In the scientific world, gold medals can be bought, like a common pipette, from unscrupulous actors. This is what a team from New York University, with its branch in Abu Dhabi (United Arab Emirates), first shown in a preprint currently under review by a journal, reported by Nature in August. She called on a company that does not sell the “medals” directly, but a doping product allowing to obtain, namely citations. The latter are references to authors and their articles that have been useful for one’s own work. They are indicated in the bibliography. In return, one hopes that one’s text will be cited by others, establishing one’s reputation.
The count of citations has become one of the major criteria for evaluating researchers, journals or research institutions, while also attracting its share of criticism. “Tricks” for inflating one’s CV have flourished: self-citing, suggesting to colleagues to add one’s articles as a reference, dividing a work into several articles citing each other…
“We had come across suspected cases of manipulation and we wanted to study further how certain profiles accumulate citations”explains Yasir Zaki, a professor and researcher in Abu Dhabi. With $300 (270 euros), his team quickly obtained fifty citations benefiting an invented researcher, from a fictitious university, created for the needs of the experiment and having written (thanks to ChatGPT) twenty articles on the subject of fake news. This bonus of fifty citations comes from only five articles, four of which were published in a chemistry journal, probably infiltrated by the company to easily pass articles. The researchers also suspect that this whole system was used to boost the rankings of other clients. “We were shocked to see how indicators could be manipulated”notes Talal Rahwan, also a professor in Abu Dhabi. “We assumed it, now we have confirmation”, notes Cyril Labbé, professor of computer science at the University of Grenoble-Alpes, who, in 2010, also created a fictitious profile and fake articles boosting his number of citations.
Faux articles
The preprint of the researchers at New York University goes beyond this “successful” experience. First, by surveying 574 scientists from the ten largest universities in the world, they found that 60% of respondents evaluate candidates on the criterion of citations. An identical share uses the Google Scholar tool for this purpose rather than commercial competitors, Web of Science or Scopus. However, the latter rely on a corpus of around twenty thousand journals, therefore much narrower than Google Scholar, which also harvests preprint sites, researchers, specialized social networks (ResearchGate, Academia, etc.) or collaborative platforms (Open Science Framework or OSF, Authorea, etc.). This is where the flaws lie.
You have 36.18% of this article left to read. The rest is reserved for subscribers.
Related News :