A paper mill is a business that sells authorship, ready-made manuscripts, citations or favourable peer reviews. Authorship for sale means paying to be named on work you did not do. Refuse every version of the offer: a byline slot, a ghost manuscript, guaranteed acceptance, a cooperative reviewer. Detection reaches these papers years later, and a retraction names the authors.
What a paper mill sells, and why authorship for sale is the product
A paper mill is not selling research. A paper mill is selling a line on your CV, and that line has to survive a glance from a committee: journal name, date, DOI, co-author list. So the offer is worded as a service, borrowing the vocabulary of language editing and statistical consultancy, then adding what no legitimate service supplies: a place in the author list, or a promise about the decision.
A service may work on your manuscript. A service may not be, choose or influence your authors, your reviewers or your editorial decision. ICMJE criteria set four requirements that must all be met: a substantial contribution to the work, drafting it or revising it critically, approval of the final version, and agreement to be accountable for it. None can be bought, and a paid contributor who does not meet all four belongs in the acknowledgements. The CRediT taxonomy shows where each contribution lands.
One distinction matters, because paper mill wording blurs it deliberately. A publisher invoicing its own article processing charge after acceptance is normal. A third party charging you for an acceptance, or pricing a position in an author list, is a different transaction: what is being sold is a decision, or a byline.
Five offers, how they are worded, and what each one costs your record
Each row is an approach type, not a quotation from any seller. The wording drifts; the structure does not. Read the fourth column first: detection decides how the offer ends.
| Offer type | How it is worded | What is actually sold | How it surfaces later | Consequence for you |
|---|---|---|---|---|
| Authorship slot | "Positions 2 and 3 available on an accepted oncology article. Price on request." | A byline on a paper you never read, by people you cannot name | Author lists sharing no institution, funder or prior work, and co-author combinations recurring across unrelated fields | Retraction naming you, an institutional case, and every paper you co-authored re-read |
| Manuscript to order | "We prepare the complete paper in your specialty, data and figures included." | Fabricated or recycled text, results and images, resold so the same material can appear under different author lists | Image reuse across unrelated papers, templated sentences, impossible statistics, and no raw data | Retraction for fabrication, and no way to defend data you never collected |
| Guaranteed acceptance | "Acceptance guaranteed. Fifteen-day publication. Special issue slots remaining." | A bypass of peer review, at a venue the seller controls or has infiltrated | No review record exists when the article is questioned, and a delisted venue's content is re-examined in bulk | Your work sits inside a corpus nobody trusts, and cannot be moved elsewhere |
| Cooperative reviewers | "Suggest our colleagues as reviewers and we will suggest yours. Draft reports supplied." | A rigged review, run through supplied names and contactable addresses | Reviewer addresses that do not match the named person's institution, and reports more alike than independent ones | Retraction for compromised review, which sits on the article and so on your publication record |
| Citations for sale | "Add these references at revision and your recent paper will be cited in return." | Manipulated reference lists and reciprocal citation | Reference patterns that make no argumentative sense, and citation clusters visible in open Crossref metadata | Correction or retraction, and a question mark over the metrics the offer promised to raise |
What publishers screen for when a submission may have been bought
Screening does not look for guilt. Screening looks for a mismatch between the paper, the people and the public record, then asks the authors to explain it. The checks below are ordinary publishing practice, described in general terms; what we do here is set out at the end of this post. COPE treats paper mills as a research integrity matter and publishes guidance for editors.
- Author identity and affiliation. Institutional email domains, ORCID records, ROR-registered affiliations, and whether the named authors have prior work connecting them.
- The correspondence trail. Whether the corresponding author answers substantive questions about the methods, or routes every reply through a third party. The corresponding author's responsibilities cannot be handed to an agent.
- Text and image checks. Similarity screening on submissions, plus inspection for duplicated or reused image panels.
- Reviewer provenance. Where a journal accepts suggested reviewers, the names are checked against independent sources rather than reached at the address supplied, because a supplied address is how a fabricated reviewer receives the manuscript.
- Data on request. An author who ran the study can produce the raw file, the protocol and the approval letter. A buyer cannot.
A decision rule for the moment an offer arrives
Run these seven steps in order. No step asks you to judge whether a message is genuinely a paper mill: the safe action is identical either way.
- Do not reply, not even to decline. A reply confirms a live address and marks you as someone who reads these offers.
- Do not open attachments or follow links, and do not send your CV, your ORCID iD or a draft for comment.
- Preserve the evidence. Save the message with full headers and archive any page it points to.
- Report it to your institution — the research integrity officer, or the research office. That record protects you if your name later appears on work you did not write.
- Tell the journal, if one is named. A journal used as bait has agreed to nothing, and may not know its name is in use.
- Check whether your name is already in use. Search your own name and ORCID iD for outputs you did not write, and set a standing alert.
- If a colleague or supervisor proposes it, refuse in writing and keep the dated copy. A verbal refusal leaves no evidence that you refused.
Send the text below to your integrity officer, and a version of it to any journal the offer names.
Subject: Report of an unsolicited authorship offer
I received the message below on [date]. It offers [a paid position in the author list / a prepared manuscript / guaranteed acceptance / reviewer arrangements] and names [journal or institution]. I have not replied and have sent no material.
I am forwarding it so there is a dated record that I received and declined the offer. The original message with full headers is attached.
Please tell me whether you need anything further, and whether my name should be checked against submissions I did not make.
[Name, ORCID iD, department]
What accepting an offer costs years later, and what to do if you already have
The cost of buying authorship is deferred, which is what makes it tempting. Detection is retrospective: checks that did not exist at publication are applied later, and a whole corpus is re-examined once one paper inside it fails. There is no time limit on a retraction, and an institutional inquiry can be opened years after publication, subject to whatever limitation period the jurisdiction applies.
What survives is the record, not the explanation. A retraction notice is registered against the original article's DOI through Crossref, states the reason, and reaches the databases and reference tools that ingest that metadata. "I was added without knowing" is a hard case to make from inside an author list you agreed to join.
If you have already accepted an offer, act before publication rather than after. Withdraw the manuscript, tell the editor what happened, and tell your institution before somebody else does. Self-reporting is the only version in which you are the person who stopped it.
Where Directive Publications stands, and the page we have not written
Our position is published. Review is double-blind, at least two independent expert reviewers are sought for each research manuscript, the handling editor decides, similarity screening runs on submissions, and our policies align with COPE Core Practices. Manuscripts reach us by email to the editorial office, and no third party is authorised to place or fast-track one.
The honest gap: we have no separate published paper-mill policy page. The commitments that bear on it sit across Research Integrity, Publication Ethics and Peer Review. We would rather say that than send you to a policy that does not exist. Two figures we will not give: how many bought submissions we receive, and how many we catch. Neither is knowable, and both are routinely invented.
If an offer of authorship for sale has reached you, the sender believes you are under enough pressure to consider it. That pressure is real, and it is not a character flaw. Read how to identify a predatory journal before choosing a venue, and how to handle a rejection when the honest route stalls.