal review on their own work first, reading the proposal the way a reviewer will, as a set of claims to be tested against evidence.

That makes the goal of reviewing grant proposals yourself simple. Find the weaknesses a reviewer is likely to catch, test whether each claim holds up, and fix the gaps while there is still time to act.

‍

Why Reviewers Reject Strong Science

Funders publish their own lists of the problems that sink applications, and those lists say remarkably little about wrong science. They describe failures of the case: arguments left unclear, unsupported, or unconvincing on the page. 

The National Institute of Mental Health's list of common application mistakes organizes these grant application weaknesses by review criterion, and the pattern repeats across institutes:

• Aims that are too ambitious, unfocused, or unclear about where the work leads
• Significance claimed in adjectives, with no evidence-based argument behind it
• Approach sections silent on controls, alternative interpretations, and pitfalls
• Feasibility doubts, from scope that outruns the timeline to methods with no demonstrated expertise behind them

Each of these is a failure of presentation and reasoning, and each is findable before submission. A reviewer can score only what's on the page. That constraint works in your favor exactly once, before you submit.

‍

‍

How Reviewers Actually Read a Proposal

Assigned reviewers read your application alongside a stack of others, on top of their own research, teaching, and deadlines. 

The reading is structured and criterion-driven. Under NIH's simplified framework, the scored grant review criteria are the importance of the research, rigor and feasibility, and expertise and resources. Reviewers are instructed to evaluate the application as presented and are explicitly told to avoid redesigning it.

The reading is also far less uniform than applicants tend to assume. When researchers replicated the NIH review process with 43 reviewers evaluating the same 25 applications, they found no agreement among reviewers in their qualitative or quantitative evaluations, including in how reviewers translated a given set of strengths and weaknesses into a numeric score.

Current procedures raise the stakes further. Under temporary modifications in place through the October 2026 council round, roughly the top third of applications are discussed in the study section, the middle third is designated competitive but not discussed, and summary statements have been pared back to a consensus sentence, the main score-driving points, and the three assigned critiques.

The practical implication is signposting. Put the significance argument where a fast reader will meet it, in the abstract and the opening of the aims page. Make each aim legible on its own. If understanding your argument requires assembling pieces from different sections, different reviewers will assemble different arguments, and the panel will score the versions they assembled.

‍

The 7 Weak Points Reviewers Find First

NIH's Center for Scientific Review publishes the problems reviewers cite most often when critiquing applications. Consolidated, they come down to seven weak points, each visible on the page before submission.

‍

1. No New or Original Idea

The top-cited problem on CSR's list. It looks like an incremental extension of published work, presented as if the increment were self-evidently important. The fix is a sentence you should be able to write in one breath: what is new, relative to which prior work, and why the field moves if you are right. If that sentence will not come, the aims need rethinking before they need rewording.

‍

2. A Rationale That Doesn't Hold

Significance sections fail when they name the gap in knowledge but never demonstrate it. Reviewers look for a chain from the documented gap to the hypothesis to the expected payoff. Build each link on a source that genuinely supports it, and state the payoff in terms of what the field can do once the work is complete.

‍

3. Unfocused Aims and Unclear Direction

This appears in two ways: aims that overlap or hinge on one another succeeding, and aims so narrow that reviewers question where the work leads. Write aims that are independent, falsifiable, and achievable within the project period, each with one sentence on what success opens up next.

‍

4. Questionable Reasoning in the Approach

Reviewers flag experiments that sit adjacent to the hypothesis without testing it directly, and approach sections that never mention alternative interpretations. Map every experiment to the specific claim it tests. Name the plausible alternative explanations and show how the design discriminates between them.

‍

5. Missing Experimental Detail

Controls left implied, power calculations unstated, analysis plans compressed to a sentence. Give a skeptical methodologist enough to judge feasibility: controls, sample sizes, statistical approach, and the criteria you will use to interpret each outcome.

‍

6. Gaps in the Published Literature

At least one panel member will know the recent or contradictory work your background section skipped. Engage inconvenient findings directly. A paragraph explaining why your hypothesis survives a contradictory result reads as command of the field. Omission reads as evasion.

‍

7. More Work Than the Team Can Credibly Deliver

Overambitious scope shares a root with its quieter sibling: an essential method with no demonstrated expertise behind it in the biosketch or publication record. Cut scope until the timeline is believable, or recruit a collaborator with the track record and include the letter. Reviewers score the team they can verify.

‍

How to Pressure Test Your Evidence Base

The least-checked weakness lives in the citations. 

A 2025 meta-analysis spanning 46 studies and roughly 32,000 quotations found that 16.9% of quotations in the medical literature were incorrect, with about half classified as major errors, meaning the cited source does not support the claim at all. 

Grant rationales draw on the same literature and the same citation habits. A major quotation error under a load-bearing claim hands the reviewer a score-driving gap.

The manual check is straightforward and slow because every claim needs three passes:

• List the load-bearing claims: the statements in your significance and approach sections that, if wrong, collapse the argument.
• Reread each source against the claim: confirm the paper shows what your sentence says it shows, at the strength you say it.
• Check the scope of every finding: a result in one cell line, cohort, or model does not automatically carry to yours. Mark where you are extrapolating, and say so in the proposal.

Under a deadline, few applicants complete this for every claim. QED Science built Grant Review to do it systematically. It breaks a proposal into its individual claims, explicit and implicit, and tests each against the published literature, surfacing where the evidence supports the argument and where it falls short. 

Grant Review uses the same validated approach to claim-level quality assessment published in the QED Score white paper to review the science and the claims, leaving every decision about what to change with the researcher.

‍

What to Do With the Weaknesses You Find

Sort the list into two piles. The first pile can be fixed before the deadline: missing detail, unsupported citations, overlapping aims, scope cuts, a clearer significance argument. The second pile is structural: a preliminary-data gap that needs months at the bench, or an expertise gap that needs a collaborator you have yet to recruit.

For the structural pile, disclose and address. A limitation you acknowledge, paired with a mitigation plan, reads as rigor. The same limitation discovered independently by a reviewer reads as a blind spot and colors the rest of the critique.

Part of triage is also deciding which flagged weaknesses are genuine. Some reflect deliberate trade-offs made under page limits and budget realities, and separating genuine gaps from noise is its own discipline in review work. Address those with a sentence of justification, and save the redesigns for the weaknesses that threaten the argument itself.

Build the list expecting no help from the process itself. NIAID notes that once reviewers find a fatal flaw, they may stop discussing an application, so the feedback ends there, and you can fix everything in a summary statement and still be caught by problems no one wrote down. A pre-submission review exists to produce the inventory the review process will never hand you.

‍

Build the Complete Weakness Inventory Before You Submit

Every proposal eventually gets a rigorous read. The question is whether the first one happens while you can still act on it.

Each of the seven weak points above is, at bottom, a claim that has to survive contact with the evidence. That is the test Grant Review runs. It breaks your proposal into its individual claims, checks each against the published literature, and shows you where the case holds and where it needs work, delivering in minutes an assessment that typically takes months to come back.

Upload your proposal to Grant Review and see each claim tested against the published literature before a study section reads a word, so you can find the gaps, fix what can be fixed, and submit knowing how your case holds up.

‍

FAQs