What?

(Paper) mills are fraudulent companies that produce and sell scientific manuscripts or derivatives thereof. For example, they offer unjustified authorship of a (often fake) paper, you can buy a fake (positive) peer review or extra citations to falsely increase the impact of your work. You can even buy completely fabricated articles, often based on fraudulent data or plagiarised research.

You can read more about what (paper) mills are and how they work on the intranet. 

In addition to researchers who consciously participate in (paper) mills activities, researchers also become unknowingly involved, e.g. because their name, prestigious status or institution is used as a kind of bait to entice others to buy. Both cases (conscious or unconscious) involve numerous violations of research integrity:

• identity fraud, whereby the names of bona fide and unsuspecting scientists are misused
• unlawful authorship, because no contribution was made
• unlawful use of data, text or images without reference to the original (plagiarism)
• data falsification and fabrication
• image manipulation, whereby the same images are used in different unrelated studies or where the data/images are completely fabricated and/or manipulated
• ... . 

 

How can you recognise mills activities? 

Some characteristics of mills activities can only be detected (and investigated) by editors. However, 'regular' researchers can also identify “red flags” (Parker et al, 2024). The presence of (one of) the characteristics below does not guarantee that the source is fraudulent, but caution and extra verification are required if you want to use and/or integrate these sources in your own work. 

 

Regarding the content:

• Description of very generic study hypotheses and experimental approaches.
• Often in the field of cancer or rather unusual or little-known topics.
• Frequent occurrence of references that have little or nothing to do with the topic, especially when the same references appear in different papers.
• No reference to funders. 
• Frequent use of so-called “tortured phrases”:

From a presentation by Lisa Parker, WCRI 2024.

 

• Paper is not indexed in a bibliographic database.
• DOIs used do not link to existing papers.

 

Regarding data and visualisation:

• Broad datasets, which are almost always poorly managed, labelled or incomplete.
• Data does not fully correspond to the types of experiments described in the methods.
• Requires specialised knowledge or skills, either to open the datasets (specialised software) or to read them (specialised knowledge or expertise to make sense of them).
• Manipulated or edited images (cropped, rotated, duplicated)
• Frequent use of stock images (too “neat” background, unlabelled Western blots, scatter plots that look too “neat”, etc.)

 

Regarding authorship:

• Email addresses without institutional format or link to the authors' names.
• Less obvious collaborations between authors who (1) may not know each other; (2) do not have a common research interest; (3) are affiliated with different universities; (4) specialise in very different disciplines; and (5) are not specialised in the topic of the paper. E.g. a collaboration between an economist and a physician where the paper is not about the economics of healthcare (Abalkina, 2023).
• Papers with only one author, in disciplines where this is not common, e.g. biomedical sciences. It is possible that the other slots were not sold or that one author bought all the slots (Abalkina, 2023).
• Extremely high number of changes in authorship, often in different rounds of peer review, even after the article has been accepted for publication. 

 

Individually, these factors are not always (immediately apparent) problematic, but together they can form a pattern that points to milling activities. Mills are also becoming smarter and sometimes limit the number of papers submitted to journals each year, not to arouse suspicion. Or they spread the number of submissions across many different journals and publishers. This makes them more difficult to detect (Abalkina, 2023).

Increasing efforts are being made to make information about suspicious papers available more quickly. Platforms such as PubPeer and Retraction Watch play an important role in this. The tools are now increasingly being integrated into databases (such as Web of Science) and reference software (such as EndNote and Zotero), and users can also receive notifications themselves when papers are contested via plug-ins.

If you would like to know more about how to quickly find retracted or contested publications, read the research tip: Resources: where to find information on retracted or contested publications?

 

As a researcher, how do you deal with mills activities?

• Do not respond to advertisements for fraudulent products and services. United2Act has compiled the following overview to help you recognise them:  

Figure by United2Act, Paper mills: the essentials 

 

• Discuss mills activities with colleagues: many researchers are not yet aware of the problem.
• Search your name and publications: are they being used unlawfully?
• Remain critical of scientific work and always carry out a number of checks if you want to use papers:

▪ Verify the identity of the authors: Is there an ORCID?  Do the contact details match those stated in the paper? Can you find them online?

Please note! Mills often link existing names to fake email addresses, which makes it difficult to contact them and verify the actual situation, because you never know who is on the other end of the email. It could be a representative of the paper mill who uses all and different addresses.

▪ Check the affiliation to verify whether authors belong to the institution listed.

▪ Does a researcher's publication list (e.g. via their ORCID profile) include the article in question?

▪ Screen for (some of) the above-mentioned issues that indicate mill activity. Does anything seem out of the ordinary?

▪ Is the underlying data available for consultation?

▪ Check the references. Can you find them and are they relevant to the topic?

▪ Check whether the paper/author is mentioned on PubPeer or Retraction Watch or, better still, make sure you receive automatic notifications of contested papers.

▪ Check the status of the article on the journal's website: if Mills activity is suspected, the editor may add a comment to the paper, known as an “Editor's Note”. This note is often only visible on the website and is not indexed, which means that readers who access a paper via other platforms will not see it. This is different from a Retraction Notice (when mill activity has been confirmed), which is a permanent document with its own DOI, indexed and linked to the paper (Parker et al, 2024).

▪ Want to look further? Check papers by the same author, similar papers in the same journal (using the same keywords, title, topic, etc.).

 

• By applying good research practices, we indirectly make it more difficult for mills. When the standards of good research are clear and adhered to, deviations from them will be more readily apparent and less tolerated. 

▪ Always use your institutional email address and your ORCID.

▪ Apply the principles of Open Science in your work.

▪ Use preprints, preregistration and registered reports.

▪ Make raw and processed data available (with metadata and documentation).

 

• Do you have a well-founded suspicion of mills activity? Contact the editor of the journal, the editor-in-chief and, if available, the in-house research integrity team. Share well-founded questions and doubts on platforms such as PubPeer and Retraction Watch.
• If you need help, contact the secretariat of the Committee for Research Integrity or the faculty contact points research integrity.