Scientific sleuths use ‘tortured phrases’ to find research fraud

Impenetrable jargon in scholarly papers can often cause readers to tune out, but might academic gobbledegook hold the key to smashing research fraud? The answer, it seems, is overwhelmingly “yes” after a number of journals were forced to retract papers following the detection of bizarre, unexpected and sometimes hilarious scientific terms, known as “tortured phrases”.

This month, the Institute of Physics announced that it had pulled almost 500 papers, many of which had been flagged for plagiarism and citation manipulation owing to suspicious phrases, while the Problematic Paper Screener, which uncovers such terms, has been credited in hundreds of retractions made in the past two years.

Among the more unusual phrases to catch the attention of keen-eyed scholars are “fluffy rationale” (fuzzy logic), “bosom peril” (breast cancer), “kidney disappointment” (renal failure), “shrewd devices” (smart devices) and “subterranean insect province” (ant colony).

Data scientists who were baffled by references to “uncooked records” and “crude information” later realised the authors were talking about “raw data”, while “secret word recuperation” was used instead of “password recovery”.

Computer experts mulling the repeated mention of a “steering convention” worked out that it was a synonym for a “routing protocol” used by internet servers. Meanwhile, those reading a paper on MRI scanning technology eventually realised that “attractive reverberations” did, in fact, refer to “magnetic resonance”.

Respiratory medics puzzled by the mention of a “valuable hack” later deduced the authors meant a “productive cough”, while Mitt Romney, the Utah senator who lost to Barack Obama in the 2012 US presidential poll, may be disappointed to learn that an allusion to “Glove Romney” and his ineffective use of big data while electioneering was published uncorrected in a 2019 conference paper.

While some tortured phrases may be explained by inadvertent use of autocorrect or poor English, nonsensical text is often a marker of plagiarism and other serious types of academic wrongdoing, said Guillaume Cabanac, associate professor of computer science at the University of Toulouse, whose software skills have been used to scan academic literature globally for tortured phrases.

“The real authors – or scammers – are usually asked to deliver papers quickly. So they will find five or six papers of interest, plagiarise their abstracts and use paraphrasing software to avoid detection by plagiarism checkers,” explained Dr Cabanac.

While Dr Cabanac and his immediate collaborators, Cyril Labbé and Alexander Magazinov, have led the hunt for “tortured phrases”, they have increasingly been joined by scientific sleuths around the world, with this global “invisible college” keen to uncover and share many more problematic phrases and papers on PubPeer, which facilitates post-publication peer review, he explained.

“It’s like a lab without borders – we trust each other and there are many people spending many hours a day reading papers who are keen to help,” said Dr Cabanac.

“We’re also seeing more amateur sleuths getting involved, and that could be a game-changer for detecting research fraud.”

jack.grove@timeshighereducation.com