How to Respond to Reviewer Comments

First, read what the decision actually says

Before reacting to the reviews, find the editor's decision in the letter. The four common outcomes mean very different things:

• Minor revision — a limited set of changes. Often re-checked by the editor alone rather than sent back out for another round of review.
• Major revision — substantial rewriting, missing detail, or additional analysis. The revised paper is likely to go back to the original reviewers, so they will read your response letter with your first version fresh in mind.
• Reject and resubmit — the current submission is closed. A resubmission is usually processed as a new manuscript with a new ID and no commitment to acceptance, though the editor is signalling that a much-changed paper would be considered.
• Reject — the journal will not reconsider it. Your next move is a different journal, not a rebuttal.

The editor's own remarks outrank the reviewers'. If the editor highlights two of six comments as essential, those two decide your paper.

Sleep on it, then build the comment list

Read the reviews, close the file, and come back a day later. Then do something mechanical: paste every comment — editor first, then each reviewer — into a fresh document and number them. Long paragraphs that raise three separate issues get split into three numbered items, so nothing is quietly skipped.

That numbered list is now your to-do list and the skeleton of your response letter. Work through the manuscript first, then write the letter from what you actually changed.

The structure of a response letter

The convention is a standalone document that reproduces each comment verbatim, followed by your reply. Don't paraphrase the reviewer — quoting exactly lets the editor match your response to the review without cross-referencing.

Reviewer 2, Comment 3. "The sample is described as representative, but no response rate is reported."

Response. Thank you — this was missing. We have added the response rate and the non-response comparison to Section 3.1, and softened the claim of representativeness accordingly.

Change (p. 8, lines 214–219). "Of 1,240 invitations, 383 questionnaires were returned (30.9%). Respondents did not differ significantly from non-respondents by age or region…"

Three moving parts, every time: the comment, what you did, and where to find it — with the new text quoted so the reviewer never has to open the manuscript to check.

Three responses you will write over and over

1. You agree and changed it. Keep it short. Thanks, what changed, where. Padding a simple fix with three sentences of gratitude makes the letter harder to audit.

2. You partly agree, or solved it differently. Say so explicitly rather than letting the editor discover it:

We agree the mechanism needs more support, but a mediation analysis is not possible with cross-sectional data. We have instead added the two robustness checks in Section 4.3 and flagged the causal limitation explicitly in the Discussion (p. 17).

3. You disagree. Legitimate, as long as you argue from evidence and not from irritation. Give the reasoning, cite support, and still concede something — usually a clarification, a limitation, or a sentence that prevents the next reader from making the same reading:

We have retained the original coding scheme. Recoding to a binary measure would discard the graded responses the instrument was designed to capture (Author, 2021). To prevent this being read as an oversight, we now justify the choice at the start of Section 3.2 and report the binary specification in Appendix B.

Note what that last one does: it treats the reviewer as a consultant rather than an opponent, and it converts a refusal into a visible improvement.

When the reviewers contradict each other

It happens often, and it is not your job to reconcile it privately. Lay both comments out, explain the route you chose and why, and hand the decision to the editor: "Reviewer 1 asks us to expand the theory section; Reviewer 3 asks us to shorten it. We have kept the expansion but moved the historical background to an appendix, and we are happy to follow the editor's preference." Editors arbitrate; they would rather be asked than discover the conflict after the next round. Correspondence goes through the editorial office — don't contact reviewers directly.

What you actually upload

Most journals expect three files:

• a clean copy of the revised manuscript, changes incorporated, no markup;
• a tracked-changes or highlighted version showing what moved;
• the point-by-point response letter.

Requirements vary between publishers, so re-read the decision letter and the journal's author guidelines before uploading. The deadline is in that letter too — if you need more time, ask the handling editor before it passes rather than after.

The mistakes that cost you a round

• Silent non-compliance. Claiming a change you didn't make, or ignoring a comment, is the fastest way to a second major revision.
• No locations. A response with no page or section numbers forces the reviewer to hunt, and hunting produces new objections.
• Arguing tone. "The reviewer clearly did not read Section 2" loses you a reader you need.
• Unverified new citations. References added in the last hours before resubmission are where fabricated or mis-attributed sources creep in — particularly if AI helped draft the additions. AI models invent plausible-looking references, and a reviewer who already knows the field will notice.

Frequently asked questions

Keep reading

• How to Write a Literature Review Without Fabricating Sources
• Why ChatGPT Makes Up Citations
• In-Text Citations Explained: APA, MLA, Chicago, Harvard & IEEE
• Free Citation Generator — All Styles