How to Respond to Peer Reviewers: Structure, Tone and Handling Disagreement

Decision letters use a small vocabulary with important differences. Minor revision usually means the editor expects to accept the paper once specific points are fixed, often without sending it back to reviewers. Major revision means the paper could become acceptable but needs substantial work, and it will usually return to the same reviewers. Reject and resubmit, used by some journals, means the paper will be treated as a new submission. Rejection with an invitation to submit elsewhere, or transfer to a sister journal, is a different route altogether.

Pay attention to the editor's own comments. Editors often signal which reviewer points they consider essential, which they regard as optional, and whether a requested analysis is a condition of acceptance. When reviewers disagree with each other, the editor's steer is the tie-breaker.

Note the deadline and whether the journal allows extensions. If a requested analysis or new data collection will take longer than the deadline allows, write to the editor early rather than submitting a rushed revision.

Almost everyone finds critical reviews frustrating on first reading. It is worth waiting a day before drafting anything, then rereading with the assumption that each comment points to something that confused or unconvinced at least one careful reader. Even a comment based on a misunderstanding tells you that the manuscript allowed the misunderstanding.

Next, copy every comment into a working document and number them, splitting long paragraphs into separate points. Then classify each one. A simple scheme works well: clarification or presentation changes, additional analyses or data, changes to framing or literature, points you intend to contest, and points outside the paper's scope. Estimate the effort for each and identify dependencies, for example a new analysis that will change the discussion.

Journals differ in format, and some provide a template or a text box in the submission system, but the most widely used structure is a point-by-point response. It begins with a short cover note to the editor, then reproduces each reviewer comment in full, followed immediately by your response and a precise description of what changed and where.

Make changes easy to verify. Give page and line numbers for the revised manuscript, quote the new or changed text, and submit a tracked-changes or highlighted version if the journal asks for one. Editors and reviewers are busy; a response that says only "We have addressed this" forces them to hunt through the manuscript and tends to irritate.

Use visual distinction between the reviewer's words and yours, for example italic or indented reviewer comments and plain-text responses. Keep numbering consistent with the reviewer's own points. If one change answers several comments, answer each one briefly and cross-refer rather than skipping any.

• Opening note to the editor: thank the editor and reviewers, summarize the main changes in three to six bullets, and mention any points on which you respectfully disagree.
• Reviewer 1, Comment 1.1: the full comment as written.
• Response: what you did and why, in one to three short paragraphs.
• Change in manuscript: section, page and line numbers, plus the revised text quoted.
• Repeat for every comment from every reviewer, including those that required no change.

The right tone is professional and appreciative without being servile. A single sentence of thanks at the start of each reviewer's section is enough; thanking reviewers again in every answer adds length without adding substance. Focus the response on the evidence and on what changed.

Avoid defensive phrasing such as "As we clearly stated" or "The reviewer has misunderstood". Reframe these situations around your own text: "We have revised the paragraph to make the sampling frame explicit, since the original wording allowed two readings." This concedes nothing about the science while acknowledging a communication problem you can fix.

Be precise about what you did. "We added a robustness check excluding the 2020 observations; results are reported in Supplementary Table S4 and the main estimates are unchanged in direction and significance" is far more persuasive than "We performed additional analyses as requested."

You do not have to accept every suggestion, and editors expect some disagreement. What they do not accept is silence or dismissiveness. When you disagree, acknowledge the concern, explain your reasoning with evidence or references to the methodological literature, and, where possible, show that you have taken the concern seriously in some way, for instance by adding a sensitivity analysis, a limitation, or a sentence explaining the choice in the manuscript.

A useful pattern is: acknowledge, explain, act. "We appreciate the suggestion to use a random-effects model. We retained the fixed-effects specification because the Hausman test rejects the random-effects assumption (reported in the new footnote 7). We have added a sentence to the Methods explaining this choice."

When two reviewers ask for opposite things, say so openly and explain which option you chose and why, referring to the editor's guidance if there was any. When a request would change the paper into a different study, explain the scope decision and point to the limitation or future-work paragraph where you address it.

The difference between an adequate and an excellent response is usually specificity. The invented examples below respond to the same reviewer comment: "The sample size seems small for the number of predictors in the regression model."

The strongest version concedes the legitimate concern, adds evidence, and changes the manuscript so future readers see the answer, not only the reviewer. It does not apologize excessively or argue with the reviewer's motives.

With several co-authors, revisions can stall because nobody owns particular comments. Assign each numbered comment to a named author with a draft deadline, keep one master manuscript file to avoid conflicting versions, and designate a corresponding author to assemble the final response in a single voice.

Agree before drafting on points of disagreement with reviewers, since co-authors may hold different views. All authors should approve the final revised manuscript and response, as most journals require, and any change in authorship during revision needs explanation and signatures according to journal policy.

• Keep a tracking table: comment number, owner, status, location of change.
• Freeze the manuscript before generating final line numbers for the response.
• Run a final consistency check on abstract, results and discussion after new analyses.
• Check that figures, tables and supplementary files are updated and renumbered.

Most avoidable delays come from incompleteness or from a mismatch between what the letter says and what the manuscript shows. A careful final comparison between the response and the revised file catches most of them.

• Skipping or merging comments so that some points receive no explicit answer.
• Claiming a change that was not actually made, or giving line numbers from the old version.
• Adding new analyses to the response letter but not to the manuscript, so future readers never see them.
• Over-long responses that bury the key change in a page of explanation.
• Introducing new unrequested material that raises new questions without flagging it to the editor.
• Letting the word count or reference limit drift beyond the journal's maximum during revision.
• Changing the author list or data without the journal's required process and explanations.

Peer review relies on confidentiality. COPE's ethical guidelines for peer reviewers ask reviewers not to disclose or use information from manuscripts under review, and authors should treat reviewer reports as confidential correspondence too, not posting them publicly without the journal's permission. If you suspect a reviewer has a conflict of interest or is asking you to cite their own work inappropriately, raise it politely with the editor rather than in the response letter.

If you receive help drafting a response, whether from a colleague, an editor or an AI tool, remain responsible for every scientific claim, check journal policy on declaring assistance, and never state that you ran an analysis you did not run.

The short note that opens the response package is often read first and sometimes read most carefully, because it tells the editor whether the revision is substantial and whether any points remain contested. Keep it to roughly half a page. Thank the editor for handling the manuscript, summarize the main changes in a short list, and flag, neutrally, any requests you did not follow and why.

If the revision changed something significant, such as the main analysis, the title, the sample or the author list, state it explicitly here rather than letting the editor discover it. Mention any new supplementary material and confirm that all authors have approved the revision. If the journal asks for a clean version, a tracked-changes version and the response as separate files, name them clearly so the editor can find each one quickly.

Avoid using the cover note to argue the paper's merits again or to criticize reviewers. A calm, organized note sets the tone for the detailed responses that follow and makes the editor's job easier, which is always in your interest.

A rejection after review still gives you valuable, free expert feedback. Before sending the paper elsewhere, address the comments that would apply at any journal, such as unclear methods, missing analyses or weak framing, even though the next journal will not see the previous reports. Many papers are reviewed by the same people at a second journal, and an unchanged manuscript can create a poor impression.

Reconsider fit as well as quality. A rejection that says the paper is sound but of limited interest to the readership points to a scope or audience mismatch, which is a journal-selection problem rather than a writing problem. Some publishers offer transfer services that carry reviews to a partner journal; weigh the time saved against whether that journal really suits the work. Appeals are appropriate only when you can show a clear factual error in the review process, not simply because you disagree with the decision.