The Complete Guide to Responding to Peer Review Comments

You just received peer review comments on your manuscript and your stomach dropped. That is normal. The dread, the indignation, and (if you are honest) the uncomfortable recognition that some of the comments have a point. Every researcher knows the sequence. What separates papers that survive revision from papers that don't is rarely the quality of the science. It is how you respond to peer review comments in the document that accompanies your resubmission.

Here is the uncomfortable part most guides avoid stating directly. Strong revisions with weak response letters get rejected more often than decent revisions with excellent response letters. The response letter is the argument that the paper is ready. It has to be systematic and specific, and it has to know what the editor is actually deciding.

What follows is the whole sequence, from opening the decision letter to clicking submit.

What Do Reviewers Actually Want When They Read Your Response?

Before you write a word, understand the psychology on the other side. Reviewers are researchers too, usually overworked, almost certainly unpaid for the review itself, and in most cases genuinely trying to improve the paper even when their tone suggests otherwise.

When a reviewer opens your response letter, they are looking for four things:

• Evidence of engagement. Not agreement with everything, but engagement with the substance. They want to see that you read the comment, understood its intent, and took it seriously.
• Specificity. "We revised the manuscript" is not a response. "We added a subsection (3.2) with a sensitivity analysis across three parameter ranges, reported in new Table 4" is a response.
• Intellectual honesty. If you cannot do what they asked, they would rather you say so and explain why than watch you perform a hollow gesture that does not address the concern.
• Findability. They want to locate their comment, read your response, and reach the corresponding change in the manuscript without hunting through 40 pages.

Most peer review frustration on both sides comes from a mismatch in these expectations. Reviewers expect substantive engagement. Authors supply deflection dressed as compliance. Recognizing that asymmetry is half the battle. For the formatting dimension, see the response to reviewer comments template in Word and LaTeX.

Step 1: Read the Reviews. Then Close Your Laptop.

This is not performative advice. Read the decision letter and all reviewer comments once from beginning to end. Do not open your manuscript. Do not start revising. Do not compose draft responses in your head (you will do this anyway, but recognize it as emotional processing instead of strategy).

The first read is for absorbing the blow and letting the sting dissipate. Give yourself 24–48 hours before you begin working. Every senior researcher will tell you this, and every early-career researcher ignores it at least once, usually to their regret.

After the cooling period, read the reviews again. The second read is different. The comment you first found outrageous now looks like it identifies a genuine weakness. The comment you first agreed with turns out to rest on a misreading of your method. Emotional distance changes the signal-to-noise ratio dramatically. If a review has crossed from harsh into personal, see the Reviewer 2 survival guide before you draft anything.

Step 2: Categorize Every Single Comment

Open a spreadsheet or document and list every individual comment from every reviewer. Assign each one to a category. This categorization is the foundation of your revision strategy. Skip it and you will spend three weeks reacting instead of executing.

Must-fix: The reviewer is right and the change will make the paper better. Failing to address these is the fastest path to rejection. Typical examples: statistical errors, absent references to directly relevant work, and method descriptions too vague to tell readers what you actually did.

Negotiable: The reviewer has a legitimate underlying concern, but the specific fix they suggest is not the only option, or the best one. You can address the concern without doing precisely what was asked. The classic case: "You should run a new experiment." Maybe a reanalysis of existing data answers the same worry.

Misunderstanding: The reviewer misread something. This is still your problem. If an expert reader in your field misinterpreted the text, so will others. Correct the misunderstanding in your response and fix the writing that caused it. Never answer a misunderstanding without also revising the manuscript to prevent the next one.

Out of scope: The reviewer is asking you to write a different paper. This happens more often than it should. You can decline these requests, but the declination has to be justified. Explain why the requested change falls outside the paper's stated contribution and acknowledge it as a direction for future work.

For a printable version of this categorization framework inside a full revision checklist, see the revise and resubmit playbook.

Build the Comment Log Before You Build Anything Else

The categories above only pay off if they live somewhere you can sort. Make a spreadsheet with one row per comment and these columns:

• ID (R1.4, R2.1, ED.2 for editor comments) so every response has a stable handle
• Verbatim text pasted from the decision letter and never summarized
• Category (must-fix, negotiable, misunderstanding, out of scope)
• Type (analysis, framing, prose, factual)
• Hours your honest estimate, before you start
• Owner which co-author does the work
• Manuscript location filled in at the very end, after the text is frozen
• Status (not started, in progress, drafted, verified)

Four rows from a real-feeling log look like this. Comment R1.2, "the sample size justification is missing," must-fix, analysis, 4 hours, first author, Section 3.1. Comment R1.7, "consider citing the 2023 replication literature," must-fix, factual, 1 hour, second author, Section 2.2. Comment R2.3, "validate on a second dataset," negotiable, analysis, 40 hours, owner still unassigned. That single row decides whether your revision fits the deadline. Comment R2.9, "Figure 2 axis labels are unreadable," must-fix, prose, 30 minutes, first author.

