Grant Proposal Review

Produces a section-by-section, criterion-anchored critique of a research grant proposal: structured summary, criterion-by-criterion evaluation, panel-process awareness, major/minor issues, budget and feasibility concerns, and an overall-impact recommendation. Adapted by Pradyumna Jayaram from public funder guidance (NIH, NSF, ERC) and reviewer-training resources (AAAS, ESF).

When to use

You have been invited by a funding agency or study section to review a research grant proposal.
You are running an internal mock review for a colleague, trainee, or institutional mock-panel before submission.
You are auditing an unfunded proposal retrospectively to revise and resubmit.
You are evaluating a private-foundation letter of intent or full proposal and the foundation has published its review criteria.
You are serving on a reviewer panel and need a structured template for your written critique and pre-meeting notes.

When NOT to use

For reviewing a journal manuscript (use ors-peer-review-manuscript-review).
For evaluating a researcher's overall record (use ors-peer-review-scholar-evaluation).
For writing the applicant's reply to reviewer critiques on a revised proposal (use ors-peer-review-reviewer-response).
For evaluating a single specific aim or sub-project in isolation from the full application.
For evaluating industry-sponsored research contracts or clinical-trial agreements (different review process).

Prerequisites

Recognized expertise in the proposal's topic area or methodology.
A complete, unmarked copy of the proposal, including the research plan, biosketches, budget, facilities, and any data-management plan.
Access to the funder's specific review criteria, panel description, and reviewer guidelines.
Familiarity with the funder's mission and portfolio (read the program announcement / funding opportunity announcement in full).
Time to read the proposal at least twice and to consult 2–3 key cited references where the approach rests on prior work.

Core workflow

1. Pre-flight checks

Confirm the panel, funding mechanism, and any special instructions (e.g., early-career vs. established-investigator track, new-investigator eligibility, multiple-PI rules).
Read the program announcement / funding opportunity announcement (NIH FOA, NSF solicitation, ERC work programme) in full. Review criteria vary by mechanism.
Declare any conflict of interest (financial, personal, institutional, recent collaboration, mentorship, family) before reading the proposal in depth. If a disqualifying conflict is identified, decline immediately. Funder review-management systems (e.g., NIH's Internet Assisted Review, NSF's proposal-management system, ERC's submission platform) collect COI declarations up front.
Note the funder's confidentiality rules: most panels operate under strict non-disclosure; proposals, panel discussions, and reviewer identities (in double-blind or single-blind setups) must not be discussed outside the panel.
Note the review format: written critique only, written critique plus preliminary score, or written critique with discussion at a panel meeting.

2. Initial assessment (read-through)

Read the proposal once without note-taking to gauge scope, feasibility, and fit.

What is the central scientific question or problem?
Is the proposed work within the funder's mission and the panel's scope?
Is the magnitude of the advance plausible, or are the claims over-stated?
Are the team, environment, and budget commensurate with the aims?
Is the proposal an administrative resubmission (which may have an introduction page with prior critiques and a response)?

3. Criterion-by-criterion evaluation

Map your critique to the funder's published review criteria. The mapping below covers the three most common public funders; private foundations usually publish analogous criteria in their request-for-proposals documents.

NIH-style criteria (research project grants, R-mechanism)

Criterion	What to assess
Significance	Does the project address an important problem? Will the proposed work advance the field? What is the burden of the problem being addressed, and to whom?
Investigator(s)	Are the PD/PI(s) and key personnel appropriately trained and capable? Is the track record commensurate with the proposed work? For early-stage investigators, is the institutional support adequate?
Innovation	Does the application challenge existing paradigms? Are the concepts, approaches, or methods novel? Is refinement, improvement, or new application of existing methods sufficient for the mechanism?
Approach	Are the overall strategy, methodology, and analyses well-reasoned and appropriate? Are potential problems and alternative approaches identified? For aims that depend on a risky or unpublished method, is there a fallback? Are the timeline and milestones realistic?
Environment	Are the institutional support, equipment, and other physical resources adequate? Are unique features of the scientific environment leveraged?

In addition, NIH reviewers consider Overall Impact — the synthesis of the five criteria above, weighted by the reviewer's judgment, reflecting the likelihood that the project will exert a sustained, powerful influence on the field. Overall impact is not an arithmetic average of criterion scores.

NSF-style criteria (Merit Review Principles)

NSF articulates two Merit Review Principles (intellectual merit and broader impacts) and several criteria that flow from them.

