Scientific fraud, poor research and honest mistakes lead to thousands of retractions

This is one of a five-part series from Cosmos Weekly, investigating the peer review process.

On 10 October 2022, there were 36,104 retractions listed on a website specifically built to collate journal retractions and editors’ “expressions of concern”.

Some of them are the result of honest mistakes (many of which are subsequently corrected by the authors and reflected in the journal database). Some are deliberately fraudulent.

If you were to look at the most recently retracted article listed on the Retraction Watch database or check out the latest queried article on Pubpeer – a website through which members of the scientific community voice their concerns about specific journal articles – you would probably click on a paper you’ve never heard of.

This is to be expected, given the enormous number of papers published in a diverse range of specialised fields. Every so often, however, the maelstrom of outrage becomes impossible for the media to ignore. In most of these cases, it’s not just the specialist field that suffers.

Cases such as Sylvain Lesné’s alleged manipulation of images in a key Alzheimer’s research paper from 2006, which then diverted research funding and drug trials, have potentially  indirectly impacted the lives of many patients living with dementia.

The integrity of science and the process of peer review is also, quite reasonably, brought into question – and this can have some pretty dire consequences.

What if the science is faulty, but the outcome is beneficial?

In 2016, a marine biology paper published by Dr Oona Lönnstedt and Professor Peter Eklöv in Science claimed to find definitive evidence that microplastics (such as exfoliating microbeads popular in cosmetics and toiletries at the time) were devastating to marine life even at relatively low concentrations. It had been previously demonstrated that fish ate microplastics and that at high concentrations (exceeding levels found in natural environments) they were harmful to the fish.

In the paper, Lönnstedt, a PhD graduate of James Cook University (JCU) then employed at Uppsala University, Sweden, claimed that not only did the microplastics affect the behaviour of larvae fish, making them slower and more “stupid”, but that the fish preferred to consume the microplastics over their normal diet.

Lönnstedt’s work was picked up by media worldwide and added to a large body of evidence regarding the harms caused by microplastics in the marine environment. Legislation was quickly introduced in many places around the globe (including Australia, Canada, Italy, Korea, New Zealand, Sweden, the UK and the US) to control the use of these microplastics. But alarm bells began to ring.

Colleagues of Lönnstedt, present at the time she conducted her experiments at a research station on an island in the Baltic Sea, reported discrepancies in the data and timeline of the experiments. Somehow most of the data wasn’t backed up and was lost when a laptop was stolen, and there didn’t seem to have been enough time for Lönnstedt to have done the experiments during her stay in the remote lab. The paper was subsequently retracted from the journal after a recommendation by the Central Ethical Review Board in Sweden.

Overwhelmingly, scientists agree that a ban on the use of microplastics was desperately needed, and that controls introduced don’t hinge on Lönnstedt’s research alone. But it remains an important and undeniable lesson: the failure of peer review to detect fraudulent science has implications reaching far beyond the population of fish being studied, spanning oceans and even penetrating our courts of law.

The retraction of the microplastics paper resulted in an increased interest in Lönnstedt’s earlier works, including research performed as a PhD candidate at JCU. Questions were raised about the data collected within one of her papers resulting from her PhD research, particularly a paper published by Biology Letters on the cooperative behaviour of hunting lionfish. Corrections have been made to the original paper to address an editorial expression of concern.

According to a report of proceedings, an independent external research misconduct inquiry into Lönnstedt’s research while at JCU was conducted by the university. The JCU panel reported that despite there being a “number of breaches of the Research Code” and “inadequate reporting of data and lodgement and securement of data post-PhD”, these constituted poor research practice or professional standards rather than research misconduct.

For the whistleblower who initially cast doubt on the lionfish data, the outcome was anything but satisfactory. This individual, who asked not to be identified because of concerns doing so might hurt their career, told Science, for whistleblowers, “it’s exhausting. And depressing.”

Overwhelmingly, scientists agree that a ban on the use of microplastics was desperately needed – but it remains an important and undeniable lesson: the failure of peer review to detect fraudulent science has implications…

Sadly, this is not the only case of fishy science, and neither is it the only case with links to an Australian university.

An extremely high-profile recent case that has also been covered in depth by Science, refers to allegations of fraud and misconduct of ‘star’ marine ecologist, Associate Professor Danielle Dixson by the University of Delaware (UD). Although the findings pertain to research performed by Dixson while employed at UD, she, like Lönnstedt, also completed her PhD studies at JCU.

During her PhD, Dixson investigated the effects of ocean acidification on the behaviour of coral reef fish. The research claimed that acidification resulted in fish becoming confused and swimming towards chemical cues emitted by predators, rather than away from them.

In 2020, Australian scientist Timothy Clark (now an associate professor at Deakin University in Victoria) and a group of researchers, published an article in Nature that sought to replicate Dixson’s studies, but could not. Within the paper, the authors write “we additionally show that the large effect sizes and small within-group variances that have been reported in several previous studies are highly improbable”, suggesting that data within the Dixson paper was unreliable.

The authors requested an investigation into scientific conduct by Dixson, leading to UD seeking a retraction of three of Dixson’s papers on fish behaviours and coral reefs, following an earlier retraction of a separate article published in Science. Like Lönnstedt, this has launched further interest into other Dixson papers.

Data that bites

In a blog post from 2020, Kate Laskowski, now an assistant professor at the University of California in the US, details how, as a final year PhD student, she had been excited to confirm a theory on spider behaviour. The theory, which suggested that niches develop in certain spider colonies following repeated social interactions, used data provided by behavioural ecologist Jonathan Pruitt.

According to Laskowski’s blog, “the paper was honestly a breeze for me to write up as the predictions were so clear and the results so straightforward”. She published the work with Pruitt as second author in 2014 in the Proceedings of the Royal Society B.