Two things fall out of the log immediately. The first is your critical path, which is always the small number of rows with double-digit hour estimates. The second is your completeness check: at submission, every row needs a drafted response and a verified location, and an empty cell is a comment you were about to skip. Sort by Status before you upload anything.

The log also settles arguments with co-authors. "Did we deal with Reviewer 2's point about attrition?" has a one-cell answer instead of a thread.

Step 3: Prioritize by Impact on the Accept/Reject Decision

Not all reviewer comments carry equal weight. Some are gatekeepers: if you do not address them, the paper is dead regardless of everything else you do. Others are cosmetic suggestions no editor would reject a paper over. Your revision time is finite (usually 60–90 days for major revisions), and spending it on the wrong priorities is a common and avoidable failure.

Prioritize in this order:

1. Structural and methodological concerns. These take the most time and carry the most weight in the decision. If a reviewer questions your sample selection or your statistical approach, this is where your first week goes.
2. Requests for new analyses or evidence. Start these immediately, because they may involve running computations, re-analyzing data, or producing new figures. Waiting until week three guarantees a last-minute crisis.
3. Framing and argumentation. How you position your contribution, what you claim and do not claim, how you discuss limitations. Intellectually demanding, quick to execute.
4. Clarity and writing quality. Sentence-level edits, paragraph reorganization, better figures. Important, and still lower priority than the structural issues above.
5. Minor corrections. Typos, formatting, missing references. Do these last.

The error to avoid is working bottom-up. Spend week one polishing prose and fixing citations, then discover in week three that the new sensitivity analysis the reviewer requested produces unexpected results, and you have left yourself no room to think.

Step 4: Write the Point-by-Point Response

The standard format for responding to peer review comments is not optional in most fields. For each reviewer comment, the entry looks like this:

Reviewer 1, Comment 3. [Quote or faithful paraphrase]

Response: [What you did and why]

Changes in manuscript: [Section, page, line numbers]

Three principles govern the quality of each response.

Be specific. The sentence "We agree and have revised accordingly" is a failure of communication and no act of concision. Write instead: "We agree that the sample size justification was insufficient. We added a formal power analysis (Section 3.1, lines 112–118) showing that n=340 provides 80% power to detect a medium effect size (d=0.5) at α=0.05. The study is therefore adequately powered for our primary hypothesis."

Lead with agreement when possible. Even when you are about to disagree, an opening line that acknowledges what is valid in the comment costs nothing and signals good faith. "The reviewer raises an important point about external validity" stays true even when you then explain why external validation data are unavailable.

Provide evidence instead of assertions. "We believe our approach is correct" is an assertion, and a weak one. "Our approach follows the recommendation of [Smith et al., 2023], who demonstrated that method Y outperforms method X for datasets with these characteristics (see their Table 3)" is evidence. Evidence persuades. Assertions irritate.

For ready-to-use scaffolds you can drop your individual responses into, see the copy-paste rebuttal letter templates.

How to Handle Contradictory Reviewers

Reviewer 1 says your paper is too long. Reviewer 2 says you need more detail. Reviewer 1 praises your theoretical framework. Reviewer 3 wants it deleted. The situation is stressful and completely ordinary, and editors expect it.

Flag the contradiction explicitly. Write it out: "We note that Reviewer 1 and Reviewer 3 offer opposing suggestions regarding the theoretical framework (Reviewer 1 wants it retained and expanded, Reviewer 3 wants it removed entirely). We have attempted to balance both perspectives by…" Pretending the contradiction does not exist insults everyone's intelligence.

Look to the editor's letter for prioritization signals. The editor often tips their hand. If the decision letter says "please pay particular attention to Reviewer 1's methodological concerns," that is a hierarchy and not a suggestion. Follow it.

Find the underlying convergence. Contradictory surface-level suggestions frequently point at the same underlying problem. "Too long" from Reviewer 1 and "needs more detail" from Reviewer 2 might both be solved by cutting the weakest sections and strengthening the most important ones. The reviewers diagnosed the same issue (uneven depth) from opposite directions. On how to tell that case apart from genuine expert disagreement, see what to do when reviewers contradict each other.

