Peer Review Checklist

---
name: peer-review-checklist
description: Conducting rigorous peer review — structured evaluation, major vs minor issues, and writing reviews authors can act on.
category: scientific
---

## Overview

Peer review is quality control for science: the reviewer's job is to
judge whether the claims are supported and the work is sound — not to
rewrite the paper or gatekeep the field. This skill covers structuring
a review (summary, major issues, minor issues), evaluating novelty,
methods, statistics, and interpretation separately, and writing criticism
that improves the paper.

## When to use

- Reviewing a manuscript for a journal or conference
- Deciding on a recommendation: accept, minor/major revision, reject
- Separating fatal flaws from fixable issues from nits
- Writing reviews that editors trust and authors can actually use
- Training students in how to review

## Core concepts

- **The reviewer's question:** are the methods sound, do the results support the conclusions, and is it a sufficient advance? — in that order. Soundness first; novelty is secondary.
- **Major vs minor:** major = affects the conclusions (flawed design, wrong analysis, unsupported claims, missing controls); minor = doesn't (typos, figure clarity, extra citations, phrasing). Recommendations hinge on majors.
- **The summary test:** open with a 3–5 sentence summary of what the paper claims — if you can't, the paper is unclear (that's a major issue); the summary also proves to the editor you read it.
- **Evidence over taste:** "I would have done it differently" is not a criticism unless the authors' approach is wrong — distinguish preference from error.
- **Confidentiality and conflicts:** don't use the manuscript's ideas, don't delay to scoop, decline when conflicted or unqualified — the ethical infrastructure review depends on.
- **Tone:** critical but collegial — attack the work, never the author; specific and actionable beats clever and dismissive.

- **The "would I cite this" test:** after reading, ask whether you'd confidently cite the main claim — if not, articulate exactly what would make you confident; that's your major-issue list.
- **Reviewing interdisciplinary work:** mark clearly which parts are inside vs outside your expertise — editors need to know your coverage limits to assign complementary reviewers.
- **Post-publication review:** the same critical apparatus applies to preprints and published papers — public, signed post-publication review is increasingly the venue where errors get caught.

## Practical workflow

### 1. Read strategically (three passes)

1. **Pass 1 (15 min):** abstract, figures, conclusions — what's claimed? Is it interesting and plausible? Form initial questions.
2. **Pass 2 (1–2 h):** full read — methods detail, statistics, logic chain. Check every claim against its evidence; note where methods don't support results.
3. **Pass 3 (targeted):** verify the critical analyses — redo key calculations if feasible, check references for the load-bearing citations, inspect supplementary methods.

### 2. Evaluate by dimension

- **Novelty/significance:** real advance or incremental? (Be honest about your field's standards, not your pet interests.)
- **Methods:** appropriate design? Controls present? Sample sizes justified? Reproducible from the description?
- **Analysis:** correct statistics? Assumptions checked? Multiple comparisons handled? Effect sizes, not just p-values?
- **Interpretation:** do conclusions follow from results? Alternatives considered? Limitations stated? Claims in abstract match the body?
- **Presentation:** figures clear and honest? Writing comprehensible? References appropriate (not padded, not missing key prior work)?

### 3. Write the review

1. **Summary:** what was done, what was found, your overall assessment — shows comprehension and frames the critique.
2. **Major issues:** numbered, each with the problem, why it matters, and what would fix it ("the conclusion that X requires [control/analysis]; without it, consider weakening to Y").
3. **Minor issues:** typos, clarifications, figure suggestions, missing citations — grouped, not numbered with majors.
4. **Recommendation:** accept / minor / major / reject — aligned with the issues listed; "major revision" means you believe the flaws are fixable.

### 4. Calibrate

1. Ask: if the majors were fixed, would I accept? If no, the recommendation should probably be reject, not endless major revisions.
2. Check proportionality: a solid paper with rough edges deserves minor revision, not punishment for not being your ideal paper.
3. Disclose limits: "I could not fully evaluate the [specialized method] section" — editors need to know your coverage.

### 5. Deliver the review and follow up

1. Submit on time — late reviews delay science; if you can't, decline early so the editor can reassign.
2. Keep a copy of your review; when the revised manuscript returns, check whether your major issues were actually addressed (not just responded to).
3. Update your recommendation honestly on re-review — authors who did the work deserve the upgrade; authors who dodged deserve the same verdict.

### 6. Quick-reference checklist

- [ ] "Would I cite this claim?" test applied to the main result
- [ ] Expertise boundaries stated to the editor
- [ ] Major vs minor issues separated clearly
- [ ] Every criticism paired with an actionable suggestion
- [ ] Tone is professional; no ad hominem or contempt
- [ ] Conflicts of interest declared
- [ ] Review submitted on time (or declined early)
- [ ] On re-review: major issues checked as addressed, not just answered

## Common pitfalls

- **Reviewing the paper you wish they'd written:** demanding different experiments that aren't needed to support the actual claims.
- **Major/minor confusion:** listing 30 nits as major issues — dilutes the real problems and exhausts the authors.
- **Vague criticism:** "the statistics are questionable" without saying which analysis and why — unactionable; cite the specific test and the violated assumption.
- **Novelty gatekeeping:** rejecting sound work for "insufficient novelty" when the journal's bar is lower — apply the venue's standard, not Nature's.
- **Adversarial tone:** sarcasm and contempt — unprofessional and makes authors defensive rather than responsive.
- **Confidentiality breaches:** discussing the manuscript publicly or using its ideas — the fastest way to end a reviewing career.
- **Reviewing to display expertise:** demanding citations to your own work or unrelated methodological purity — review the paper's claims, not your CV.
- **Confidentiality lapses:** discussing the manuscript's unpublished ideas with colleagues or using them — the ethical bright line; never cross it.