October 2, 2012
Fraud, Not Errors, Account For Large Percentage Of Retractions In Scientific Papers
Lawrence LeBlond for redOrbit.com - Your Universe Online
A new study finds that fraud in scientific research is growing at a troubling rate, with an average of 300 papers per year being retracted for some form of scientific misconduct.
It had been generally perceived that most peer-reviewed medical, biological and scientific papers retracted have been done so due to unintentional errors. However, authors conducting the largest-ever study of retractions, have found that the percentage of papers withdrawn due to fraud or suspected fraud has jumped significantly since the mid-1970s. In fact, in 1976, there were fewer than 10 fraud retractions per 1 million studies published, compared to 96 per million in 2007.
Study authors are not quite sure why this is occurring, but point out that study authors are under constant pressure to make a big splash in science, both on the funding and attention level. And its possible science is mirroring a subtle increase in deception observed in overall society.
The new study, published in the journal Proceedings of the National Academy of Sciences (PNAS), has found that 66 percent of retracted life-sciences papers were done so due to misconduct. It follows another study published last year by the journal Nature, that reported an alarming increase in the number of retractions of scientific papers--an increase ten times that of the previous decade.
And while other studies have suggested many retractions have resulted from unintentional errors, the new study dug deep to challenge that common comforting assumption. The authors analyzed more than 2,000 retracted papers in the biomedical and life sciences categories and found that misconduct was the reason for 75 percent of the retractions seen.
The authors, including senior author Arturo Casadevall, MD, PhD, professor of microbiology, immunology and medicine at Albert Einstein College of Medicine in New York, noted that misconduct included fraud, suspected fraud, duplicate publication and plagiarism.
“Biomedical research has become a winner-take-all game—one with perverse incentives that entice scientists to cut corners and, in some instances, falsify data or commit other acts of misconduct,” said Casadevall.
To determine the reasons for the retractions, the authors consulted several secondary sources, including the National Institutes of Health (NIH), Office of Research Integrity and Retractionwatch, which investigates scientific misconduct.
The researchers found that about 21 percent of the retractions were from unintentional error, while 67 percent were due to misconduct--of which 43 percent due to fraud, 14 percent due to duplicate publication, and 10 percent plagiarism--and an additional 12 percent of retractions were due to unknown reasons.
“What´s troubling is that the more skillful the fraud, the less likely that it will be discovered, so there likely are more fraudulent papers out there that haven´t yet been detected and retracted,” said Casadevall.
Earlier studies that reported on retraction accounts relied solely on the journals´ retraction notices, which are written by the papers´ authors, noted Casadevall. Many notices are wrong, because authors commonly write ℠We regret we have to retract our paper because the work is not reproducible,´ which is not exactly a lie, Casadevall continued.
“The work indeed was not reproducible – because it was fraudulent. Researchers try to protect their labs and their reputations, and these retractions are written in such a way that you often don´t know what really happened,” Casadevall explained.
The study also found that journals with a higher impact factor (a measure of a publication´s influence in the scientific community) had significantly higher rates of retractions. Casadevall attributes this to prevailing culture in science, which rewards scientists for publishing large numbers of papers and getting them published in prestigious journals.
“Particularly if you get your papers accepted in certain journals, you´re much more likely to get recognition, grants, prizes and better jobs or promotions,” he said. “Scientists are human, and some of them will succumb to this pressure, especially when there´s so much competition for funding. Perhaps our most telling finding is what happened after 2005, which is when the number of retractions began to skyrocket. That´s exactly when NIH funding began to get very tight.”
Casadevall has proposed a number of solutions to the problem of scientific misconduct, and published such solutions in a recent article in the journal Infection and Immunity. Among possible solutions he outlined were the following: more emphasis placed on the quality of publications rather than quantity; less emphasis on impact measures when rating journals; fostering a cooperative and collaborative culture in the research community; developing more stable and sustainable sources of research funding; and creating more flexible career pathways to prevent the ongoing loss of capable scientists due to inadequate funding.
