Paper mills and the erosion of research credibility: Researchers beware
If you asked a researcher about “paper mills” 20 years ago, they would envision a facility where wood pulp is processed to produce paper. But the paper mills we are talking about here are something altogether different.
Paper mills in academic publishing are profit-oriented, dubious organizations that use sophisticated marketing tactics and professional-looking websites to attract researchers seeking to expedite publication. They sell authorships, ghostwritten papers on nonexistent research, and fake data and images to researchers and guarantee publication in reputable journals. Career advancement often requires academics to churn out papers rapidly. Add to that the high rejection rates of top-ranking journals, and therein lies the nudge towards falling for predatory publishing practices—on purpose or unwittingly, it can be hard to tell.
This piece aims to raise awareness about the phenomenon, tracing its origins and modus operandi and bringing to light the red flags to be wary of.
A brief history of paper mills
“Term paper mills” were documented in the 1960s, and these targeted undergraduate students to “buy” readymade term papers. While the exact origins of paper mills in academic publishing are not clearly defined, their proliferation can be linked to the digital age and the ease of communication and information sharing online. The growth of the internet and online publishing platforms has made it easier for individuals or organizations to set up paper mills and offer services that cater to researchers looking to boost their publication record quickly.
In 2013, Science reported the existence of “publication bazaars” in China, where authorship was sold to researchers (in one case, first co-authorship was up for grabs at $14,800). The experiments described may or may not have actually been performed. Some of these paper mills may produce actual data, but then sell the same images and results to multiple authors reporting different experiments. Three years later, Science again reported on papers for sale, this time in Iran. Another disturbing finding was that young academics struggling to find employment often ended with jobs in the paper-selling business.
Jana Christopher reported a set of 12 manuscripts submitted to FEBS journals that contained suspiciously similar western blot images with regularly spaced bands, lack of background noise, and identical backgrounds, although the manuscripts came from different research groups at different institutions. Around the same time, Jennifer Byrne and her team raised concerns about fraudulent gene knockdown studies appearing to come from paper mills, sharing highly similar figure layouts and orders.
Elisabeth Bik, a microbiologist-turned research integrity sleuth, began posting about image duplications in 2013 and took up identifying and reporting potential research misconduct full time in 2019. In their quest to trace fraudulent practices, Bik other image forensics experts uncovered a peculiar phenomenon: hundreds of papers from different authors and affiliations, all suspiciously similar and appearing to have been “generated” by a single source (à la tadpole-like western blot bands and Death Star flow cytometry plots).
Retractions galore
These exposés highlight systematic efforts by paper mills to produce fraudulent scientific papers through data fabrication and image manipulation. Worryingly, many of these papers get published successfully, eventually making their way into the reference lists of credible papers. To put a stop to perpetuating this fraud, journals have been initiating retractions. After finding astounding similarities in texts, charts, and titles, the Royal Society of Chemistry announced the retraction of 68 papers from three of its titles for being products of paper mills. Early this year, the publisher Taylor and Francis had 80 papers retracted from their journal Acta Agriculturae Scandinavica, Section B — Soil & Plant Science for the same reason. As recently as last month, the Scottish Medical Journal retracted 13 papers because they showed “indicators of paper mill activity.”
While whistleblowers, sleuths, and discerning readers unmask these alarming cases, the papers may take years to be retracted, persisting in the literature and misinforming researchers for long.
Hallmarks of paper mills for authors
Awareness of the warning signs can help authors, editors, and peer reviewers identify problematic manuscripts from paper mills. This awareness is crucial for safeguarding unsuspecting authors and readers and for accelerating retraction efforts.
Red flags for authors
Authors should be aware that authorships cannot be purchased, neither can data, images, or entire manuscripts. Tempting as these offers might sound, authors must disregard—or possibly, report—invitations to such scams (e.g., see below).
Source: https://twitter.com/acochran12733/status/1786773475933225044
Young researchers should also be careful to not get picked up by such organizations that might offer a role to play in these deceitful practices in the guise of employment.
Red flags for peer reviewers and editorial staff
Thousands of mass-produced papers pass peer review and editorial screening. A higher level of awareness of the tell-tale signs of papers coming from paper mills will enable swifter action upon suspicion or detection of paper mill activity.
Here are some hallmarks of mass-produced papers (based on findings by Bik, Byrne and Christopher, Gaby, etc.):
1. Images that are a bit too perfect
- Identical or suspiciously similar images used across different papers from different authors/institutions.
- Clean western blots with perfect bands and no background noise or artifacts.
- Reused stock images across unrelated papers.
2. Déjà vu: Highly similar formats
- Unusual levels of textual similarity across papers from different authors/institutions, suggesting the use of manuscript templates.
- Papers from different labs following eerily similar figure layouts, order, and formatting.
3. Boilerplate hypotheses
- Superficial or generic hypotheses being advanced to justify analyses on different topics, lacking specific reasoning.
4. Problems in methodological details
Incorrect or misidentified descriptions of nucleotide sequences, cell lines, models, or reagents, potentially due to replacement or copying from other sources.
- Lack of disclosure or omission of important reagent details, nucleotide sequences, and raw data supporting the reported experimental results.
5. Unrealistic productivity
- Unusually high number of publications from the same authors/institutions in a short time.
- Trials with an unusually large number of participants and/or implausibly short recruitment period for the trial.
- Submission to a journal unusually rapidly after the study is completed, or in some cases before it would have been possible to have completed the trial.
- Unbelievable magnitude of a reported improvement, much larger than is typically seen in trials.
6. Subpar writing quality
- Poor writing quality with unclear methods, inconsistent data presentation, or nonsensical text.
- Papers littered with “tortured phrases” as a result of the use of probabilistic text generators (e.g., “bosom peril” for breast cancer and “profound neural organization” for deep neural network!)
7. Problematic affiliations
- Authors affiliated with companies/institutions without research programs or from multiple disparate locations.
- No funding source is listed or the study is listed as self-funded, which is particularly fishy when the sample size or study design suggests that the study was expensive.
Addressing the paper mill crisis
The Committee on Publication Ethics (COPE) recently issued a new position statement on paper mills and announced their support towards the United2Act consensus statement on action against paper mills. COPE has also recommended several actions, such as training journal editors and editorial staff in identifying fake papers, development of tools to identify suspicious manuscripts, a shift in incentives for researchers to favor quality over quantity, and streamlining the retraction process to adapt to paper mill papers.
With advancements in AI, paper mills might be harnessing sophisticated tools to make their outputs more believable. Traditional stock photos might become a thing of the past with AI-powered image generation. GPT-like tools are making it easier to fabricate manuscripts with natural-sounding text. As paper mills become better at producing convincing outputs, we urgently need technology-based solutions. Last year, a new tool was launched by the STM Integrity Hub to detect products of paper mills: Clear Skies Papermill Alarm. This alarm system provides a simple color-coded evaluation of research papers based on similarity to content known to be produced by paper mills.
Concluding notes
Paper mills not only enable academic misconduct but also devalue the hard work of legitimate researchers. This complex problem requires the collective effort of the entire academic community. Addressing the paper mill issue involves technological solutions, institutional policies, and a cultural shift towards upholding academic honesty.
It is encouraging to see concerted efforts from academic institutions, publishers, and technology companies to tackle this menace head-on. However, in the war against unscrupulous practices that prey on the competition and pressure faced by authors, vigilance and unwavering resolve from the authors themselves are particularly crucial.
Comments
You're looking to give wings to your academic career and publication journey. We like that!
Why don't we give you complete access! Create a free account and get unlimited access to all resources & a vibrant researcher community.
Subscribe to Conducting Research
