The Scholar's Vault · DOCX template with annotated examples
The letter that gets a revision read as a revision.
Response to reviewers: structure, decision rules and two annotated examples
A revise-and-resubmit is decided twice. Once by the reviewers, who read the manuscript, and once by the editor, who reads your letter first and the manuscript second. This template is built for the second reading.
No ghostwriting, ever · Human-led, PhD-signed
The page is the content. The download is the working copy — a DOCX with the response block already styled, the five-class triage table as a fillable grid, and the two annotated examples in comment balloons you delete as you go. One email covers all six Vault resources.
What this is, and what it is not
A structure, decision rules, and a bank of sentence shapes.
You write the substance, in your words, and you sign it.
We do not draft a response letter for anybody’s signature, and this file will not do it either. The two annotated examples below are worked through in full so you can see the reasoning, and both stay empty at the point where your evidence goes, because that is the only part that decides anything.
The three moves that matter are not writing moves. They are: sorting the objections, deciding which ones you lose on purpose, and putting the decisive sentence where the editor actually reads.
The night-before version
If your resubmission is due tomorrow, do these three and skip the rest. They carry most of the effect; everything else on this page makes them better.
- Triage. Split every comment into its own row and mark each one Fatal, Structural, Evidential, Presentational or Wrong.
- Verdict first. Open every response with “Accepted”, “Accepted in part” or “Declined”, and give a location for every claimed change.
- One disagreements block. Move every decline into a single block and name it in your opening paragraph.
01 · The triage pass, before you write a word
Sort the objections first.
Budget 45 minutes. This step, not the prose, is what separates an accepted revision from a second round.
- Copy every reviewer comment into a numbered list, one objection per row. A comment containing three complaints becomes three rows. Reviewers write in paragraphs and editors check against lists.
- Classify each row against the table below. Write the class in a column beside it.
- Count the Fatal rows. If the count is zero, this is a revision. If the count is above two, book a conversation before you spend three weeks on it.
The five classes of objection
| Class | What it means | Default move | Typical share of a review |
|---|---|---|---|
| Fatal | The claim cannot survive the objection with the data you have | Narrow the claim, or run the analysis | 0–2 rows |
| Structural | The argument is there but the reader cannot follow it | Move, split or signpost. No new data | 2–5 rows |
| Evidential | A specific check, control or robustness test is missing | Run it, report it, say where | 3–8 rows |
| Presentational | Table, figure, notation, length, reference formatting | Do it without discussion | 5–15 rows |
| Wrong | The reviewer has misread, or is asking for a different paper | Defend, once, with a quotation | 0–3 rows |
The count in the right-hand column is what a normal review looks like. A review with fifteen Presentational rows and one Fatal row reads as brutal and is close to an acceptance.
02 · The concede-and-defend rule
Concede everything that costs you nothing.
You are not scored on how much you defend. You are scored on whether the editor can tell, in one pass, that you understood what was asked.
The working rule: concede everything that costs you nothing, so that your two defences are visible. A letter that fights on twelve fronts has no front.
| Situation | What to do | What it costs you |
|---|---|---|
| The reviewer is right | Change it, say so in one sentence, give the page and line | Nothing. Spend these freely |
| The reviewer is right about the problem, wrong about the fix | Concede the problem in your first clause, then propose your fix | Two sentences, and it reads as expertise |
| The reviewer asks for an analysis you can run in under a week | Run it. Put it in an appendix if it says nothing | A week, against an eleven-week second round |
| The reviewer asks for data you do not have and cannot get | Narrow the claim in the manuscript to what the data supports, then report the narrowing | Scope. This is the honest trade, and reviewers accept it far more often than authors expect |
| The reviewer has misread the paper | Defend, and treat the misreading as evidence that the text is unclear. Change the text as well | One paragraph, and you keep the finding |
| The reviewer wants a different paper | Decline in two sentences, name what the paper is about, and offer the alternative as future work | The editor decides. Say it once and stop |
| Two reviewers contradict each other | Name the contradiction, choose one, and give the reason in a sentence the editor can adopt | Nothing. Editors are relieved when authors do this for them |
The rule about the last row. When reviewers disagree, the editor has to break the tie, and will use whatever reasoning is available. Put yours in the letter, and it is usually the reasoning that gets used.
03 · The letter architecture
What goes where.
Six blocks, in this order. The order is the argument.
01 · The opening paragraph
Three sentences maximum
This is the sentence that decides whether the editor reads the rest, and it is the last thing you should write.
