6.0 KiB
| title | chunk | source | category | tags | date_saved | instance |
|---|---|---|---|---|---|---|
| Research paper mill | 1/1 | https://en.wikipedia.org/wiki/Research_paper_mill | reference | science, encyclopedia | 2026-05-05T04:28:30.198721+00:00 | kb-cron |
In research, a paper mill is a business that produces poor quality or completely fraudulent journal papers that seem to resemble genuine research, as well as sells authorship on such papers. Paper mills are an example of academic dishonesty that affects all fields of academic publishing including academic writing, scientific writing and medical writing, and represents a failure of research ethics and research integrity.
== Activities == In some cases, paper mills are sophisticated operations that sell authorship positions on legitimate (but poor quality) research, but in many cases the papers contain fraudulent data and can be heavily plagiarized or otherwise unprofessional. These services are advertised on platforms such as Facebook, Telegram and WhatsApp. Activities include contract cheating, academic ghostwriting and medical ghostwriting. It may include data fabrication, leading to junk science and retractions in academic literature.
== History and prevalence == Research paper mills first emerged as a significant problem during the 2010s, though the earliest known papers attributed to paper mills are known from the 2000s, at least as early as 2004. In 2013, Science published an investigation of various kinds of research misconduct in China, including noting the practice of buying papers from an online catalogue. In 2014 COPE wrote an editorial raising concerns on agencies selling services, including authorship of pre-written manuscripts. According to a 2021 report in Nature, thousands of papers in academic journals had been traced to paper mills from China, Iran and Russia, and some journals were revamping their review processes." Chinese researchers have been identified as particularly prevalent customers of paper mill services. Differing estimates put the share of paper mill productions between 2% and 20% of published academic papers, with particularly severe problems in some areas of biomedicine. A 2026 study estimated using machine learning analysis that approximately 10% of recently published literature of cancer may be the product of paper mills, rising to over 20-22% for liver, gastric and bone cancer, with the number of published suspect papers haven risen exponentially from essentially zero flagged papers in 2000. China was found to have the proportionally highest number of flagged papers with 36% of the total Chinese output being considered suspect, followed by Iran with 20%, Saudi Arabia with 16%, Egypt with 15%, and Pakistan and Malaysia with 13% each. The apparent prevalence of paper mills in China has been attributed to the heightened "publish or perish" pressure placed on academics and other scientific professionals in China, particularly on medical doctors and nurses, who are required to publish academic papers in order to receive promotions despite research not being a significant part of their job.
== Detection == A 2024 peer-reviewed forensic study showed that provenance-based image analysis can automatically cluster manuscripts that originate from the same paper mill, providing scalable evidence of systematic production.
== Examples ==
In early 2022, Times Higher Education and the Science Magazine News department covered a report exposing a Russian paper mill company International Publisher Ltd. The report identified hundreds of published academic papers where positions for authorship had been sold through a Russian website allowing researchers to pay for academic prestige without requiring legitimate research contributions. During the three-year period analyzed, 419 articles were identified that were matched to manuscripts later published in many different academic journals, with a significant bias towards publications in predatory journals. While the paper mill targeted various journals, almost 100 papers were published in International Journal of Emerging Technologies in Learning (Kassel University Press) alone, seemingly coordinated through the involvement of journal editors hosting Special Issues with space for coauthors auctioned off for anywhere from $180–5000 USD. In a separate network, guest editors and salaried academic editors for MDPI were found to coordinate sale of authorship across four different MDPI journals, totalling over 20 papers (picture, right). Beyond collusion between editors and International Publisher Ltd., many legitimate research papers also sold authorship unknown to the journal editors, and were ultimately accepted in journals published by Elsevier, Oxford University Press, Springer Nature, Taylor & Francis, Wolters Kluwer, and Wiley-Blackwell. As of April 6, 2022, many of these publishers have opened an investigation into the matter. In May 2024, the Wall Street Journal published a report on fake studies that affected New Jersey publisher Wiley. More than 11,300 papers were retracted, and 19 journals were reportedly closed. The problematic papers were linked to Hindawi, an Egyptian publisher of about 250 scientific journals that Wiley acquired in 2021. The article detailed how the research paper mill fraud worked, and highlighted individual efforts to identify and prevent future fraud. The article also warned that artificial intelligence was going to make fraud more difficult to detect. By analysing duplicated images, Richardson et al hypothesize that paper mills may have used a large bank of images for mass publications and may have collaborated with editors in targeted journals such as PLOS ONE and Hindawi for expedited publications. Paper mills can switch journals quickly if a partner journals get deindexed from literature aggregators. The researchers cited "Academic Research and Development Association" (ARDA), based in Chennai, India, as an example of such a paper mill.
== See also == Academic mill (disambiguation) Diploma mill Essay mill Research Integrity Risk Index Predatory publishing Publish or perish
== References ==