Be transparent about the trade-off. When no synthesis is possible, say so: "Given the opposing recommendations, we chose to [action], because [reason]. We are happy to adjust this if the editor prefers a different approach." That puts the decision where it belongs, with the editor who has read all the reviews.

How to Disagree with a Reviewer Without Getting Rejected

You are allowed to disagree. Editors expect it, and a response that agrees with everything (including mutually contradictory comments) signals a lack of critical engagement. But how you disagree determines whether the disagreement reads as confidence or as arrogance.

The Formula That Works

1. Acknowledge the comment and the concern beneath it
2. Explain your reasoning, backed by evidence (citations, data, logical argument)
3. Offer something: a clarification in the text, a partial analysis, a new sentence that addresses the concern even if you did not make the requested change
4. Defer when appropriate: "We are happy to make this change if the editor feels it is warranted"

Before/After: Disagreeing with a Methodological Suggestion

Reviewer comment: "The authors should use method Z instead of method Y. Method Y has known biases."

Weak response:

We disagree. Method Y is widely used and appropriate for our data.

This is an assertion without evidence. "Widely used" is an appeal to popularity and not a defense. The reviewer will not be persuaded, and the editor has no basis for siding with you.

Strong response:

We appreciate the reviewer's concern about potential biases in method Y. We selected method Y over method Z for two reasons: (1) method Z requires stationarity assumptions that do not hold for our dataset (see diagnostic tests in new Appendix C), and (2) recent work by [Chen & Park, 2024] demonstrated that method Y's bias is negligible for sample sizes above n=200, which applies here. To address the concern directly, we re-ran the full analysis with method Z in Appendix D. The results are qualitatively identical (new Table D1), which supports the validity of our primary analysis. We added a note in Section 3.2 (line 134) pointing to that parallel analysis.

Before/After: Handling a Request You Cannot Fulfill

Reviewer comment: "The authors should validate their model on a second dataset from a different geographic region."

Weak response:

We do not have access to another dataset.

Factually true, strategically disastrous. It reads as unwillingness to engage with a legitimate concern about generalizability.

Strong response:

This is an excellent suggestion that would strengthen the paper's generalizability claims. Unfortunately, datasets with comparable variables for a different geographic region are not currently available in the public domain, and new data collection falls outside the scope of this revision. We addressed this concern in three ways: (1) we added a cross-validation analysis using a held-out 30% subset of our data (new Section 4.4, lines 245–260), (2) we expanded the Limitations section (lines 312–320) to discuss geographic generalizability explicitly as a direction for future work, and (3) we compared our coefficient estimates with those reported by [Rivera et al., 2023] for a European sample (new Table 5). The patterns are broadly consistent.

Before/After: Reviewer Wants Something Removed

Reviewer comment: "The literature review is too extensive and detracts from the paper's contribution."

Weak response:

We believe the literature review is necessary to contextualize our work.

Strong response:

We take this point seriously and have shortened the literature review from 2,400 words to 1,600 words. Specifically, we removed the review of tangentially related work on [topic] and condensed the discussion of [second topic] into a single summary paragraph. The material that remains directly supports the development of our hypotheses in Section 3. We retained the subsection on [core topic] because it introduces the theoretical gap that our study addresses. Without it, the research question in Section 3.1 would lack sufficient motivation.

Special Cases the Standard Process Misses

Editor-only comments

Some decision letters carry a paragraph from the handling editor that never appears in any reviewer report. Give those comments their own numbered block (ED.1, ED.2) at the top of your response, ahead of Reviewer 1. Editors notice when their own points are buried at the bottom, and their points are usually the ones that decide the case.

Comments hidden in an annotated PDF

Reviewers sometimes attach a marked-up manuscript instead of writing a list. Those margin notes count. Transcribe every one into your comment log with an ID like R2-PDF-14 and a page reference, and respond to them in the letter the same way you respond to numbered comments. Editors rarely reopen the annotated file to check, but the reviewer who wrote the notes certainly will.

The reviewer who says they skipped something

"I did not have time to check the derivations in Appendix B" is not a comment you can answer, and it is a warning. That appendix is going to a fresh reader, or to a statistical reviewer, in round two. Use the revision window to tighten it even though nobody asked.

The terse review from a non-native English speaker

Short, blunt comments are not always hostile. "Method unclear. Rewrite." carries the same information as three polite paragraphs. Read for the technical content and respond at full length. Answering terseness with terseness is the mistake.

The very short major revision window

If the editor gave you 14 days instead of 60, the categories do not change but the triage does. Ask for an extension within the first two days if you need one, because a request on day 12 reads as panic. Where no extension is coming, spend the time on comments that gate acceptance and answer everything else correctly and briefly. Answer all of them. Never fewer.

