Scientists whose work is frequently cited by other authors are generally seen as influential and productive—there are even rankings to recognize and celebrate them. But there is also evidence that not all of them adhere to the most rigorous standards of integrity. In a survey published in the journal PLOS Biology on January 30, epidemiologist John Ioannidis, an ethics expert from Stanford University, USA, found that at least 8,000 researchers on a list of highly cited researchers had a retraction on their résumé, meaning they have had at least one paper withdrawn after publication due to error, plagiarism, or data manipulation.
Ioannidis and his group reached this figure after combing through a database maintained by scientific integrity news website Retraction Watch, which had 55,237 records of retractions at the time of the study. They then compared this list of retractions with citation data from Elsevier’s Scopus database. The team discarded cases in which the retraction was due to factors unrelated to author behavior (such as problems in the editing process), as well as cases in which the articles were republished with corrected data or were not indexed in Scopus. They were left with 39,468 papers to evaluate.
Finally, the authors’ names were cross-referenced against two different segments of a list of highly cited scientists known as the “Stanford List,” extracted from a database maintained by Ioannidis himself since 2016. One of the segments, containing 223,152 names, featured authors from among the most cited 2% in the world in their fields in 2023—8,747 of them, or 4% of the total, equivalent to one in every 25, had had an article retracted. The other segment featured 217,097 names from the 2% most cited throughout their careers: 7,083 scientists, or 3.3% of the total, had suffered a retraction. There were 1,199 and 957 researchers who had had two papers retracted in the two segments respectively.
According to the study, retractions were most common in the life sciences, with clinical medicine and biomedicine leading the way, each accounting for around 5% of all retracted articles. There were very few papers retracted in subfields of disciplines such as the arts, humanities, and social sciences, probably because they more often publish results in other formats, such as books and reports. In absolute numbers, the USA led the way with 2,332 retractions of articles by authors among the 2% most cited over the course of their career, followed by China (877), the UK (430), Japan (362), and Germany (336). In terms of the proportion of highly cited authors with a retraction from each country, the highest percentages were observed in Senegal (66.7%), Ecuador (28.6%), and Pakistan (27.8%). “This may reflect problematic research environments and incentives in these countries, several of which are also rapidly growing their overall productivity,” Ioannidis wrote. Of the countries with a high level of scientific production, the highest proportions were seen in India, where 9.2% of highly cited researchers have had a retraction, then China (6.7%), Taiwan (5.7%), and South Korea (5.3%). Brazil, which historically has had fewer retractions than other nations, is not mentioned in the study.
The PLOS Biology article did not look into the causes of the phenomenon, but it did spark a debate about what might be going wrong for highly prestigious scientists. “The analysis provides a bit of weak circumstantial evidence that this is a real issue and that we should look carefully at authors with high citation rates, and not just assume they are great researchers,” neuropsychologist Dorothy Bishop, an expert in scientific integrity and retired professor from the University of Oxford, UK, told the journal Nature.
Ioannidis was measured in his interpretation of the results. “Caution is needed when interpreting retractions, as they do not always signify misconduct. Further analysis on a case-by-case basis is essential,” he wrote in the paper. It is natural, he points out, for prolific scientists to have more retractions—since they publish more, there is a greater chance of there being a problem in one of their papers than among colleagues who publish less. According to the study, scientists with retracted articles had a greater number of publications and a higher number of coauthors than others. Ioannidis emphasizes the importance of having a broad overview of the characteristics and vulnerabilities of the most influential scientists in various disciplines and places around the world. It was for this reason that he included data on retractions for each author mentioned in the most recent version of the Stanford List, published in September 2024.
Ivan Oransky, cofounder of the website Retraction Watch, told Nature that the results highlight the problems with ranking scientists based on citations. He says the study data is not surprising, especially with respect to countries that reward scientists based on academic impact indicators. “My guess is that this behavior is somehow tied to citation manipulation.” Oransky’s assumption is corroborated by the coordinators of another annual list of highly cited researchers, published each year by the company Clarivate Analytics. To ensure the credibility of its list, Clarivate analyzes each case individually to weed out any attempts at manipulation.
Researchers who abuse self-citation or make citation agreements with other authors—methods commonly used to artificially expand scientific impact indicators—are excluded. Any being investigated for misconduct at their institutions are also removed, as are those who have had articles retracted, and even anyone who has signed papers with hundreds of coauthors, since it is impossible to assess the individual contribution of each author in such cases (see Pesquisa FAPESP issue nº 323).
The progressive addition of these filters has caused the number of people disqualified from the list to grow each year. In 2021, just over 300 scientists were removed due to behavior linked to the manipulation of performance indices. The figure was 550 in 2022 and just over 1,000 in 2023. Last year, a record of 2,000 names were eliminated—from a list of 6,600 researchers. “We continue to refine our evaluation and selection policies to address the challenges of an increasingly complex and polluted scholarly record,” said David Pendlebury, Clarivate’s head of research analysis, in a statement.
Republish