Paper mills: see the wood for the trees (Part 2)

Ten signs you might be dealing with a “paper mill” product

How to detect fabricated and manipulated submissions

Working with internal and external editors in the most affected fields, who have become adept at weeding out suspected paper mill submissions, we’ve gathered this list of characteristics of such works to help others in detecting them. While there are sometimes clear signs there is a problem with a submission, more often suspicion comes from several clues that, while individually might not be a cause for concern, signal a problem when combined. Furthermore, the problem might not be apparent when looking at an isolated paper; it is only when the same characteristic is seen across multiple papers, across multiple journals, that the full picture begins to emerge.

In the game of “cat and mouse” against paper mills, it would be counterproductive to publicly share every detail, but here are a few better-known tells that may help you spot problematic papers before publication. If you are an editor in a field that is currently targeted by paper mills, notably the biomedical sciences and the physical sciences, this is what we’d suggest looking out for:

1. Duplicate (parts of) submissions Work submitted to multiple journals simultaneously or in close succession, as well as identical or similar submissions, or elements of submissions e.g., cover letters and conflict of interest statements, made to the same journal from different author groups, are all hallmarks of paper mills. Duplicate submissions, or elements of submissions, to the same journal are shown in the Editorial Manager duplicate submission check, but – more significantly – Elsevier has recently rolled out a check for potential duplicate submissions between journals across most of our journals.

2. High textual similarity, for example in Crossref Similarity Check Paper mill submissions sometimes show high overall similarity with previously published works from other author groups, that are not cited. The similarity is usually detected throughout the manuscript. However, cases have been seen where translation software is used to translate and then reverse translate texts to evade detection by software.

3. “Cookie cutter” manuscripts The same structure and formatting time and time again, from different author groups, may suggest a paper mill operation is behind the submissions.

4. Submissions making no attempt to adjust to the journal’s style or formatting requirements While authors are not always required or expected to provide a cover letter or meet journal formatting requirements in their first submission, paper mill submissions tend to make no attempt to customize for the journal in question.

5. “Pick from the dropdown options” titles

   A subset of paper mill submissions correlate a gene, protein or drug with human disease, often cancer. In these cases, formulaic titles are common(abre em uma nova guia/janela). While this does not always indicate a problem, in particular for papers written by authors for whom English is not a first language, in combination with other signals, it may indicate a paper mill manuscript.

6. Stock images, graphs and tables Identical images, graphs and tables, perhaps with modifications to axis titles, for example, can crop up across paper mill submissions on completely different research topics and hypotheses.

7. Suspicious experimental results Especially for experiments involving western blots, flow cytometry data and microscopy, the results in paper mill submissions have often been duplicated or manipulated e.g., very neat, cropped, high-contrast western blot bands. Incorrect nucleotide sequences and errors in reagents are also common. When coupled with authors who are unable or unwilling to share raw data, when requested, suspicions are raised even further.

8. Authorship changes Author and/or email changes at revision or post-acceptance, flagged in the details screen, may indicate the authorship is not genuine and that authorship may have been sold.

9. Suspicious email addresses Alphanumeric emails, bearing no resemblance to author names or institutes, with non-institutional domain names, and which appear to have been used rarely before i.e., are not registered with the journal and rarely or never show up in “hits” for other papers or profiles, indicate they may have been created specifically for the paper.

10. Suspicious reviewers Reviewer reports that are similar in structure and content, offer very positive comments, and that are returned rapidly can signal something untoward. If these are provided from author-suggested reviewers, repeated across several submissions, or are from recent eager volunteers on specific topics, and their email addresses display the same red flags described in the preceding point, you might have a fabricated review on your hands, designed to give the paper mill product an easy route to publication.

We encourage you to bear in mind the signals mentioned above as you consider submissions, to learn from the experiences and best practices of other journals, and to take advantage of the new tools and services that we make available on an ongoing basis to help you validate the trustworthiness of submissions.

Read part 3: Practical steps in pursuit of paper mills (coming soon!) to hear what actions Elsevier and some of our journals are taking.