It has to do three things: state what changed structurally, name the number of substantive changes, and say what you did not do and why in half a clause.
Template shape
“We thank the reviewers for [SPECIFIC THING THEY CAUGHT — not ‘their helpful comments’]. The revision makes [N] substantive changes, of which the largest is [ONE-CLAUSE DESCRIPTION]. We have not [THE ONE THING YOU DECLINED], for the reason set out at point [N].”
The third sentence is the one nobody writes. It is also the one that stops an editor discovering your refusal on page eleven and reading everything before it as evasion.
02 · The summary of changes
Between three and six bullets
Each names the change and the location, and nothing else.
An editor who reads only this block must be able to decide whether the revision is serious. Write it so that it is true.
03 · The point-by-point response
One block per numbered objection
In the reviewers’ order rather than yours. Reviewers check their own comments against your numbering, so never renumber, never merge two comments into one response, and never answer out of sequence.
The repeating unit is below, and it is the part of this resource you will use most.
04 · The disagreements
Two paragraphs is the ceiling
Every point where you decline goes here, together, and nowhere else. Scattered disagreement reads as defensiveness. Concentrated disagreement reads as a position.
If you need more than two paragraphs, the objection is fatal, and you are arguing with the wrong problem.
05 · The changes nobody asked for
Three to five lines
A short block listing what you improved on your own initiative. This block tells an editor the manuscript was reread rather than patched.
Keep it factual. A new sensitivity analysis belongs here; a rewritten sentence does not.
06 · The close
Two sentences
What the revision now supports, and a plain offer to do more if the editor names it.
No gratitude inflation, no restating the summary. The letter ends on the finding, not on your feelings about the process.
04 · The three phrasings
Three sentence shapes, and when to spend them.
Each one gives the editor language to use in the decision letter. That is why they work, and it is the only reason they work.
Phrasing One · The attribution move
Give the reviewer authorship of the improvement, in specific terms, once.
“Reviewer 2 is right that [SPECIFIC CLAIM] was not supported by [SPECIFIC EVIDENCE]. We have [WHAT YOU DID], and the result is reported at [LOCATION]. The revised claim is narrower and better evidenced than the one we submitted.”
Why it works. A reviewer who reads their own objection resolved in their own terms has an outcome to point at. Vague thanks give them nothing to point at, so they go looking for a new objection.
Use it on your two or three most substantial concessions. Use it on all fifteen, and it stops meaning anything.
Phrasing Two · The scope move
For the objection you cannot satisfy with the data you have. Concede the boundary rather than defending the claim.
“We agree that [OBJECTION] limits what these data can establish. Rather than [WHAT THE REVIEWER ASKED FOR, WHICH YOU CANNOT DO], we have narrowed the claim at [LOCATION] to [NARROWER CLAIM], and added [ONE SENTENCE] to the limitations. The narrower claim is the one the design supports.”
Why it works. It converts a demand for evidence you cannot produce into a decision about scope, and scope is a decision the author is entitled to make. Reviewers rarely insist once a claim has already been narrowed.
Phrasing Three · The misreading move
For the reviewer who is arguing with a paper you did not write. Take the blame for the ambiguity, keep the finding.
“This may reflect a lack of clarity on our part. [SECTION] was intended to establish [WHAT IT ACTUALLY DOES], not [WHAT THE REVIEWER READ IT AS]. We have rewritten the opening of [SECTION] and added a signpost at [LOCATION] so the distinction is explicit at first mention.”
Why it works. It removes the contest. The reviewer is not told they were careless; the text is told it was unclear, and both parties keep their standing.
Never use the phrase “the reviewer has misunderstood”. It is true perhaps half the time, and it costs you the room every time.
05 · The response block, as a repeating unit
The unit you will use fifteen times.
Copy this once per objection. Keep the four parts in this order.
Comment [R#].[N] [Reviewer's words, quoted in full, italicised. Never paraphrased.] Response [One sentence: accepted, accepted in part, or declined. Say which, first.] [One to three sentences: what you did, or why you did not.] Change [Exact location: section, page, line, table or figure number.] [The revised text, quoted, where it is under 40 words.] Status Accepted · Accepted in part · Declined (see Disagreements)
Three constraints on this block
- Quote the comment in full. An editor comparing your letter to the review must not have to switch documents.
- Lead with the verdict. “Accepted.” then the explanation. A response whose verdict arrives in sentence four reads as a negotiation.
- Give the location as a number. “We have clarified this” without a line number is unverifiable, and an unverifiable claim in a response letter is the same error as an unverifiable claim in a paper.
