8.2 Peer Review, Rebuttals, Revisions, and Appeals
Key Takeaways
A point-by-point response should reproduce or identify each comment, state the action taken, cite revised locations, and explain respectfully when the authors decline or modify a request.
Reviewer requests do not override protocol, prespecification, participant protection, data integrity, authorship, or accurate reporting; label new analyses and their limitations.
Every substantive revision should be verified across the manuscript, abstract, tables, figures, supplement, registry, reporting checklist, and related assets as applicable.
Appeal grounds and deadlines are journal-specific; factual or procedural error can be a strong basis, but a reasoned disagreement may also be considered when the venue permits it.
“Open peer review” is an umbrella for several practices—open identities, published reports, open participation, or interaction—and does not guarantee that every identity and report is public.
Peer Review, Rebuttals, Revisions, and Appeals
Peer review is an editorial evaluation, not an instruction to satisfy every request at any cost. A strong revision answers the underlying scientific concern, protects the integrity of the analysis, and leaves a clear record of what changed.
Triage the decision letter
Read the editor's letter before individual reviews. The editor may prioritize comments, reject a reviewer request, set a deadline, or condition reconsideration on specific work. Build a comment log with owner, classification, response, manuscript location, dependencies, and approval status.
Classify comments as:
- editorial or presentation;
- factual clarification;
- reporting omission;
- statistical or methodological;
- request for new data or analysis;
- interpretation or tone;
- scope-changing request;
- ethical, privacy, authorship, or integrity concern;
- inconsistent or impossible request.
Similar comments can be grouped internally, but the response letter should make it easy for each reviewer and editor to find the answer.
Construct the response
A point-by-point response normally:
- thanks the editor and reviewers professionally;
- quotes, numbers, or clearly paraphrases each comment;
- states what the authors did;
- provides the revised text or a concise description;
- identifies page, paragraph, line, table, or figure in the clean revision;
- explains the scientific basis when a request is not followed exactly.
Do not call a reviewer careless, speculate about identity, or claim a change was made when it was not. “Done” is weaker than a precise answer. If tracked changes are requested, provide them along with the clean file and response.
Requests for additional analysis
A reviewer may identify a valuable analysis, but authors should first ask whether it is supported by the data, consent and permitted use, protocol and analysis plan, statistical assumptions, resources, and article scope.
For a post hoc subgroup request:
- explain whether the subgroup and interaction were prespecified;
- assess sample size, multiplicity, missingness, and biological rationale;
- do not infer a subgroup effect merely because one subgroup is significant and another is not;
- label an added analysis exploratory;
- report the interaction or other appropriate comparison when justified;
- avoid causal or confirmatory language beyond the design;
- decline the request with a reason when it would be invalid, misleading, disclosive, or impossible.
Supplementary placement does not cure a flawed analysis. Conversely, “not prespecified” does not automatically mean an analysis can never be informative; it changes its evidentiary status.
Protect consistency and authorship
A revision can alter the abstract, conclusions, tables, figures, plain-language summary, disclosure, data-sharing statement, and reporting checklist. Use a cross-document check after material changes. Verify values against current approved outputs and update the participant flow or registry when required.
Named authors should review material scientific revisions and approve the final version to be published. Do not let a coordinator infer approval from silence. New contributors may affect acknowledgments or authorship, and any byline change must follow the journal's process.
If a reviewer asks for identifying participant data, confidential sponsor information, or use outside consent or governance, explain the constraint and discuss a privacy-protective alternative with the editor. Peer review does not waive law, consent, contracts, or ethical oversight.
Conflicting or inappropriate comments
When reviewers conflict, describe the conflict and propose a resolution to the editor. Follow the editor's explicit direction if it is scientifically and ethically acceptable. If a comment contains an error, correct it with evidence rather than dismissing it.
A reviewer request to cite their own work should be assessed for relevance. Add genuinely relevant sources; explain why irrelevant citations were not added. Report suspected coercive citation, discriminatory language, confidentiality breach, or conflict through the journal's process.
Appeals
An appeal asks a journal to reconsider a decision; it is not a second routine rebuttal. Read the journal's grounds, deadline, word limit, and submission route. Strong appeals are concise and evidence based. They may identify a material factual misunderstanding, procedural failure, unmanaged conflict, overlooked evidence, or another ground the journal accepts.
Do not teach that factual error is the only ethical basis for appeal. Some journals allow a reasoned challenge to editorial or scientific assessment; others state that priority decisions are not appealable. The venue decides. Repeated lobbying, contacting reviewers, or submitting the same manuscript elsewhere while an appeal is active can violate policy.
If the team chooses not to appeal, revise for another venue. Do not submit simultaneously. Reformat and recheck the work rather than carrying forward an obsolete cover letter or an unsupported claim added during the first review.
Review models
Common labels are not perfectly standardized:
- Single-anonymized: reviewers usually know author identity; authors do not know reviewer identity.
- Double-anonymized: author and reviewer identities are concealed from each other during review to the extent feasible.
- Open peer review: may involve signed reviews, public reports, open identities, author–reviewer interaction, or open participation in different combinations.
- Post-publication review: commentary or formal review occurs after public release.
Check the journal's definition. Open reports do not necessarily reveal reviewer names, and signed review does not necessarily mean the full decision history is published.
Important
On exam scenarios, answer the concern rather than obeying mechanically: protect data integrity, label exploratory work, document every change, obtain author approval, and follow the journal's own appeal and review policy.
A reviewer requests an underpowered, non-prespecified biomarker subgroup analysis. What is the best response?
Run it and present it as confirmatory.
Ignore the comment without explanation.
State that reviewers may never request post hoc work.
Explain prespecification and power limits, perform a valid exploratory analysis only if scientifically justified and clearly labeled, or give a reasoned explanation for declining.
Which action is most likely to support a journal appeal?
Follow the venue’s appeal policy and provide concise evidence of a material factual, procedural, conflict, or other accepted error in the decision.
Email the anonymous reviewer directly.
Appeal every priority rejection as a matter of routine.
Submit the manuscript to another journal while representing that no appeal is pending, regardless of both policies.
What does “open peer review” mean?
It always means anonymous reviewers and confidential reports.
It is an umbrella for practices such as open identities, published reports, open interaction, or open participation; the journal must define which features it uses.
It guarantees acceptance after public comments.
It eliminates editorial decision-making.
Sections you finish are checked off in the contents.