How to Disagree With a Reviewer Politely (and Still Get Accepted).

Disagreeing with a reviewer politely comes down to three moves: concede the part of the comment that is true, contest the part that is not, and support your position with evidence rather than adjectives. Say plainly that you did not make the requested change, then give a reason the editor can verify, such as a citation or a reanalysis. Editors accept well-argued refusals all the time. What sinks revisions is defensiveness, not disagreement.

You are allowed to say no. Peer review is an argument between experts, not a compliance exercise, and the response to reviewers letter is where that argument happens. The rest of this post covers when to have the argument and how to win it without making an enemy of Reviewer 2.

When is disagreeing the right call?

Push back when the requested change would make the paper wrong, or when it demands work your study cannot support. Accept everything else.

Most comments deserve acceptance, including the annoying ones. A reviewer who catches a mislabeled axis or an overreaching claim in the abstract has done you a favor at someone else's expense. The comments worth contesting are rarer. A request to reframe your findings around a hypothesis you never tested. A demand for a new experiment that would take a year and leave the conclusion untouched. A suggested citation that does not say what the reviewer thinks it says.

Think in proportions before you commit. Declining two of twenty comments reads as judgment; declining fifteen reads as a fight. The decision letter sets the ceiling on all of this: major and minor revisions mean different things for how much argument the editor will tolerate.

Concede the observation, contest the prescription

Almost every reviewer comment bundles two claims: an observation about your paper and a prescription for fixing it. The observation is usually right. The prescription is where you have room to say no.

Take the classic: "the sample is small; the authors should collect additional data." The observation is true. Your sample is small, and pretending otherwise costs you credibility you will need later in the letter. The prescription can still be wrong. Perhaps the study was powered for exactly this design. Perhaps the data collection window closed two years ago and a new cohort would be a different paper. Concede the first half in your opening sentence, then contest the second half with the power analysis, not with an adjective.

The split has a side effect: it shows the reviewer you read them carefully. A rebuttal that opens by agreeing is hard to read as hostile.

Build the rebuttal on evidence, not adjectives

An evidence-first rebuttal has a fixed shape: quote the comment, concede what is true, state your position in one sentence, then give the evidence. Close by pointing at whatever did change in the manuscript, even if it is one sentence in the limitations.

In the point-by-point response format, that looks like this:

The reviewer is correct that our sample (n = 42) is small. We did not collect additional data, because the study was powered for a within-subject design: the a priori power analysis (Section 2.3) required n = 38 for 80 percent power at the expected effect size. We now state this limitation explicitly in the Discussion (p. 14, lines 310 to 314).

Nothing in that paragraph performs confidence. Every sentence carries a fact the editor can check against the manuscript.

Phrases that sound firm without sounding defensive

The phrases that work share one property: they carry information instead of emotion. "We respectfully but strongly disagree" carries neither.

A few that hold their ground:

• "The reviewer is right that X. We kept Y because..." followed by the reason itself.
• "We considered this change and decided against it. Two things drove the decision."
• "This is a fair concern, and our data cannot resolve it. We now say so directly in the limitations."
• "We could not find support for this interpretation; the closest study we found reports the opposite pattern, and we now cite it in Section 4."

Then strip the intensifiers before you send. "We strongly believe" is weaker than "we believe," and both are weaker than the citation that follows the claim. The instinct that pads manuscripts with hedging language pads rebuttals too, and it reads the same way to an editor. The rule that catches it in a Waraq review is one line long. Filler / hedging: "Remove empty intensifiers and filler." Response letters earn that flag as often as manuscripts do.

Retire three phrases entirely. "With all due respect" convinces nobody. "As we already explained" is a scolding. "The reviewer misunderstands" blames the reader for the text; write "the original phrasing was unclear, and we have rewritten it" instead, even when you privately disagree.

When should you offer a compromise change instead?

Offer a compromise when the disagreement is about emphasis rather than substance, or when the reviewer's confusion is itself evidence that readers will be confused. You keep your analysis, and the paper still moves.

Compromise moves cost little. A sentence in the limitations. A supplementary table. A citation to the position the reviewer favors, placed next to your own. Each one lets the reviewer see their fingerprint on the paper without changing what the paper says, and experienced authors budget for it.

The one compromise to refuse is the false one: running an analysis you believe is inappropriate just to make a comment go away. That analysis will sit in the published record with your name on it, long after the reviewer's name is forgotten.

What if the reviewer pushes back in the next round?

Answer once more, briefly, and only with new material. If you have no new evidence, do not restate the old rebuttal in firmer language. Write to the editor and ask them to adjudicate the point.

A second-round disagreement is no longer between you and the reviewer. It is a question for the editor, and framing it that way is the polite exit: "We and the reviewer read this literature differently. We have set out both positions and defer to the editorial decision." Editors arbitrate disputes like this constantly. Asking for a ruling admits nothing; it is how the process is designed to end.

And if the editor sides with the reviewer, make the change or take the paper elsewhere. Both are respectable outcomes. Arguing a third time is not a strategy. It is a delay, with your manuscript as the collateral.