Lies, deceit and the work of magazines

Abstract

In 2011 a professor from the Department of Gynecology and Obstetrics at the University of Melbourne (Australia) came across a removal notice from a scientific article. This kind of news always catches the attention of researchers and he started reading. The notice clarified that the data presented in the withdrawn article were identical to those presented in a previous publication, but from Spain. The difference was that the original article studied polyps and the retracted article, myomas.

Same data, different pathologies. The decision of the journal editors was the right one, since, although nothing is impossible, it is unlikely that this degree of similarity was the result of chance. Just to put them in context, the probability of two twenty-word sentences being written identically in two different journals (by different authors) is in the range of 1 x 10-80. In the realm of statistics, this is equivalent to "almost impossible". If we add numerical data to this, you can imagine that the number of 0's will be much higher and the chance much lower.

The professor, who was at that time editor of a prestigious journal in the field, took it upon himself to be more cautious with the manuscripts received. Shortly thereafter, he received a manuscript with, in his opinion, manipulated data and rejected it, only to discover that the same article, with the data made up, had been published in another journal a year later.

What does all this lead us to? To remember that the articles we read have passed through the hands of many people, who are still human. If on one end we have someone with a desire to lie, there is a chance that the lies will go unnoticed until they are printed on paper and even then it may be a long time before someone discovers the deception.

How can we avoid this? First, on the investigator's side. Remember why you are investigating and that lies can have consequences in the lives of many patients. If, despite this, your prestige, academic position or money are more important, then you are in the wrong profession. Let your conscience be your jury and executioner.

On the receiving end, detection will depend on factors such as the level of plagiarism (the more extensive, the more likely); the experience of the reviewer and, if possible, the use of artificial intelligence tools. Although they tend to be stigmatized, they can be very useful in this process. With their ability to review millions of pages of data, published articles and websites in seconds, the possibility of lying and successfully escaping is reduced.
It is currently estimated that 60 to 90% of plagiarism will be detected before publication.

Almost 100, but it is not an absolute victory. It only remains for us from this platform to advocate their scientific side, support ethics in research and remind them of the consequences of fraud. After all, today's lies may be tomorrow's rules.