Managing the Revision Timeline

For major revisions you typically have 60–90 days. That feels generous until you notice that new analyses take weeks, co-author feedback cycles are unpredictable, and the response letter alone needs substantial focused time. Realistic timelines look something like this:

• Days 1–3: Read reviews, process emotions, begin categorization
• Days 4–7: Complete categorization, build revision plan, identify new analyses needed
• Days 8–30: Execute major revisions (structural changes, new analyses, rewritten sections)
• Days 31–45: Draft the point-by-point response letter
• Days 46–55: Revise response letter, align with manuscript changes, check for consistency
• Days 56–60: Co-author review of both documents, final proofreading, submission

The mistake that derails the most revisions is underestimating the time the response letter requires. It is not a summary of what you did. It is a persuasive document with its own drafting and polishing cycle. Budget at least 10–15 days.

One step belongs at the very end and is easy to lose. Verify every manuscript location you cited. Line numbers move every time you edit. Freeze the text, regenerate the line numbers, then walk the comment log row by row and confirm that each "Changes in manuscript" field points where you say it does. For a checklist built around this timeline, see the day-by-day revise and resubmit checklist.

The Second Round After You Submit

After resubmission, one of several things happens. The editor may accept without further review, which is common for minor revisions. The editor may send the paper back to the original reviewers, which is standard for major revisions. Or the editor may invite a new reviewer, which feels like starting over but usually means they want a tiebreaker.

In the second round, reviewers evaluate one thing above all. Did you take their comments seriously and answer with substance? They are not usually re-reviewing the paper from scratch. They are checking whether their concerns were addressed. That is why the response letter carries so much weight. It is the document they read first, and it sets the frame for how they read the revised manuscript.

Second-round comments come in two flavors, and telling them apart matters. "The authors' response to my point 3 does not address the issue" means you have a live problem and one more chance to fix it. "The authors have addressed my concerns. I have one remaining minor suggestion" means the reviewer has moved to accept and is tidying up. Answer the first kind at length. Answer the second kind in two sentences, and do not reopen arguments the reviewer has just let go.

Facing 30+ reviewer comments and a 60-day deadline? ReviewPanel.ai analyzes your manuscript alongside every reviewer comment and generates a structured, point-by-point response with specific edit suggestions. One paper, one price, $29.

Frequently asked questions

How long should I take to respond to peer review comments?

Use the full window the journal gives you, minus a few days of buffer, but start immediately. For a 60-day major revision, the first week goes to reading and categorizing instead of writing, weeks two through four go to analyses and structural changes, and the last three weeks go to the response letter and co-author review. Submitting in 10 days does not impress anyone. It usually means new analyses were skipped, and reviewers notice that in round two.

Do I have to respond to every single comment?

Yes, every one, including comments you think are wrong and comments that are pure typo corrections. Editors check completeness before they check quality, and a missing response is the easiest failure for them to spot. Responses to comments you reject are usually shorter than concessions: two or three sentences that acknowledge the concern, give your reason, and point at the manuscript location that resolves it. Silence, by contrast, reads as either carelessness or evasion, and editors cannot tell which.

What if the reviewer is wrong about a fact in my paper?

Correct the record once, politely, with a specific pointer, then fix the writing that made the error possible. "We believe there may be a misunderstanding" followed by "our approach is described in Section 2.3, lines 88–95" does the work without a fight. Then add a clarifying sentence at that location and say you did. If one expert reader misread the passage, other readers will too, so the correction is also a genuine improvement to the paper.

How do I handle two reviewers who want opposite things?

Name the conflict in your response, by reviewer and comment number, before you defend your choice. State what each reviewer asked for, say which way you went and why, then invite the editor to overrule you. Editors resolve these conflicts constantly and much prefer doing it in round two over discovering in round three that you quietly satisfied one reviewer and stonewalled the other. Silent selection is the only genuinely bad option here.

Should I email the editor before I submit my revision?

Usually no, with two exceptions. Write if you need an extension, and write early instead of on the deadline. Write if a reviewer request is impossible for a reason the editor should weigh before you spend six weeks on it, such as data that no longer exists or an analysis that would require ethics approval you cannot obtain in the window. For anything else, put the question in the cover section of your response letter, where it becomes part of the record.

What if the second round brings brand-new comments?

Answer them as ordinary comments, without irritation, in the same point-by-point format. New comments in round two usually come from a reviewer who was added late or from a reviewer reading a section that changed substantially. Where a new comment contradicts something settled in round one, say so in a single neutral sentence and point at the earlier exchange. Your round-two letter should stand on its own, because a newly added reviewer has not read your round-one letter.