06 · Two annotated examples
One where the reviewer is right. One where the reviewer is wrong.
Two exchanges, worked through. The notes beside each paragraph are the annotation, and they come out of your version.
Example One · The reviewer is right, and hostile about it
A serious objection, delivered with an edge. The letter answers the objection and ignores the edge.
Comment R2.3
“The authors claim a causal effect, but the design is cross-sectional. This is a serious overstatement, and I am not convinced the paper can be salvaged.”
Response
Accepted in part.
Verdict first. “In part” is honest here and buys the defence in sentence three.
Reviewer 2 is right that the design does not support a causal claim, and the language at [SECTION] did overstate it. We have replaced the causal framing throughout with [SPECIFIC ALTERNATIVE FRAMING], and the abstract now reports the association rather than the effect.
Phrasing one, spent on a real concession.
We have not removed the mechanism discussion at [SECTION]. It is now explicitly labelled as a proposed mechanism consistent with the association, and the two testable predictions it generates are stated at [LOCATION] so a subsequent design can reject it.
The defence, once, with the reason built in. The reviewer is offered something to want rather than something to fight.
Change
[SECTION], [PAGE]: causal language removed at eleven points, listed in the tracked-changes file. Abstract, [LINE]: revised. [SECTION], [LINE]: relabelled mechanism; added predictions.
Counted. “Eleven points” is checkable; “throughout” is not.
Status
Accepted in part.
What that exchange does. It concedes the thing that was actually wrong, keeps the contribution, and gives the editor a sentence to write in the decision letter. The reviewer’s word “salvaged” is never mentioned. Heat is not answered with heat; it is answered with line numbers.
Example Two · The reviewer is wrong
Phrasing three in use. The temptation here is to prove the reviewer careless, and that is the move that costs you the paper.
Comment R1.6
“The authors present [METHOD] as though it were validated for [POPULATION]. It is not. The entire analysis in Section 4 rests on an assumption the literature rejected a decade ago.”
Response
Declined, with a change to the text.
Verdict first, even here. The second clause stops the decline reading as a refusal to move at all.
This may reflect a lack of clarity on our part. [SECTION] does not apply [METHOD] to [POPULATION]; it applies it to [WHAT YOU ACTUALLY DID], which the validation in [CITATION, YEAR] covers, and the distinction was only made in the appendix.
Phrasing three. The text is told it was unclear, and the reviewer keeps their standing.
We have added a sentence at [LOCATION] stating the scope of the validation at first mention, and moved [APPENDIX ITEM] into the main text at [LOCATION].
The concession that costs nothing. A misreading is evidence the text is ambiguous, whoever is right.
If the objection is instead that [CITATION, YEAR] should not be relied on at all, we would welcome the specific reference the reviewer has in mind, and will address it in the next round.
The one sentence that keeps a wrong reviewer from escalating. It offers a route rather than a verdict, and the editor can see you offered it.
Change
[SECTION], [LINE]: scope sentence added. [SECTION], [LINE]: appendix material moved into the main text, three paragraphs.
Status
Declined, with a clarifying change. Listed in the Disagreements block.
What that exchange does. It holds the finding, gives the editor a checkable reason, and leaves the reviewer a way to be useful in round two rather than aggrieved. Note the last sentence of the response: asking for the specific reference costs nothing if the reviewer has one, and closes the point if they do not.
07 · The cover note, the editor’s comments, and round two
What the point-by-point does not cover.
Three things sit outside the point-by-point document, and all three get skipped.
- The cover note to the editor. Most journals want a short letter alongside the response document. It is not a summary of the response, it is a decision aid: what changed structurally, in three sentences, and the one thing you declined. Six to eight lines, and never longer than the abstract.
- The editor’s own comments. Where the decision letter carries editorial requests, they are answered first, in their own block, before Reviewer 1. Editorial requests are usually about fit, length or compliance, and they are the requests the person deciding cares about most.
- Round two. Number your responses continuously from round one, so R2.3 in the second letter refers to the same objection it did in the first. Where a point was accepted in round one, say what you did and stop. Where a reviewer returns to a point you declined, answer the new argument rather than restating the old response, and if there is no new argument, say so in one line and refer to the previous letter by point number.