Criterion	What to assess
Intellectual Merit	How important is the proposed activity to advancing knowledge and understanding within its own field and across different fields? How well qualified is the proposer (individual or team) to conduct the project? To what extent does the proposed activity suggest and explore creative, original, or potentially transformative concepts? How well conceived and organized is the proposed activity? Is there sufficient access to resources?
Broader Impacts	To what extent will the proposed activity benefit society, advance desired societal outcomes, and broaden participation of groups underrepresented in science? How well integrated are the broader-impact activities with the project's intellectual-merit goals?
Additional context-specific criteria	Some solicitations add criteria such as data-management plan, postdoctoral-researcher mentoring plan, or specific workforce-development goals. Read the solicitation.

NSF explicitly asks reviewers to consider both criteria in every proposal and to provide a single overall recommendation. The relative weighting of the two criteria is left to the reviewer's judgment.

ERC-style criteria (Frontier Research Grants)

ERC panels are organized by domain (Life Sciences, Physical Sciences and Engineering, Social Sciences and Humanities) and by panel; reviewers are asked to assess both the PI and the project.

Criterion	What to assess
Ground-breaking nature and ambition	Does the proposed research aim to deliver significant advances at the frontiers of the discipline? Does it challenge current paradigms?
Scientific approach	Is the methodology — including the combination of approaches — appropriate to achieve the goals? Is the work feasible within the funding period? Is the PI's track record commensurate with the proposed work?
Principal Investigator	For Starting, Consolidator, and Advanced Grants: does the PI have the intellectual independence, the profile, and the leadership to drive the project? (For Synergy Grants: the team as a whole.)
Research environment	Does the host institution provide the resources, infrastructure, and intellectual environment to enable the project?

4. Section-by-section evaluation

Beyond the formal criteria, walk through the proposal as the applicant wrote it.

Specific Aims / Project Summary / Abstract

Is the central hypothesis clearly stated?
Is there a defensible "why this, why now" rationale?
Are the aims testable, distinct, and not a methodological laundry list?
Is the abstract free of jargon that obscures the goal to a non-specialist reader?

Background and Significance

Is the prior literature presented in a balanced way (contrary findings cited, not just supporting ones)?
Is the "gap" real, or is it a manufactured framing of well-trodden territory?
Is the potential impact articulated in concrete terms (who benefits, what changes, by when)?

Preliminary Data / Progress Report

For new proposals: are the preliminary data sufficient to justify the proposed aims?
For renewals: has the prior funding period been productive in a way that warrants continued support?
For resubmissions: has the applicant addressed the prior critiques substantively?

Approach / Research Plan / Methodology

For each aim: is the experimental design sound? Are controls identified? Is the sample size justified (formal power analysis, prior published data, or a pre-specified stopping rule)?
Are biological / technical replicates and statistical tests named with version?
Are potential pitfalls and alternative approaches identified?
For work involving human participants, vertebrate animals, or hazardous materials: are the relevant approvals or planned submissions described?
For computational / data-intensive work: are software, data, parameters, and reproducibility addressed?
For clinical / translational work: is the trial or study registered or planned for registration? Are inclusion / exclusion criteria and recruitment plans realistic?

Investigator(s) and Team

Are the roles of co-investigators, collaborators, and consultants clearly delineated?
For multiple-PI applications: is the leadership plan credible?
For early-career investigators: is the mentoring plan and institutional commitment adequate?

Environment and Resources

Are the facilities, equipment, and core support adequate for the proposed work?
Are any dependencies (e.g., a collaborator's lab, a shared instrument, a patient cohort) clearly committed?

Budget and Timeline

Is the budget commensurate with the scope of the work? Are any line items unusually large for the proposed activities?
Is the timeline realistic given the aims, with milestones or go/no-go decision points?
For modular budgets, is the total consistent with the funder's caps?
Are subcontracts to other institutions justified and reasonable?

Data Management, Sharing, Ethics

Are data-sharing plans consistent with the funder's policy (NIH DMS Policy, NSF DMP, ERC requirements)?
For research with human participants or vertebrate animals, are protections and approvals addressed?
Are conflicts of interest, foreign-component rules, or other compliance items handled per the funder's terms?

References, Biosketches, Letters of Support

Are key recent papers cited, including contrary findings?
Are biosketches current, in the funder's required format, and accurate?
Are letters of support specific, or are they generic boilerplate that does not commit to real resources?

5. Feasibility, risk, and realism checks

Is the timeline realistic, or is it over-ambitious for the team size and scope?
Are the proposed aims dependent on a single risky method, with no fallback?
Is the budget front-loaded in a way that signals over-commitment?
Are the "expected outcomes" described in vague generalities, or are they testable predictions?

6. Tone and ethics checks

Constructive framing: every criticism is paired with a suggestion or a question.
No personal attacks, sarcasm, or pejorative adjectives about the applicant or institution.
No "the applicant should have done my own work" or out-of-scope experiments.
Do not reveal your identity in single-blind reviews via language that could only come from a specific lab.
Do not share proposal content, panel deliberations, or reviewer identities outside the panel.