A few years later, acting on the concerns of a colleague, Laskowski went through the data provided by Pruitt and discovered blocks of potential duplication. Ultimately, she moved to retract co-authored affected papers from academic journals, feeling she could no longer trust the data or conclusions.

At least 13 retractions of Pruitt’s work have followed, and six papers have been labelled with “expressions of concern” by journals. Pruitt has since resigned from his academic position, although the situation remains unsatisfactorily resolved with Pruitt frequently threatening and pursuing legal action.

For Pruitt’s past collaborators and former lab members, there’s an ongoing stigma of being associated with him, not to mention a reluctance to trust other collaborators – certainly a major drawback in a global knowledge environment where collaboration is key.

Salting the earth with fraudulent fossils

The examples discussed above might lead you to the reasonable, but definitely erroneous conclusion that scientific misconduct is a relatively new problem. History is rich with stories of scientists seeking to undermine the integrity of the peer review system, with one from the world of palaeontology being a particularly concerning – and long-running – example.

In 1989, Professor John Talent, a palaeontologist at Macquarie University in Sydney, published an article with Nature. The case of the peripatetic fossils publicly cast doubt on the research of a famous Indian palaeontologist, Viswa Jit Gupta of Punjab University, by suggesting that Gupta had created “a quagmire of palaeontological disinformation”.

At the time, Gupta had published over 300 articles in scientific journals around the globe, working with a large number of international co-authors, so Talent’s allegations caused quite a stir. In the paper, Talent alleged that Gupta would provide potential co-authors with fossils that came from shops and research laboratories, rather than the areas in the Himalayas that Gupta claimed he’d found them. Gupta also doubled up on images used for samples in papers, claiming they were different specimens. Talent suspected they were the identical artefact.

Glenn Brock, honorary professor of palaeobiology at Macquarie University and a PhD student of Talent’s, explains his supervisor “wanted to go to a locality to recollect some of the fossils Gupta had published on, so as to better date them”.

“But when Talent got there,” says Brock, “he discovered that instead of Devonian fossils, he found only Precambrian metamorphic layers”.

Talent assumed he’d made a mistake at the time, but as he became more familiar with Southeast Asian geology, he began to wonder. When Talent came across a paper co-authored by Gupta and a German palaeontologist by the name of Heinrich Erben, “instead of visiting the locality, Talent contacted Gupta’s co-author and asked about the material. He said: ‘Are you sure that it’s from the Himalayas?’ And Erben replied: ‘Well, it’s been said to me, and I take it that this is the truth, because scientists don’t lie’,” says Brock.

Brock has a very personal connection to what is now known as the Gupta saga. Not only was Talent Brock’s PhD supervisor, but Brock worked as a research assistant in Talent’s lab. Brock himself had a hand in joining the Gupta dots: “John asked me to dig deeper into Gupta’s papers. There was one on trilobites and corals, and another one on brachiopods. It was simple. I read the papers, saw the images Gupta had published as samples he’d collected [from the Himalayas], and then I’d check the reference list.

“I would cross-check each reference. There were some from 1908 and 1911 from some British palaeontologist who’d been there at the turn of the 20^th century. So, I’d go to these volumes – often in big Indian memoirs that were hard to access – and I’d pull them off the shelf and open up the plates. And I’d say, ‘That is exactly the same as the one that Gupta has published’.”

“These papers were quick and easy. He didn’t have to go into the field, he’d just use someone else’s work and essentially pass it off as his own.”

Professor Glenn Brock

Considering the tools available at the time, the methods Gupta used were quite crude, explains Brock: “I remember seeing a famous English geologist who’d published plates of some brachiopods he’d found some 60 years before Gupta. And all that Gupta had done was take some scissors and cut out the specimens, put them down on a new plate with a new number on them and claim them as his own – and these were samples from somewhere very different, from parts of Somalia. So, he actually incriminated himself, by putting the source of his fraudulent material in his reference lists, probably thinking that no one would ever look.”

It was a monumental task for Talent and Brock, but some 2,530 papers later, an obvious trend emerged. “It was very clear,” says Brock, “that this was part of his modus operandi, especially during his PhD. These papers were quick and easy. He didn’t have to go into the field, he’d just use someone else’s work and essentially pass it off as his own.”

At first, Talent did not want to be involved in the saga. There is certainly a stigma that exists, even today, amongst scientific whistleblowers. The threat of being sued – or even worse, of being accused of academic jealousy – was real. But, Brock says, Talent “was so invested in the Indian geological story because he was building a paleobiogeography of all of Southeast Asia, including Russia and China, and these anomalies just didn’t fit. Without continental blocks being taken out of one spot and placed in another, it was just completely improbable.”

Gupta managed to survive for the most part, despite legal threats from other authors and the scientific community. After an initial cover-up attempt via a media ban by the Indian government, he was made to retire early, but with full benefits.

Gupta also attempted to sue Talent, who in an interview with the ABC said he’d received death threats from Gupta, and that people with proof of Gupta’s deceit and their family members were targeted by hired hitmen, with one lab technician dying from a hit-and-run and a family member of one of the report’s co-authors severely injured in another road accident.

The Gupta saga is chronicled and narrated by distinguished Australian Science communicator Robyn Williams in a 1991 film called The Professor’s New Clothes, archived at the ACMI.

---

This story is part of a five-part series from Cosmos Weekly on peer review. Read the other four:

• Is it time to review, the review?
• When peer review fails, people get hurt
• Putting science under the microscope reveals not all is well
• Peer review: where does it lie in the scientific process?

Next week: Clare Kenyon interviews Elisabeth Bik and Ivan Oransky on Monday. Subscribe to our free daily newsletter to be the first to see it.