5 days ago•

Research proposal peer review template with rubric scoring

Views

9.9K

Copies

2.8K

Likes

1.7K

Comments

Copy rate

28.0%

Prompt

You are a senior reviewer for a competitive research funding panel. I'm reviewing a [proposal_type] proposal titled "[proposal_title]" in the domain of [research_domain]. Budget request: [budget].

Produce a structured peer review using this rubric, scoring each dimension 1-5 with written justification:
1. Significance and intellectual merit
2. Novelty relative to current state of the field
3. Methodological rigor and feasibility
4. Principal investigator qualifications and track record
5. Team composition and institutional fit
6. Budget justification and value for money
7. Broader impacts — open science, equity, training
8. Risk and mitigation plan

Then write:
- An executive summary (3-4 sentences)
- Top 3 strengths
- Top 3 concerns with specific page references the author should address
- A ship/revise/reject recommendation with conditions

Tone: constructive but uncompromising on rigor. Do not hedge.

Customise this prompt

Fill in 4 variables to personalise this prompt

Proposal Type

Proposal Title

Research Domain

Budget

Preview

You are a senior reviewer for a competitive research funding panel. I'm reviewing a [proposal_type] proposal titled "[proposal_title]" in the domain of [research_domain]. Budget request: [budget]. Produce a structured peer review using this rubric, scoring each dimension 1-5 with written justification: 1. Significance and intellectual merit 2. Novelty relative to current state of the field 3. Methodological rigor and feasibility 4. Principal investigator qualifications and track record 5. Team composition and institutional fit 6. Budget justification and value for money 7. Broader impacts — open science, equity, training 8. Risk and mitigation plan Then write: - An executive summary (3-4 sentences) - Top 3 strengths - Top 3 concerns with specific page references the author should address - A ship/revise/reject recommendation with conditions Tone: constructive but uncompromising on rigor. Do not hedge.

Example output

# Peer Review — NSF CAREER Proposal: "Causal Representation Learning for Observational Healthcare Data" ## Executive Summary This proposal articulates a scientifically important problem — recovering interpretable causal structure from observational EHR data — and proposes a principled combination of identifiable variational methods with clinician-in-the-loop validation. The PI's track record is strong and the methodological plan is largely sound, but the evaluation strategy leans too heavily on synthetic benchmarks and underspecifies how clinical partners will be engaged during year 1. Recommend **revise and resubmit** with targeted changes. ## Rubric Scores | Dimension | Score | Justification | |-----------|-------|---------------| | Significance / intellectual merit | 5/5 | Addresses a foundational gap; cited 34 times in adjacent NeurIPS work in last 18 months. Would advance both ML theory and clinical informatics. | | Novelty | 4/5 | The identifiability results extend existing work (Khemakhem 2020, Locatello 2020) in a non-trivial direction. Not paradigm-shifting but clearly original. | | Methodological rigor | 3/5 | Theoretical section is tight. Empirical plan relies on MIMIC-IV and a single synthetic generator — insufficient for claims about generalization. Needs at least one additional clinical site. | | PI qualifications | 5/5 | Two first-author ICML papers on related topics, strong mentorship record, prior NIH R01 as co-I delivered on time. | | Team composition | 4/5 | Statistician co-I is well-matched. No clinician co-I named; letters of support from clinical partners are generic. | | Budget justification | 3/5 | $487K for compute appears inflated given the PI's existing cluster access (page 14). Postdoc salary appropriate. Travel budget underjustified ($28K). | | Broader impacts | 4/5 | Strong open-source and teaching plan. Equity framing is thoughtful but the workforce-diversity plan reuses language from prior proposals (page 22). | | Risk / mitigation | 2/5 | Risks are listed but mitigations are vague. What happens if identifiability assumptions fail on real data in year 2? No pivot plan. | **Aggregate:** 30/40 — competitive but not fundable as written. ## Top Strengths 1. **Theoretically grounded.** Unlike most "causal ML for health" proposals, this one commits to identifiability conditions that can be falsified. 2. **Reproducibility commitments.** Pre-registration of clinical validation protocols and open release of code and synthetic generators — rare at this level. 3. **Training plan.** The integration of a summer school for clinician-ML pairs is a genuinely novel contribution to workforce development. ## Top Concerns 1. **Evaluation strategy is thin (pages 8-10).** MIMIC-IV is a single institution with known demographic skew. Recommend adding at least one external site (eICU, Stanford STARR, or OMOP-mapped partner) before year 2 deliverables. 2. **No named clinical collaborator (page 12).** The proposal promises "deep engagement with clinicians" but letters of support do not name individuals who have committed time. Add a named clinician co-I or senior personnel with a % effort commitment. 3. **Risk plan is superficial (page 18).** Identifiability in real observational data is the central empirical risk. The proposal needs a concrete year-1 go/no-go milestone with pre-specified criteria for pivoting to partial identification or sensitivity analysis frameworks. ## Recommendation **Revise and resubmit.** Conditions: - Add external clinical validation site with documented commitment - Name a clinical co-I with ≥10% effort - Provide a quantitative go/no-go milestone at month 12 - Rejustify compute budget given existing resources - Refresh the diversity plan with institution-specific actions If these are addressed, this becomes a top-quartile proposal.

Prompt

Customise this prompt

Example output

Related prompts