7. Finalize score and recommendation

Translate the criterion-by-criterion evaluation into a single overall-impact score, using the funder's published scale.

NIH 9-point scale (1 = exceptional, 9 = poor). Criterion scores on the same 1–9 scale feed into the overall impact; the final overall-impact score is a holistic judgment, not a formula.
NSF categories: Excellent / Very Good / Good / Fair / Poor (with optional numeric anchors in some panels).
ERC: scored 1–5 (or the panel's published scale) for each sub-criterion plus an overall score, with the published scoring descriptors (e.g., "fully meets the criterion" through "does not meet the criterion").
Private foundations: typically use a categorical scale (Highly Recommended, Recommended, Not Recommended) defined in the RFP.

Write a defensible rationale for the overall score that synthesizes the criterion scores and any feasibility, budget, or environment concerns. The score is a recommendation to the funder's council or board, not the funding decision itself.

Code patterns

Standard written-critique skeleton (NIH-style, R01)

================================================================================
REVIEW OF [Application ID / PI / Project Title]
================================================================================

1. Overall impact (and summary)
---------------------------------
[2-4 sentence synopsis. State the central question, the approach,
 the team, and your overall assessment. Provide the overall-impact score
 and a one-sentence rationale for it.]

2. Criterion scores
-------------------
Significance:           [1-9]
Investigator(s):        [1-9]
Innovation:            [1-9]
Approach:              [1-9]
Environment:           [1-9]
Overall Impact:        [1-9]

3. Strengths
------------
- [Specific strength tied to a criterion. Cite aim, page, or section.]
- ...

4. Weaknesses
-------------
- [Specific weakness, why it matters, what a fix would look like. Cite
   page or aim.]
- ...

5. Detailed comments on the approach
-------------------------------------
[Aim-by-aim walk-through. For each aim, summarize the design, identify
 specific concerns (controls, sample size, alternative interpretations),
 and note feasibility.]

6. Budget and timeline
----------------------
[Comments on budget reasonableness, timeline realism, and any
 specific line items that seem over- or under-funded.]

7. Protections and compliance
------------------------------
[Human subjects, vertebrate animals, biohazards, data sharing,
 clinical-trial registration, etc.]

8. Overall recommendation
-------------------------
[Defensible narrative tying criterion scores and strengths/weaknesses
 to the overall-impact score. Note whether the proposal is a
 resubmission and whether prior critiques have been addressed.]

NSF-style critique skeleton

================================================================================
REVIEW OF [Proposal ID / PI / Title]
================================================================================

1. Summary
----------
[2-4 sentence synopsis in your own words. State the question, the
 approach, the team, and the broader-impact activities.]

2. Intellectual merit
---------------------
Strengths:
- [...]
Weaknesses:
- [...]

3. Broader impacts
------------------
Strengths:
- [...]
Weaknesses:
- [...]

4. Summary statement
--------------------
[Synthesis explaining the overall recommendation. Reference both
 intellectual merit and broader impacts.]

5. Overall recommendation
-------------------------
[Excellent / Very Good / Good / Fair / Poor]

ERC-style critique skeleton

================================================================================
REVIEW OF [Proposal ID / PI / Title]
================================================================================

1. Summary
----------
[Brief synopsis of project, PI, and host institution.]

2. Ground-breaking nature and ambition
---------------------------------------
[Comments on the project's potential for significant advance at the
 frontier of the discipline.]

3. Scientific approach
----------------------
[Comments on methodology, feasibility, timeline.]

4. Principal Investigator
-------------------------
[Comments on intellectual independence, track record, leadership.]

5. Research environment
-----------------------
[Comments on host institution, resources, infrastructure.]

6. Overall assessment
---------------------
[Synthesis and overall score on the panel's published scale.]

Reusable criterion-by-criterion rubric

Section	Look for	Common failure mode
Specific Aims	Testable, distinct, ambitious but realistic	Three aims that all depend on Aim 1; "fishing expedition" framing
Significance	Concrete impact; real gap	Restates prior work as if it were a gap
Approach	Adequate controls; sample-size justification; pre-specified analysis plan	"We will determine the optimal X" with no definition of optimal
Investigators	Track record matched to the work; clear roles	Senior PI with no time committed; junior PI with no mentoring plan
Environment	Facilities and resources committed, not aspirational	"Access to the core facility" with no letter
Budget	Commensurate with scope; milestones named	Five-year project with no go/no-go decisions
Data / Ethics	Sharing plan; registrations; approvals	"Data will be made available upon request"

Red-flag phrases to surface to the panel chair (confidential section)

Substantive overlap with another pending or active application from the same lab (potential duplicate submission).
Concerns about scientific integrity, data fabrication, or authorship dispute in cited prior work.
Conflicts of interest not declared by the applicant.
Budget items that are inappropriate for the mechanism (e.g., capital equipment in an operating grant).
Commitments to resources or patients that are not substantiated by a letter of support.
Foreign-component issues (subawards, significant performance outside the applicant institution) not addressed in the application.

