The scientist who uncovers errors in research articles admits that there is a significant amount of carelessness involved.
When Sholto David left his job last fall, he had several options. He could have searched for a new position, gone on a trip, or gone camping with his bike. However, David, a biologist residing in Pontypridd, Wales, dedicated his time and energy to a rather niche hobby: identifying errors in scientific publications and working to correct them.
According to David, the job is mostly unappreciated. Many scholars became defensive when faced with his critiques or simply chose not to address them. Editors of academic journals also tended to dismiss or ignore his letters, or would take an extremely long time to investigate them.
However, an announcement was made that caused a stir. The highly regarded Dana-Farber Cancer Institute, affiliated with Harvard University and located in Boston, revealed their intention to withdraw six research papers and revise 31 others. This decision was prompted by concerns raised by David in a blog regarding numerous studies conducted by prominent leaders at the institute.
David identified some additional papers that require attention. According to Dr. Barrett Rollins, the integrity research officer at Dana-Farber, one of these papers is still being investigated and three do not require any further action. Out of the remaining sixteen papers, the data was produced in external laboratories. The leaders of those labs have been contacted and steps will be taken to ensure that any necessary corrections are made to the literature.
David expressed that the decision came as a significant shock. He commended the Dana-Farber Cancer Institute for taking responsibility and making corrections and retractions. However, he also noted feeling disappointed by the situation. He mentioned that it is uncommon for institutions to take immediate action and that the investigation process can often be prolonged, while journals are slow to respond.
Per Rollins, the institute has a standard procedure of thoroughly reviewing any potential data errors and making necessary corrections. As a result, in 97% of the cases that David had identified, where the institute’s scientists were the primary authors, the institute had already taken swift and definite action. Rollins, who is also an author on some of the papers in question, had abstained from participating in any investigations related to them.
David, who completed his doctoral studies in molecular biology at Newcastle University, has always been intrigued by discrepancies in scientific research. Initially, he uncovered mistakes in systematic reviews, meta-analyses, and clinical trials, and brought them to the attention of the responsible scientists and the publishing journals. However, when his attempts were unsuccessful, he resorted to posting his findings on PubPeer, a platform where fellow scientists can provide feedback on published papers.
It evolved beyond a mere leisure activity. The biologist has reported approximately 2,000 articles on PubPeer, primarily due to suspicions of image manipulation. Initially, he manually detected instances by visually examining for duplicates and suspicious alterations such as stretched or cropped images, or spliced sections. Presently, he receives assistance from Imagetwin, an AI-powered program that cross-checks images against a library of over 25 million photos published in open access journals.
David stated that the reason for many mistakes in scientific research may be due to carelessness, and he believes that people have high expectations for the scientists conducting the research. When individuals contribute to scientific and cancer-related causes, they expect the researchers to have strict standards and be meticulous in their work.
A wave of withdrawals and amendments has occurred in response to a significant study in 2021 that revealed only half of the leading pre-clinical cancer studies could be replicated by researchers. In numerous instances, the results were significantly less significant than initially reported when the experiments were repeated. David suggested that this could be attributed to a lack of quality in the work being conducted, ultimately leading to a lack of reproducibility.
David refrained from accusing anyone of wrongdoing and emphasized that he did not want to create a hostile atmosphere for scientists where they would feel intimidated or afraid to publish their work. However, he also expressed concern about the repeated mistakes that were being made.
“How many mistakes can we overlook before considering a more concerning issue?” inquired the speaker. “While reviewing numerous papers, it’s inevitable to come across errors. However, there must be a threshold for the number of careless mistakes one can make before it becomes more than just an honest slip-up.”
According to Rollins, finding discrepancies in an image within a paper does not necessarily prove that the author intended to deceive. This determination can only be made after a thorough and fact-based investigation, which is a crucial part of our response. Our past experience has shown that these errors are typically unintentional and do not constitute misconduct.
David plans to go back to a traditional job within the next year or two. He explained that he is currently using his savings and may do some traveling this year. Despite appearing financially secure, he clarified that he is simply making do with his limited savings and taking advantage of low rent prices in the area.