Casadevall noted that the findings were not completely appalling. He and his colleagues reported that 43 percent of all retractions came from just 38 of the thousands of periodicals worldwide. “So while we´re not looking at a systemic disease, so to speak, in the scientific community, our findings do indicate a significant problem that needs to be addressed.”
“Very few people are doing it, but when they do it, they are doing it in areas that are very important,” said Casadevall. “And when these things come out, society loses faith in science.”
Casadevall added that he hadn´t set out to research study fraud. His initial plan was to examine the most common avoidable errors that caused retractions.
Nicholas Steneck, director of the research ethics program at the University of Michigan, said this is the first study to show scientific misconduct as the leading cause of retractions in scientific papers. It shows that there is a need for better, more honest reporting of retractions by science journals themselves, noted Steneck, who was not part of the new study.
Casadevall noted that most “scientists out there are well meaning and honest people who are going to be totally appalled by this.”
In the study, Casadevall worked with Dr. Ferris C. Fang of the University of Washington and R. Grant Steen, a medical communications consultant in Chapel Hill, North Carolina. The trio gathered all the retraction notices published before May 2012 by searching PubMed, a database of scientific literature maintained by the National Library of Medicine.
“I guess our O.C.D. kicked in and we started trying to look at every paper we could look at,” Fang said.
After reviewing all the data, the trio compiled a listing of the journals with the most retracted articles due to misconduct.
At the top of the heap was The Journal of Biological Chemistry with 37 retractions for fraud, followed closely by Anesthesia & Analgesia with 33 retractions and Science with 32 retractions; The Journal of Immunology had 30 retractions due to misconduct. Surprisingly, the Proceedings of the National Academy of Sciences, the very periodical that published the latest fraud study, ranked 5th with 27 retractions.
Other journals with retracted papers due to misconduct include Blood, Nature, the Journal of Clinical Investigation, Cancer Research and Cell rounding out the top ten. Casadevall and his colleagues found that 38 research groups with five or more retractions accounted for 44 percent of articles linked to fraud or suspected fraud.
“We haven´t seen this level of analysis before,” said Dr. Ivan Oransky, an author of Retraction Watch and the executive editor at Reuters Health. “It confirms what we suspected.”
Oransky told Carl Zimmer of the New York Times that he expected the rise to continue in the near future. He and his co-author, Adam Marcus, have been scrambling to keep up with new cases of fraud.
One example comes from the Japanese Society of Anesthesiologists, who reported in July that Dr. Yoshitaka Fujii had falsified data in 172 papers, most of which have not been officially retracted.
Dr. Benjamin G. Druss, a professor of health policy of Emory University, said he found the statistics in the paper to be sound but added that they “need to be kept in perspective.” Only about one in 10,000 papers in PubMed have been officially retracted, he noted. By contrast, 112,908 papers have had published corrections.
Casadevall disagreed. “It convinces me more that we have a problem in science,” he said.
While the occurrence of fraudulent papers remains relatively few, the rapid increase in retractions due to fraud is a sign of a winner-take-all culture in which getting a paper published in a major journal can catapult a career. He added that this incentive could be driving some authors to break the rules.
“We need to look at how we have structured the system, so scientists are not given incentives to [commit fraud] quite as strongly,” noted Fang.
Zoe Corbyn of Nature reported that the study also found some significant geographical differences. Retracted papers with lead authors based in historical scientific superpowers, such as the United States and Germany, were more likely to be linked to fraud. In emerging scientific powers such as India and China, however, plagiarism and duplication caused more of the retractions.
“These trends may reflect differences in incentives, cultural norms and proficiency in English among these countries,” remarked Fang.
Casadevall said the latest work is about science trying to clean its own house. And because it's about fraud, he said he did one extra thing with his study: He sent reviewers not just a summary of their work, but all the data, “so they can check on us.”