An editorial decision letter tells authors what the editor decided, why, and what happens next. Put the decision in the first sentence after the greeting. Give the reasons the outcome rests on, pass on the reviewer reports, say exactly what the author must do and by when, and promise no outcome that has not been decided.
Neither the Committee on Publication Ethics (COPE) nor the International Committee of Medical Journal Editors (ICMJE) sets a format for these letters. The structure and templates below are general editorial practice.
For a rejection after peer review, adapt the letter in what a handling editor does when two reviewers disagree. When the reviewers agree, leave out its paragraph about the positive report, and give the appeal route your journal states. That post also covers checking a reviewer's facts and editing an abusive report.
Every editorial decision letter needs the same five parts
| Part | What it says | Without it |
|---|---|---|
| 1. Decision | The outcome, in the journal's own category name | Criticism reads as rejection, or praise as acceptance |
| 2. Decisive reasons | The one to three points the outcome rests on | Every reviewer remark looks equally important |
| 3. The reports | Reports attached as Reviewer 1 and Reviewer 2; your own points marked as the editor's | Your view reads like an anonymous extra reviewer |
| 4. What to do | The actions, the response format and how to send it | The revision arrives in the wrong shape |
| 5. Deadline and next step | The date, what happens next and, for a rejection, the appeal route | The next step and the date are unclear |
Four rules apply to every letter:
- Promise nothing. The Principles of Transparency and Best Practice in Scholarly Publishing say journals should not guarantee acceptance of initial manuscript submissions. A revision letter should not imply it either, so cut "we look forward to accepting".
- Keep money out. A decision must never depend on an author's ability to pay, as the handling editor's role sets out, so no charge belongs in the letter.
- Own the reasons. The ICMJE Recommendations say a journal need not follow reviewer recommendations, favourable or negative. Write the outcome in the first person, as your own judgement, and give the reasons for it.
- Protect confidentiality. Remove anything written for the editor only and, under anonymous review, anything that identifies a reviewer. Do not upload the manuscript to an artificial intelligence (AI) system that cannot keep it confidential, unless the authors have given explicit permission.
Desk reject: name the failed check and say no peer review took place
Authors are told to read a desk rejection for the check that failed, as our guide to the checks editors make before peer review explains. Name it as a fact about the manuscript: scope, article type, a missing ethics statement, or a report incomplete against its EQUATOR Network checklist. A guideline governs reporting, so say the report is incomplete, not that the study "did not follow" it. Never give a negative result as the reason; the ICMJE says such studies should not be excluded.
If a check finds copied text, COPE's flowchart on plagiarism in a submitted manuscript has the editor query the author before any decision. COPE recommends no similarity percentage as a benchmark for investigating overlap.
Subject: Decision on [manuscript number], "[short title]"
Dear [author name],
I have decided not to send "[title]" for external peer review, and [journal] will not consider it further.
The reason is [one ground, stated as a fact about the manuscript, and where the requirement is published].
As no reviewers read the manuscript, there are no reports, and this is not an assessment of [the findings].
[choose one: A new submission that fixes this would be welcome and would be handled as a new manuscript. / [journal] does not publish this type of work. A journal that publishes [type of work] may suit it; the choice is yours, and nothing has been sent anywhere on your behalf.]
If you wish to appeal, [appeal route as stated in the journal's policy].
Yours sincerely,
[name], [role], [journal]
Major revision: merge the reports into one ranked list
"Please respond to all comments" leaves the author to rank two reports. Merge duplicate points, number them once, and mark each Required or Optional. COPE's ethical guidelines for peer reviewers ask reviewers to separate additional work that is essential to support the claims from work that would only strengthen or extend it. Apply that test across all the reports.
Dear [author name],
My decision on "[title]" is major revision. This decision does not commit the journal to accepting a revised version.
The decision rests on [the one to three decisive concerns].
Resolve each Required point below, or explain why it cannot be resolved. You may decline an Optional point with a short reason.
Required
1. [point] (Reviewers 1 and 2; [section])
2. [point] (editor)Optional
3. [point] (Reviewer 2)Where the reviewers conflict, I have said which request to follow at [point numbers]. Acceptance is not contingent on adding any reference suggested by reviewers or editors.
