The word “bubble”, Paolo Crosetto did not want to write it in his study. But it is this vocabulary borrowed from financial crises that this economist chooses when we ask him what all these curves that he sent us by email at the beginning of September mean. With three of his colleagues, he has just signed an analysis on the health of the scientific publishing sector, published in Quantitative science studies. This appears to be positive: obtained by analyzing the online registers of the main players in the sector, its data shows a surge in the number of scientific articles published between 2016 and 2022, of the order of +47 %. But according to the presentation of the researcher, affiliated with INRAE (national research institute for agriculture, food and the environment), this apparent dynamism actually masks a serious risk of crisis in the sector.
Because by examining the data more closely, a large part of this growth seems to be based on the proliferation of products that are particularly risky on a scientific level, threatening, in addition to the publishing economy, the quality of the accumulated knowledge. Paolo Crosetto spins the metaphor: “It’s as if we were in the spring of 2007, just before the crash. Except that instead of the financial system and purchasing power, it’s confidence in research that could be affected “.
The rush for “special issues”
Over the period studied, many scientific journals rushed to produce issues of a new type: “special issues”, special editions organized around a theme rather than a discipline. Normally exceptional, these publications now account for 38% of the scientific production of the main publishing houses, guarantors of the probity and rigor of scientific literature. This is three times more than in 2016.
Paolo Crosetto compares these special issues to “subprimes”, these apparently safe batches of credit but including risky loans, which flooded the financial markets in 2008: “These scientific productions are backed by well-rated journals and seem of good quality, but in reality they are particularly at risk of conveying poor quality content, plagiarism or fraud”, summarizes the researcher.
In a traditional edition, independent researchers commissioned by the journal reread the work. The special issues are constituted by a “guest editor”, an external scientist, responsible for recruiting his own review committee and attracting contributions. A win-win subcontracting, in theory: “Flattered, the guests mobilize their network to fill their pages, and the magazines develop,” continues Paolo Crosetto. The system makes it possible to attract more publications, and therefore more money for journals which make money by charging researchers for publication costs.
First bankruptcies
An attractive system, but more volatile. Difficult to criticize a study that comes from knowledge. “By delegating, publishers are more exposed to predatory behavior from scientists or structures who seek to publish en masse to increase their statistics, without worrying about the content of their articles,” notes Hervé Maisonneuvemember of the European Association of Science Editors. Within this association which brings together the editors of the main European newspapers, the specialist also saw “the bubble” coming. He now fears that part of the sector will collapse.
A disaster scenario, but not improbable. The first bankruptcies have already taken place: in 2022, the Hindawi publishing house, whose astonishing growth essentially depended on these products, saw its production diverted by “paper mills”, these structures which place bogus paper for scientists desperately seeking notoriety, and who have succeeded in polluting at least 2% of the scientific literature. When the markets learned of the deal, the stock’s owner, Wiley, who had bought Hindawi for $298 million in 2021, lost $400 million in one day. He had to close Hindawi to avoid collapsing.
The sinking brings to mind these banks which, upon realizing the toxicity of subprimes, sank in 2008. If scientific studies are not financial assets, the risk of contagion is very real. Hindawi’s fall caused the largest “retraction” in history: 8,000 papers were, at once, withdrawn from literature, the harshest sentence for scientific work. Frontiers And MDPItwo publishers who had invested massively in the production of special editions – much more than Hindawi – suddenly slowed down their production following this incident.
A sudden slowdown
These houses, relatively young in the sector, showed staggering growth until 2022: the number of works published by these three structures jumped by 675 and 1079%, between 2016 and 2022, according to the work of Paolo Crosetto. This exponential increase – 6 to 10 times that of Hindawi – allowed them to move from the Top 100 to the Top 6 in the world of the most productive scientific publishing houses. For comparison, the most prestigious of their competitors, Springer, Nature (a branch of the Springer group) or BMCsaw their production increase by only 51, 32, and 3%.
After the dismantling of Hindawi, the number of articles published by Frontiers And MDPI fell by 14% and 21% respectively in the second half of 2023. In these houses, special issues had proliferated so much that they far exceeded normal production. We must count them, to realize the real scale of the phenomenon: in 2022, MDPIthe undisputed champion of the practice, has published around 188,000 articles, in “exceptional” issues, compared to only 25,400 in normal editions. Exposing itself, in fact, to an immense risk in the event of fraudulent practices.
The ship has already shook, twice. In 2023, two newspapers MDPIJournal of Risk and Financial Management and International Journal of Environmental Research and Public Health (IJERPH), have been “delisted” from scientific encyclopedias, these directories which aggregate all productions. Special editions on themes very far from their initial expertise had infiltrated the middle of volumes containing science that was nevertheless suitable. Thus made invisible, these two titles, although well integrated into research, often cited in works, had to suddenly scale back, going from several thousand published articles to a few hundred.
The discontent of researchers
Alerted by these shocks, more and more scientists are making their discontent known. “I do not know a researcher who has not been invited in special issues on subjects on which he had no expertise,” laments Megha Sud, doctor of geography and member of the International Science Council, the union International Union of Scientists. Many researchers report the same stories: off-ground, purely commercial canvassing.
Scientists have even decided to pack up because of the risks to the financial health of publishing houses and, above all, to the quality of scientific literature. This was the case for Gemma E. Derrick, part of the editor-in-chief of Publications MDPI because of this “predatory” behavior. And around forty reviewers of the journal NeuroImage of Elsevier, who left the ship after an increase in prices deemed inadequate with the proposed review.
In 2022, theUnescothe body of the United Nations among others responsible for science, saw itself obliged to publish a guide to discern risky initiatives from others, as the system is so permeable. In Switzerland, the National Research Fundthe main public funder of science, decided last year to reserve its publication support for normal editions. “Everyone was amazed at the speed with which the special editions spread,” comments its president, Matthias Egger.
Towards a reform of the model?
If the national research agencies of major scientific nations, such as the United States, the United Kingdom or France, have not followed this movement, considering that freedom of publication takes precedence, they have banned the use of “bibliometric indices”. , indicators that compile the number of articles published or the number of times they are cited. “This helps prevent scientists from being overly encouraged to publish and turning to suspicious numbers,” relates Patrick Couvreur, president of the Evaluation and Open Science committee of the Academy of Sciences.
The measure did not resolve the problem: “Bibliometric indices are, in fact, still over-watched. Because the juries which examine promotions, recruitments, or scientific projects cannot be experts on all subjects”, continues the specialist. Paolo Crosetto’s curves did not surprise him: once, a magazine included him in a review committee without even informing him. “Bad practices are too widespread, with many journals in the gray zone,” regrets the specialist.
In May 2024, Paolo Crosetto was invited by the Global Research Council, the forum of scientific institutions, to present his study. Since then, his inbox has been filled with collaboration proposals from national agencies wanting to benefit from his ideas to combat bad editorial practices. The economist would like publishers, largely private companies but whose profits depend on public funds, to be obliged to publish their statistics. The number of special editions published, but also the time spent rereading articles, has declined sharply in recent years.
For his part, Frontiers refutes the idea that special editions are risky products. The publisher prefers to see them as “innovation” rather than as commercial operations. “Protection against fraudulent research is strictly a matter of editorial quality control, linked to the policies, processes and practices of each publisher, and does not depend on special issues,” assures the group to L’Express. Frontiers is publicly attacked the seriousness of Paolo Crosetto’s studyconsidering that it was truncated. Which did not prevent the review Quantitative Science Studies to publish it. Many independent scientists confirmed the seriousness of his analysis to L’Express. MDPI did not wish to answer our questions.
.