More than 120 papers have been withdrawn from subscription databases of two high-profile publishers, IEEE and Springer, because they were computer generated thanks to the SCIgen software designed to generate random computer science research papers. The trouble is that they had no meaning at all. All of them were labelled as peer reviewed and all of them were published in proceedings of actual conferences. In theory, these non-sense papers have been reviewed, presented and questioned by a chairman. This leads to quite an odd situation.
The two publishers that were impacted reacted in two very different ways. IEEE, with more than 100 SCIgen paper published, is not directly facing up to the problem and is just deleting the papers without notification. Titles of papers can still be found in the table of contents of conferences but are not available. The Springer response seems much more adequate. It has published a retraction notice alongside each fake paper while announcing that measures to deter future copycats have been taken.
Incidentally, most of these offending papers where related to conferences that took place in China. And the phenomenon may stem from some particularities of the way certain conferences are organised in the country. Nevertheless, all those non-sense publications have found their way through the scientific publications system, even though it should never have happened.
This can be explained by several factors, which are substantially changing the way the scientific community shares its knowledge. On the one hand, technological developments have made the writing, publication and dissemination of documents quicker and easier. On the other hand, the pressure of individual evaluation of researchers—the proverbial publish or perish—is changing the publication process. This combination of factors has led to a rapid increase in scientific document production.
In a sense, one could say that the global knowledge is growing ever faster than before. The presence of this junk science can be interpreted as a side effect of the ‘publish or perish’ paradigm combined with a blind evaluation based on abstract numbers. As a result, such situation is, in itself, is a clear incitation to fraudulence.
Text mining tools, like the SCIgen detection solution, are very efficient in identifying SCIgen papers, duplicated papers and plagiarisms. But the use of such tools is a ‘quick and dirty’ response to the problem. The situation is like if a kind of spamming war started at the heart of science. The phenomenon is taking place precisely at the very heart of science, because knowledge diffusion is at the heart of science too. It is a spamming war, because exerting high pressure on scientists mechanically leads to too prolific and less meaningful publications even if they are not non sense.
In quantum mechanics, the act of measuring a system results in that very system being disturbed. This adage is true in physics, but also in computer science and in scientometrics and bibliometrics. By aiming at measuring science, these approaches are perturbing scientific processes, particularly when used for management purpose.
Lecturer Joseph Fourier University, Computing Laboratory (CNRS, LIG), University Grenoble Alpes, France.
Featured image credit: CC BY-NC-ND 2.0 by Alby Headrick
Go back to the Special Issue: Ethics, values and culture driving research