Peer review runs on trust: an editor sends a confidential manuscript to a stranger and believes the response comes from the expert they invited. When that assumption fails, everything downstream fails with it. Over the past decade, publishers have retracted thousands of articles because the reviews that approved them were written by the authors themselves or by services selling favourable reports. It is one of the most damaging integrity problems in publishing — and one authors can help prevent.
How the fraud works
- Fabricated reviewer identities. The classic method: an author suggests genuine, well-known experts but supplies email addresses they secretly control. The invitation arrives in the author's own inbox, and a favourable review is returned within hours.
- Compromised or impersonated accounts. Real reviewer profiles in a submission system are hijacked, or near-identical addresses are registered to impersonate a known researcher.
- Review rings and paper mills. Organised networks agree to review one another's submissions favourably, or sell packages that include manuscripts, authorship slots and cooperative reviewers.
- Coerced or traded reviews. A reviewer demands citations to their own work, or a group reciprocates approvals.
An invented example of the classic method
The case below is invented to show the mechanism. An author suggests three reviewers. Two are listed with their university addresses. The third, a senior professor at a large university, is listed with a free-mail address the author created that week. The editor invites all three. The "professor" replies within a day, recommends acceptance and praises the methods in general terms. The real professor never saw the manuscript.
One step would have stopped this: inviting the professor through the university's staff page.
Why it is hard to detect
A fake review often looks plausible: competent, moderately critical, and delivered promptly. The signals are indirect — a free-mail address for a professor at a major university, a review returned in two hours, unusually uniform phrasing across several reports, or reviewers who exist but have no publication record in the field. Individually these are weak; together they form a pattern, which is why detection usually happens at the publisher level, across many submissions.
Signals and the check that settles each
| Signal | Check that settles it |
|---|---|
| A suggested reviewer's address is free-mail or on an unfamiliar domain | Find the reviewer's institutional page or ORCID record and invite them through the address listed there |
| The name matches a real expert, but the address differs from the one on their recent papers | Write to the address on their most recent paper and ask them to confirm |
| A report arrives within hours and recommends acceptance without specific comments | Look for evidence the manuscript was read: named tables, figures, methods or references |
| Near-identical wording appears in reports on different submissions | Compare the report text across the journal's recent submissions |
| The same suggested reviewers appear on unrelated manuscripts from different authors | Review those submissions together before making any decision |
What a serious journal does
- Verifies reviewer identities independently — institutional email addresses, ORCID iDs, and publication records, rather than relying on contact details supplied with the submission.
- Never uses suggested reviewers exclusively. Editors add reviewers they sourced themselves.
- Watches for behavioural signals — implausible turnaround times, duplicated review text, and clusters of submissions sharing the same suggested reviewers.
- Investigates and corrects the record when fraud is found, following COPE guidance — which can mean retraction even where the science is sound, because the acceptance decision was corrupted.
COPE publishes two flowcharts on this problem: one for peer review manipulation suspected during review, and one for manipulation suspected after publication. Both are free on the COPE website. Before publication, reports from an unverified reviewer cannot support a decision. The manuscript goes to independently sourced reviewers or is rejected. After publication, the journal first checks with the named reviewer, then explains the concern to the authors and their institution and checks whether the other reviews were adequate. The outcome can range from leaving the article as it stands to a correction, an expression of concern or a retraction.
What honest authors should do
Suggest reviewers, but suggest them properly: real experts, no close collaborators, and always a verifiable institutional address — never a personal one you can access. Decline any offer of "guaranteed acceptance", "review assistance" or authorship for sale; these come from the same ecosystem. And treat a journal that accepts a paper in days without substantive comments as a warning, not a convenience — that speed is often the sign of a review process that did not happen.
Checklist before you submit reviewer suggestions
- Every address matches one published on the reviewer's institutional page or recent papers.
- No address is one that you, a co-author or an assistant can log into.
- No suggested reviewer has published with any author within the journal's stated window.
- You have not contacted any suggested reviewer about the manuscript.
- Every co-author has seen the list before it is submitted.
Our guide to suggesting reviewers for your manuscript explains who qualifies and gives a template for each entry.
What to do if you are affected by review fraud
Your name was used as a reviewer
You may learn of it from a colleague or from an editor checking a report. Do not reply to any message that came from the suspect address. Find the editorial office address on the journal's own website. Send the manuscript title or number, the date and the address that was used. Ask the journal to record that the report was not yours.
A co-author supplied the reviewer details
A retraction for review manipulation applies to the whole article, so it carries every author's name. If you see a suggested reviewer with an address you cannot verify, raise it before submission. If you find out afterwards, tell the handling editor in writing. Reporting it yourself is far better than being found out in an investigation.
You were offered help with reviewers
Keep the message, including the sender's address. If it names a journal, forward it to that journal's editorial office, using the address on the journal's own website. The same networks sell authorship slots; our guide to paper mills and authorship for sale lists the related offers to refuse.
Updated 17 September 2026: We expanded this guide with an invented worked example, a table of fraud signals and checks, steps for affected researchers and an author checklist.