How to Respond to Peer Review Comments

Reviewer comments feel personal, but treating them as a worklist and answering point by point turns revision rounds into acceptances.

Guide · 8 min read · Directive Publications Editorial Team

Step 1: Sort every comment

Make three lists:

1. Easy fixes (typos, clarity, references)
2. Analysis / writing changes
3. Disagreements (you will explain, not ignore)

Do the sorting mechanically, before you form an opinion about the review as a whole. Number every comment the way the reviewer laid them out — R1.1, R1.2, R2.1 — and copy each one into the list where it belongs. Where a comment bundles three requests into one paragraph, split it and number it as three. You will have to answer them separately anyway.

The point of sorting is to turn a wall of text into an estimate. List one is usually most of the volume and almost none of the work: wording, a missing citation, an undefined abbreviation, an unlabelled figure axis. List two is where the time actually lives — a robustness check, a reanalysis under a different specification, a section that has to be rebuilt because two reviewers got lost in the same place. List three is usually short, often one or two items.

When you finish, you know something you did not know an hour earlier: whether this is a weekend of work or six weeks. That answer decides whether you ask the editor for more time, and asking at the start reads very differently from going quiet and missing the date.

Two things to watch while sorting. The editor's own letter outranks the reviewers' — if the editor says the statistics must be addressed and the rest is at your discretion, that is your priority order regardless of how forcefully a reviewer wrote. And where two reviewers ask for opposite things, such as cutting and expanding the same literature review, do not silently pick one. Ask the editor which direction they prefer. A short email at that point is normal and saves a round.

Step 2: Change the manuscript first

Editors trust replies that point to line numbers / tracked changes, not promises.

For each comment:

• What you changed (or why you did not)
• Where it appears in the revised file

The order matters more than it sounds. Authors who draft the letter first end up promising changes they have not attempted, and some of those promises do not survive contact with the data. Then either the letter gets rewritten, or it goes out describing a revision that is not in the file. Reviewers check. Being caught claiming a change that is not there costs a full round and some credibility.

Set the file up before you edit. Turn on continuous line numbering across the whole manuscript rather than restarting each page, so "line 412" points at exactly one place. Work in tracked changes and produce two exports at the end: a marked copy showing every edit, and a clean copy that reads normally. Journals commonly want both, and sending only one is a routine reason a resubmission is bounced back by the editorial office before any reviewer sees it.

Keep a log as you revise — one row per numbered comment: the comment number, what you did, where it landed. A three-column table in a scratch file is enough. That log becomes the response letter almost verbatim, which is why the letter takes an afternoon rather than a week.

One caution about line numbers: they move. Write the letter against line numbers, then make one more edit in Section 2, and every number below it is wrong. Do the numbering pass last, from the final file, and name the section alongside the line. "Section 3.2, lines 288-295" survives a small shift in a way that a bare number does not.

Step 3: Write a point-by-point response

Use a consistent format:

Reproduce each comment verbatim instead of summarising it. A paraphrase looks like editing the question to suit the answer, and a reviewer who meets a softened version of their own words will read everything after it suspiciously. Paste their text, keep their numbering, and answer in their order even where another order would flatter you more.

In the revision line, quote the new wording when it is short: We revised Section 2.2 (lines 141-144) to read, "Because the sample was drawn from a single site, these estimates should not be read as national." A reviewer can approve that without opening your manuscript. Telling them you clarified the limitation sends them back to the file, and going back to the file is where new objections are born.

Answer every comment, including the ones you simply accept — "Corrected, thank you" is a complete response to a typo. Keep the letter as short as the review allows. A response that runs much longer than the review itself usually means you are arguing rather than revising.

Put a short covering paragraph to the editor above the point-by-point section: thank them, confirm the manuscript has been revised in line with the comments, and name the one or two substantive changes. If you have declined anything, say so there in a sentence rather than leaving the editor to discover it on page four.

When you disagree

Disagree cleanly:

• Stay factual
• Offer a compromise when possible (clarify limitations)
• Never attack the reviewer

You are allowed to disagree, and experienced reviewers expect it. Comments get declined for good reasons: the requested analysis is impossible with the data you hold, the suggested method answers a different question, the reviewer is asking for a study you did not do. What you cannot do is skip the comment and hope nobody rereads the review.

A disagreement that works has three parts — what you understood the reviewer to be asking, the specific reason you are not doing it, and what you did instead. The third part is what makes it land. Almost every declined comment has a partial concession available. You cannot run the longitudinal analysis, but you can state in the limitations that causal direction is untestable with this design. You cannot add the second cohort, but you can name it as the obvious next study.

Watch where the wording puts the blame. "The reviewer has misunderstood our design" makes the reader at fault. "We did not make our design clear, and have rewritten the opening of Section 2 to state it explicitly" arrives at the same conclusion and hands you a revision to point at. Where a comment rests on a factual error about the field, correct it with a citation rather than with tone.

If a comment is genuinely unreadable — you cannot tell what is being asked — say so plainly to the editor, state your best interpretation, and answer that. Guessing in silence is what produces a third round.

Common mistakes that add months

• Ignoring one hard comment
• Changing text without saying so
• Writing a long argument and changing nothing
• Submitting a messy unmarked file

All four have the same shape: each forces somebody to ask you a second time for something you could have settled the first time.

Ignoring one hard comment is the costliest. Reviewers work down their own list when the revision reaches them, and the missing item is precisely the one they notice. An unannounced improvement is nearly as damaging, because an undisclosed change means the reviewer now has to compare files to trust any other line of your letter. A long argument attached to an unchanged manuscript tells the editor the round produced nothing. And a messy or unmarked file — or a mismatch between the version you describe and the one you uploaded — often stops at the editorial office and never reaches a reviewer at all.

Before you resubmit

1. Every numbered comment has a response, including the ones you accepted outright
2. Every response names a section and line, taken from the final file
3. The clean copy and the marked copy contain the same text
4. Any declined comment is flagged to the editor up front, not buried
5. Contradictory reviewer requests were raised with the editor rather than resolved silently
6. A co-author has read the letter for tone, not only for accuracy

The co-author read is worth the delay. You are too close to the review to hear your own defensiveness, and a colleague can hear it in ten minutes. Ask for one thing specifically — mark any sentence that sounds annoyed — and rewrite whatever comes back marked.

How review is structured at Directive Publications: Peer Review Process
More Academy guides: Directive Academy