The email arrives, you skim the decision letter, and your stomach drops: major revisions. Three reviewers, forty comments, and at least one of them appears not to have read your methods section at all. This is the moment that decides more publication outcomes than any experiment — and it is almost entirely about temperament and structure, not science.
What should you do first?
Nothing, for forty-eight hours. Nearly every combative, career-hurting response we have seen was drafted on the same day the decision arrived, and none of them improved the paper.
Read the letter once, close it, and step away. When you return, the comments will look less like an attack and more like what they usually are: a to-do list for acceptance.
How should you structure the response letter?
As a three-column table: the reviewer's comment quoted in full, your response, and the exact change made with a page and line reference. Number every comment, including the trivial ones.
Editors read the response letter before they re-read the manuscript. A complete, orderly table signals that every concern was taken seriously; a wall of prose signals that some were not. Keep the reviewer's original wording rather than paraphrasing — paraphrase reads as evasion.
| Comment | Response | Change |
|---|---|---|
| R2.4: "The sample size is not justified." | Agreed. We have added a power calculation and the assumptions behind it. | p. 7, lines 142–151 |
| R1.2: "Figure 3 is unreadable at print size." | Agreed. Redrawn at 600 dpi with larger axis labels. | Figure 3, p. 12 |
Should you agree with every comment?
No — but you must respond to every comment, and you should agree with most of them. Reviewers are experienced readers seeing your work fresh, and the majority of their comments identify something a reader will genuinely stumble over.
Reserve disagreement for points where making the change would make the paper wrong. Then the formula is: acknowledge the concern, state your position in one sentence, and support it with data or citations.
What if a reviewer is factually wrong?
Correct them plainly, without commentary on their competence, and treat the error as evidence that your text was unclear. A reviewer who misread your methods is telling you that readers will misread them too.
A response that works: "Reviewer 2 notes that the control group was not blinded. Blinding is described on p. 5, lines 88–92. We accept that its placement made it easy to miss and have moved it into the study design subsection." You have corrected the record and improved the paper in two sentences.
What if two reviewers contradict each other?
Say so explicitly, choose one course, and explain the choice to the editor — this is exactly the situation a handling editor exists to adjudicate. Do not silently follow one and ignore the other.
Write it out: "Reviewer 1 asks us to expand the discussion of limitations; Reviewer 3 asks us to shorten the discussion overall. We have expanded the limitations by one paragraph and cut the same length from the survey of prior work, which we judged the less essential of the two."
How long should you take over revisions?
Use the deadline the journal gives you, and ask for an extension before it expires if you need one — editors grant these routinely and think nothing of it. Missing the deadline silently is treated as a withdrawal by some journals.
How should the letter end?
By thanking the reviewers for specific improvements their comments produced. Not flattery — a record that the process worked, which editors notice.
A rebuttal handled this way turns a major revision into what it statistically is: the most common road to acceptance. If you would like a second reader on a response letter before it goes back, that is part of what we do.