Common pitfalls

Reviewing against your own research program: a request to add a control experiment that the applicant's mechanism cannot fund, or to compare against a method that requires unavailable expertise, is an unfair ask.
Scoring the writing instead of the science: a poorly written but methodologically sound proposal is not a poor proposal; a beautifully written proposal with fatal methodological flaws is still fatally flawed.
Conflating innovation with significance: an incremental advance on a high-impact question can score very well; a technically novel approach to a marginal question may not.
Drifting on the budget: review the budget, but recognize that panel discussions, not the written critique, are where major budget changes are negotiated.
Contradicting your score and your text: if your written critique identifies fatal flaws, your overall-impact score should reflect them; alignment matters.
Revealing identity in single-blind reviews: even indirect clues (citing your own work, naming a reagent only your lab makes) can unblind you.
Sharing proposal content with trainees, postdocs, or colleagues: this is a confidentiality breach and a funder-reportable issue.
Missing COI declaration: many panels ask for COI declaration both before and after reading; failing to declare late-discovered conflicts can result in the review being excluded.
Underweighting environment and feasibility: a brilliant idea from a poorly resourced team at a non-supportive institution is not the same as the same idea from a well-supported team.

Validation

A good written critique is itself a small scientific document. Before submitting, check:

Open alternatives

Closed grant-review platforms (NIH eRA Commons, NSF FastLane / Research.gov, ERC SEP) are the primary funder systems; their open-source alternatives are limited and not generally available to applicants.
Closed reference managers (EndNote) can be replaced with open tools (Zotero with Better BibTeX, JabRef) for managing the cited literature during your review.
ORCID (open researcher and contributor ID) is the de facto open standard for researcher identification and is increasingly required by funders; many journal and grant systems integrate with it.
Preprint servers (bioRxiv, arXiv, ChemRxiv, SSRN) allow reviewers to consult recent, not-yet-formally-published work that may be relevant to a proposal's preliminary data.
Open Science Framework (OSF) and Zenodo are the de facto open registries for data, code, and pre-registration linked to grant work.
AI assistants for prose polish are useful for tightening the critique, but the reviewer's judgment remains the deciding factor; do not paste unpublished proposal text into a third-party LLM without funder approval.

References

Internal cross-links

ors-peer-review-manuscript-review — analogous review workflow for journal manuscripts.
ors-peer-review-scholar-evaluation — for assessing an applicant's overall record, not a single proposal.
ors-peer-review-reviewer-response — for the applicant's reply to reviewer critiques on a revised proposal.
ors-scientific-writing-grant-proposal-structure — the applicant's side of the same process.
ors-ethics-compliance-irb — human-subjects and animal-use protections.
ors-open-science-data-sharing — data and code deposition expectations linked to grant work.

External links (verifiable, public)

NIH — Peer Review Process for Research Grant Applications: https://grants.nih.gov/grants/peer-review-process.htm
NIH — Guidance for Reviewers: https://grants.nih.gov/grants/peer-review-guidance.htm
NIH — Review Criteria for Research Grant Applications: https://grants.nih.gov/grants/peer/critiques/rpg.htm
NSF — Proposal & Award Policies & Procedures Guide (PAPPG): https://www.nsf.gov/publications/pub_summ.jsp?ods_key=pappg
NSF — Merit Review Principles and Criteria: https://www.nsf.gov/bfa/dias/policy/merit_review/
ERC — Evaluation Panels and Peer Review: https://erc.europa.eu/evaluation-panels-and-peer-review
ERC — Work Programme (frontier research grant criteria): https://erc.europa.eu/document-category/work-programmes
ESF — Reviewers' Guide: https://www.esf.org/
AAAS — Reviewer resources: https://www.aaas.org/
COPE — Conflict of Interest framework: https://publicationethics.org/resources/guidelines-0/conflict-interest
ORCID: https://orcid.org/

Changelog

1.0.0 (2026-06-10): Initial adaptation by Pradyumna Jayaram. Heavy rewrite of upstream grant-review skill: added funder-specific criterion tables (NIH, NSF, ERC), section-by-section rubric, three written-critique skeletons, budget/feasibility/environment checks, late-discovered conflict disclosure notes, and explicit "do not paste proposal text into third-party LLMs" guardrail.

skills/peer-review/grant-review