The full reports are attached. Please send a point-by-point response and a tracked-changes manuscript by [deadline], [how to send the revision]. The revised manuscript will normally be returned for re-review. If you need more time, or decide not to revise, please tell me before then.
Yours sincerely,
[name], [role], [journal]
Settle contradictory reviewer requests before the letter goes out
Responding to reviewer comments tells authors facing incompatible requests to choose a path or ask the editor. Decide each conflict yourself and write the instruction into the revision letter's numbered list.
The requests below are invented for illustration, not quoted from real reports.
| Conflict | How to settle it | Instruction in the letter |
|---|---|---|
| One reviewer asks for an analysis the other says the data cannot support | Let the study question decide, not which reviewer sounds more certain | "Do not add the subgroup analysis in point 6; the study was not designed for it." |
| A request that extends the study | Optional, unless a claim cannot stand without it; then require a narrower claim | "Point 7 is optional. Without the follow-up data, limit the conclusion to [time point]." |
| Both requests are defensible and the conclusions stand either way | Say that either is acceptable, and ask for the reason | "Points 8 and 9 conflict. Follow either one, and say in your response which you chose and why." |
| Requests to cite specific papers | Optional, and kept apart from the decision | "Cite these only where they support a statement. Acceptance does not depend on them." |
The citation row follows COPE Forum advice on a 2018 case, which suggested letters could say acceptance does not depend on adding references editors suggest. COPE notes that case advice is not formal guidance.
Minor revision: say what is left and who will check it
A minor revision is not an acceptance, so keep acceptance language out. Re-review by external reviewers is usually not needed, so say who will check the changes.
Dear [author name],
My decision on "[title]" is minor revision. The reports are attached, and the changes below are needed before I can make a final decision.
1. [change] ([section])
2. [change] ([section])[I will assess the revised version myself. / Point [n] will go back to Reviewer [n].] Please send a point-by-point response and a tracked-changes file by [deadline], [how to send the revision]. If you need more time, please tell me before that date.
Yours sincerely,
[name], [role], [journal]
Second-round decisions: judge the revision against the first letter
The points marked Required in the first decision letter are the test: check each one against the response and the tracked changes. Add a new requirement only if it is essential to the soundness of the work, and say why it was not raised before. If every Required point is resolved and any re-review raises nothing essential, the acceptance letter below may fit; otherwise use one of these paragraphs.
Further revision: My decision on the revised "[title]" is a further revision of point [n] only. Points [numbers] are resolved. Point [n] remains open: [what was asked, and what the response provides]. Please address it by [deadline].
Rejection after revision: My decision is to reject the revised "[title]". It does not resolve Required point [n]: [what was asked, and what the response provides]. Without that change, the conclusions in [section] cannot stand. Any appeal should follow [appeal route as stated in the journal's policy].
Acceptance: confirm the decision and leave the charge to the office
Say what the final version resolved and who will write next. Leave out fees, invoices and waivers, and let whoever handles the journal's finances raise any charge separately, so the editor who decides never asks for payment. Do not call acceptance final; the ICMJE notes an editor can still reject an article before publication if integrity concerns arise.
Dear [author name],
I am pleased to accept "[title]" for publication in [journal]. The revised version resolves [the decisive concerns].
The editorial office will write to you separately about the steps before publication. Thank you for your responses to the reviewers.
Yours sincerely,
[name], [role], [journal]
What Directive Publications has published about decisions
Directive Publications has not published an editor handbook, so these letters describe general practice, not our procedure. Our Peer Review Process page states that peer review is double-blind, that at least two independent expert reviewers are sought for each research manuscript, and that the handling editor decides. It says payment of any article-processing charge never influences editorial outcomes and does not guarantee acceptance. It also says reviewers and editors must not upload any part of a submitted manuscript to public or third-party generative-AI tools. Our Article Processing Charges page adds that editors and reviewers are not informed of an author's payment or waiver status during assessment. If you find an error here, report a problem to us.