Post-Publication Peer Review: A Guide for Authors

Once your paper is published, anyone can question it in public, and the way you answer becomes part of the record that readers see next to your work.

Explainer · 9 min read · Directive Publications Editorial Team

What post-publication peer review covers

The term covers two different things.

The first is a publishing model in which review happens after the article is public. F1000Research publishes articles after prepublication checks and then invites expert reviewers, whose names and reports appear alongside the article. eLife makes no accept or reject decision after review and publishes every article it reviews as a Reviewed Preprint, with public reviews and an eLife Assessment. If you are weighing a journal like this, our guide to transparent peer review covers what becomes public.

The second meaning, and the subject of this guide, is scrutiny of papers that have already passed conventional review: letters to the editor, PubPeer comments, social media threads and reanalyses in new papers. After publication, anyone with the article can check your figures, recalculate your percentages and ask why.

The practical consequence: keep the raw data, uncropped images and analysis code for every paper somewhere you can retrieve them. When a question arrives years later, the answer is usually in those files.

Where criticism of published articles appears

You cannot answer what you have not seen, so set up the basics now:

• Install the PubPeer browser plugin. It highlights discussed publications on any web page, including PubMed search results and journal websites.
• Create a signed PubPeer account early. Author replies need an account validated with a plausible institutional email address, and moderator checks can take a while.
• Name one co-author to watch for comments, so a question is not left hanging because everyone assumed someone else had seen it.

How PubPeer works from the author's side

PubPeer accepts comments on any publication with an identifier it recognises, such as a DOI or a PubMed or arXiv ID. According to the PubPeer FAQ, these are the rules that matter when your own paper is discussed:

• Commenters can be anonymous. Pseudonyms are assigned at random, and PubPeer says it will never willingly reveal information that might help identify a commenter.
• Comments must rest on publicly verifiable information. Allegations of misconduct are forbidden, and commenters are told to avoid personal remarks and speculation about motives. Anonymous comments appear only after moderation, which can take up to a week.
• Moderation is not scientific review. PubPeer does not review comments scientifically and gives no warranty that they are correct.
• You have a permanent right of reply. A reply that places responsibility on a particular co-author is accepted only if public information supports it, typically the published author contribution statement.
• Threads do not disappear. PubPeer generally will not remove a comment after a convincing author reply, or because the journal has published a correction.

PubPeer also says a comment should not be equated with an allegation of misconduct; many concern practices such as splicing gel lanes or omitting loading controls. Read each comment for what it actually claims, not for what you fear it implies.

Deciding which comments need a reply

The ICMJE Recommendations on correspondence state that authors of articles discussed in correspondence or an online forum "have a responsibility to respond to substantial criticisms of their work using those same mechanisms." On PubPeer, silence is not neutral: the thread stays, and the next reader sees an unanswered question.

The decision rule: reply when a careful reader relying on your paper would reach a different view of it without your answer.

Two distinctions help. A disagreement is not an error: the ICMJE Recommendations say matters of debate and evolving science are best handled as letters, correspondence or new publications, not corrections. And image questions often have an innocent explanation the paper did not state, such as two panels built from the same gel and therefore sharing a loading control. If that is the answer, show the originals and say so.

Writing a response that holds up

1. Go back to the source files. Open the raw data, uncropped images and analysis script before writing. Re-reading the paper tells you what you claimed, not whether it was right.
2. Agree the reply with co-authors. Decide who posts, usually the corresponding author or whoever generated the data, and circulate the draft first.
3. Answer the specific point. Restate the question in one sentence, then answer it with the figure, table or data that settles it.
4. Show the evidence. PubPeer encourages authors to share original data: post small images in the thread, and deposit larger datasets in a repository such as Zenodo, Dryad or figshare and link them.
5. Say what happens next. If the comment has found a real error, say you have written to the journal, and then actually write to it. A thread is not a correction.

An illustrative reply:

Much of the craft in responding to peer review comments carries over. The difference is that this reply is public, permanent, and read by people who never saw the original review.

When a concern reaches the journal

Journals handle two kinds of post-publication concern differently.

A formal critique. This is a letter or commentary submitted to the journal. COPE's flowchart on handling post-publication critiques notes that journals often invite the original authors to reply, that both pieces may be peer reviewed, and that journals should be clear with everyone about timelines. If you are invited to reply, meet the deadline and word limit, answer the critique rather than restating the paper, and declare competing interests as the journal asks.

A concern about integrity. This covers possible fabrication, image manipulation, plagiarism, or an error serious enough to undermine the results. ICMJE says the editor should follow procedures such as COPE's, consider informing institutions and funders, and may publish an expression of concern while those procedures run. A COPE case analysis of critiques after publication makes the key point for authors: time limits for letters do not apply to integrity concerns. In that case, an author argued that a concern had arrived after the journal's letter deadline. COPE's advice was that editors are ethically bound to act, and that if an author stays unresponsive, the institution can be contacted and an expression of concern added.

Concerns raised on social media often end up with the editor too. COPE's flowchart for that situation advises journals to acknowledge the concern publicly and then move the discussion to a journal contact. Do the same rather than argue in a thread.

If an editor writes to you about a comment:

1. Reply by the deadline, even if only to say when a full answer will follow
2. Send the original data or images the editor requests
3. Copy every co-author
4. Check whether your institution's research integrity office must be told
5. Keep every exchange in writing

Outcomes range from no action, or a published exchange of letters where no misconduct is proven, to a correction, an expression of concern or a retraction. Our guide to correcting the published record explains how each works.

If the criticism is wrong

Some comments are mistaken, and some are unfair. Either way, stay factual and let readers judge.

• Reply once, precisely. PubPeer notes that even when a commenter cannot be convinced, the wider audience evaluates the comment for themselves. If part of the critique is correct, say which part; it makes the rest of your answer more credible.
• Report comments that break the rules. Every PubPeer comment has a report button, and non-conforming comments are removed or edited.
• Do not try to unmask the commenter. PubPeer says it has been to court to defend commenters' anonymity.
• Do not rely on copyright takedowns. PubPeer says reproducing figures for criticism falls under US fair use and that it will vigorously oppose unfounded or vexatious takedown notices.
• Check where emails come from. PubPeer's alerts come from pubpeer.com, pubpeer.org or pubpeer.net; messages from other domains are not from PubPeer.

If you believe a public statement about you is defamatory, talk to your institution before doing anything in public.

Your paper is still being read

Post-publication scrutiny is not a verdict on your work; it is the literature doing its job. Authors who keep their source files and answer questions with evidence come through it with their credibility intact. Directive Publications' peer review process page explains how research articles are assessed before publication and how corrections, expressions of concern and retractions are handled afterwards. If a question about an article published with Directive Publications needs to reach the editorial office, use the contact page.