Major revision, or reject and resubmit
These read the same in an inbox and are handled differently.
| Aspect | Major revision | Reject and resubmit |
|---|---|---|
| What it is | The manuscript stays in the system | A new submission, usually a new number |
| The letter | Point-by-point, in the reviewers’ numbering | Point-by-point, plus a cover note explaining what makes this a different paper |
| What decides it | Whether the objections were addressed | Whether the objection that killed round one is gone |
| The trap | Treating it as a rejection and giving up | Treating it as a revision and changing too little |
If the decision letter names a specific fatal objection and invites resubmission, that objection is the whole assessment. Fix it first, and open the cover note with it.
Naming the files
Editors handle six versions of everything. Use these three, keep the same pattern in round two, and always supply both a tracked and a clean manuscript unless the journal asks for one.
- [AUTHOR]_[MSID]_response_R1.docx
- [AUTHOR]_[MSID]_manuscript_tracked_R1.docx
- [AUTHOR]_[MSID]_manuscript_clean_R1.docx
08 · On AI-drafted response letters
Editors have read enough of them to recognise the shape.
A response letter written by a model produces exactly the failure the table below names first: fluent thanks, equal-length answers to fifteen objections, and no location numbers.
There is a second cost. A response letter is a submission document, and a growing number of journals ask authors to disclose the use of generative tools in preparing submitted material. A disclosure you did not make is a problem that outlives the paper.
Major publishers generally enforce one of two disclosure protocols. Check the journal’s current author guidelines before you submit — this is a policy area that moves.
ElsevierVerbatim declaration
Requires a standalone section immediately above the references list entitled “Declaration of Generative AI and AI-assisted technologies in the writing process”, carrying their mandated statement:
“During the preparation of this work, the author(s) used [NAME TOOL / SERVICE] in order to [REASON]. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.”
Springer Nature and WileyContext-based disclosure
Allow disclosure inside the Acknowledgments, Methods or Introduction sections rather than requiring a verbatim block. You state the tool name, the specific tasks it assisted with, and confirm explicitly that the human authors remain fully accountable for the final content.
Basic checks of grammar, spelling, punctuation or formatting, by a standard word-processor utility or a spelling checker, are universally exempt and require no declaration under any major publisher’s guidelines.
This is why the template is empty at the point where your evidence goes. The sorting, the concessions and the line numbers are the letter. The prose around them is the part a machine can write, and also the part that decides nothing.
09 · Before you send
Nine checks, in order.
- Every reviewer comment has a numbered response, in the reviewers’ numbering.
- Every response opens with a verdict word: accepted, accepted in part, or declined.
- Every claimed change carries a location a person can open and check.
- All declines sit in one block, and there are no more than three.
- The opening paragraph names what you declined.
- Nothing in the letter is described as “clarified” without a quoted revision beside it.
- No sentence blames a reviewer for reading badly.
- The changes-nobody-asked-for block exists and is factual.
- Read it aloud. Rewrite anything that sounds like a negotiation as a report.
Failure modes
Seven ways a letter says something you did not mean.
What the letter does, and what the editor reads instead.
| What the letter does | What the editor reads |
|---|---|
| Thanks the reviewers for their “helpful and insightful comments” | A form letter. Nothing has been read yet |
| Answers fifteen objections at equal length | The author cannot tell which objection was serious |
| Defends at four separate points | Defensiveness, and now every concession is suspect |
| Says “we have addressed this” with no location | An unverifiable claim from someone submitting evidence |
| Renumbers or merges the comments | The reviewer cannot check their own points, and will say so |
| Runs to fourteen pages | The revision is being argued rather than shown |
| Never declines anything | Either the review was trivial, or the author has conceded a finding to be agreeable |
The last row is the one authors do not expect. A letter that concedes on every point, including the one where the author is right, tells an editor that the author’s judgement is available for purchase by whoever asks loudest.
Where this stops, and what we do
The one judgement the template cannot make.
It will not tell you whether an objection is Fatal.
That is a judgement about your data, your field, and what the journal has published before, and it decides the outcome. Two things help, and both are honest about their limits.
Rejection Reversal — the desk-reject or reviewer rejection turned into a ranked, fixable objection list, with a response strategy for each objection. You write the response, and we read your draft before it goes.
- Standard€549· 5 business days from written confirmation of scope
- Rush€849· 24 hours
Carries the Fit Guarantee: before work starts, a full refund or a credited rescope, your choice; after delivery, a free revision, then half the fee back if the work still misses the brief we agreed.
Or take the free route first. A written strategy brief: tell us where you are stuck, and within 3 working days we send a three-step plan, including which service fits, or that none of ours does. No follow-up sequence. Take this one if you are not sure the problem is the letter.
A reviewer’s “reject” is a list of fixable objections wearing a scary costume. This template is the costume removal, and nothing more than that.