Transparent Peer Review: What Authors and Institutions Expect

Transparent peer review is used loosely, so authors and institutions need to know which model a journal actually runs before they submit.

Explainer · 9 min read · Directive Publications Editorial Team

What it can mean

Depending on the journal:

• Publishing review reports alongside the article
• Signing reviewer identities (open identities)
• Publishing author response letters
• Sharing editorial decision letters

Treat those as four independent switches rather than one dial. Publishing the reports says nothing about whether the reviewers are named. Naming the reviewers to you says nothing about whether anything is published. One journal posts anonymous reports with the accepted article and never publishes your reply; another discloses reviewer identities to the authors at decision and publishes nothing at all.

Each switch changes something different for you:

• Published review reports. The criticisms of your manuscript sit permanently next to the final version. A reader can see the objection that was raised and judge for themselves whether your revision answered it.
• Signed reviewer identities. The name may go only to the authors, only when that reviewer chooses, or to everyone along with the published report. Ask which, because the three are very different on the day you are the reviewer.
• Published author response letters. Your point-by-point reply becomes part of the public record. Anything you wrote at the end of a bad week is quotable.
• Published editorial decision letters. These show what the editor asked for and how many rounds it took, which is usually flattering to nobody and informative to everyone.

Two further questions decide how much any of this touches you. Does publication of the file depend on acceptance — in most models nothing appears if the manuscript is rejected, so a rejection stays private. And is any part of it optional: some journals let reviewers decline to be named, and some let authors opt out of publishing the file, normally at submission rather than after a decision has landed.

Not every journal uses the same model. Directive Publications uses double-blind review — identities masked during evaluation: Peer Review Process

How to find out which model a journal actually runs

Do not rely on the word “transparent” in the journal’s description of itself, and do not rely on the publisher’s group-wide policy page. Transparency is usually set per journal, and sometimes per article type.

The fastest check takes two minutes and skips the policy prose entirely. Open a recent research article in that journal and look for a review history: a “peer review file”, “review reports” or “transparent peer review” tab on the article page, or a linked supplementary PDF. Open it. You will see at once whether the reports are there, whether names are attached, whether the authors’ reply is included, and whether the editor’s letters are in the file. Do that for two or three articles, including one of the article type you are submitting.

Then read the journal’s peer review policy for the parts a single example cannot tell you: whether the model is opt-in or opt-out, at what point you choose, and what happens if a reviewer refuses to be named.

If it is still unclear, email the editorial office before you submit and ask in one line: if this manuscript is accepted, will the review reports, the reviewer names, our response letter or the decision letters be published? That is an ordinary question, and an office that cannot answer it plainly has told you something as well.

What transparency is not

• A promise of faster acceptance
• A substitute for reviewer expertise
• Permission to publicly harass reviewers

Transparency changes what is visible about the process, not the standard your manuscript has to meet. The same kind of reviewers apply the same criteria, and a publishable file can lengthen the process rather than shorten it, because people write more slowly when a report will be read by more than one editor and two authors.

It is not a quality guarantee either. A published report by someone who was the wrong reviewer for your paper is a public bad report, and it stays there. What transparency gives a reader is the reasoning behind a decision, so they can weigh it themselves; it does not supply expertise that was never in the room.

The third bullet is the one institutions have to enforce. A signed or published report is not an invitation to answer back in public, to take the reviewer’s name to social media, or to raise it with them at a conference. It is equally not an invitation to guess an anonymous reviewer from phrasing or citation habits and then act on the guess. If a report is unfair, incompetent or hostile, the route is the same as under any other model: write to the handling editor, point to the specific claims you say are wrong, and ask for the report to be assessed or a further reviewer added. Say it once, in that letter, and let the editor handle it.

Writing a response letter that could be published

Assume a reader who is neither the reviewer nor the editor: a search committee, a competing group, the reviewer’s own doctoral students, and you in five years. Then write the letter you would be content for all of them to read.

In practice that is a handful of habits, none of which cost you anything:

• Reproduce each comment as the reviewer wrote it, number it, and answer in their order rather than yours.
• Say what you changed and where — section and line numbers from the final file — instead of asserting that a point has been addressed.
• Where you disagree, state what you understood the request to be, give the specific reason you are not doing it, and say what you did instead.
• Cut every adjective attached to the reviewer. “The reviewer has misunderstood our design” and “we did not make our design clear, and have rewritten the opening of Section 2” arrive at the same place; only one of them is quotable against you.
• Delete the sarcasm on the second pass. It always reads worse cold than it did the night you wrote it.

There is also a disclosure check that only applies when the file will be published. Read the letter for material you never meant to make public: unpublished data from a collaborator, an embargoed result, identifiable participant detail, a third party’s manuscript you have seen under review elsewhere, or the name of a colleague who ran an analysis and did not agree to be named in public. Most submission forms separate comments to the author from confidential comments to the editor. If something has to be said but should not be published, put it there and tell the editor why it is there.

What institutions should teach

• How to write responses that could be published
• How confidentiality works under each model
• How transparent reports support research assessment reform

Published review files are free teaching material, and they are the only realistic training set for the first of those. Pick two published articles in your discipline that carry open review histories — one where the exchange went well and one where it clearly did not — and run a session in which people rewrite the weakest response as though it were their own. Researchers pick up tone from examples far faster than from rules.

Confidentiality is the point that gets taught wrong, because people assume eventual publication cancels it. It does not. A manuscript under review is confidential while you hold it, whatever happens to the file afterwards: you do not circulate it, you do not use its results, and you do not paste it into a public AI tool. If you want a student to co-review for the training, ask the handling editor first and name the student — under an open model their contribution may end up on the record, which is a reason to ask rather than a reason to hide it.

The assessment point is the one a research office can act on directly. Reviewing has always been invisible labour, and a signed, published report is the first version of it a promotion panel can actually read. Decide whether your institution counts it, put that decision in writing, and tell staff which reviewing activity they may list and where. Otherwise you are asking people to attach their names permanently to their judgements for no recognised credit.

If you are the reviewer

Settle two things before you accept an invitation, because both are far harder to change later: whether the report will be published, and whether your name goes on it. Journals normally state this in the invitation; if yours does not, ask before you accept.

Then write every report as though it will be published, whether or not it will be. The standard is not complicated — specific, evidence based, addressed to the manuscript rather than to the authors, and free of anything you would not say to the corresponding author in person. Keep suspicions, integrity concerns and anything about the people rather than the paper in the confidential comments to the editor, which are not part of the published file.

Signing is a genuine choice with uneven costs. An early-career reviewer writing a critical report on a senior author in a small field is not being timid by declining to sign. Where a journal allows an anonymous report to be published, that combination is often the honest answer: the reasoning is public, the name is not.

Practical expectation setting

Read the journal’s review policy before submission. If reports will be published, write every reply as if future readers will see it.

Before you submit anywhere, settle four things:

1. Whether review reports are published with accepted articles in this journal
2. Whether reviewer names appear, and whether that is the reviewer’s choice
3. Whether your response letter and the decision letters form part of the published file
4. Whether you can opt out, and by when

Then write as though the answer to all four is yes. If it turns out to be no, you have lost nothing — the letter is better that way in either case.

More: Directive Academy · For Authors