Despite assaults on the science of global warming and evolution, the real threat to research these days may be coming from within. Last year set a record for retractions from the scientific literature, with some 400 in recognized scientific journals. This year looks on pace to meet or possibly even exceed that mark. In fact, an anesthesiologist in Japan may set a new individual record, with 193 papers under suspicion for bad data.
Retractions happen for many reasons, from honest error to plagiarism and faked data. The trend is clear: Journals today are retracting more than 10 times as many articles each year than they did a decade ago, while publishing only about 50 percent more studies than before.
This pattern has prompted a certain amount of soul-searching. Two prominent journal editors in chief — Ferric Fang, of Infection and Immunity, and Arturo Casadevall, of the microbiology journal mBio — told a National Academy of Sciences committee last month that science has become dysfunctional. They argued that scientists may feel forced to cut corners, or worse, as funding levels decline and the pressure to win grants increases. It’s tempting to make this link, but it’s impossible to prove.
The outsized increase in retractions is partly due to greater transparency, rather than more fraud; the introduction in recent years of software that efficiently detects plagiarism is responsible for many of the retractions we’re seeing, and retractions remain an infinitesimal fraction of the 1.4 million papers published every year.
Even so, the institutions of science — universities, laboratories, and journals — shouldn’t pat themselves on the back and take the public’s trust for granted. And there are distressing signs that the kind of transparency scientists need is far from universal.
Consider three recent cases at Harvard. The best-known was that of psychologist Marc Hauser, a leading behavioral psychologist who eventually resigned after a three-year investigation into the validity of his research. The results of that probe have never been fully reported. Then there was a case of alleged misconduct involving a high-profile stem cell lab that led to two retractions. Eighteen months after the article was pulled, we still don’t have a report on that, either. And the most recent case resulted in two retractions for a Harvard group studying cannabinoids in July, but the university has said zilch about its investigation.
That’s not to pick on Harvard. Earlier this year, the University of Connecticut issued a massive report saying a faculty member, Dipak Das, who studies chemicals in red wine, among other things, had fabricated much of his data. While school officials deserve praise for releasing the exhaustive document, it took them three years to make the investigation public — during which Das’s tainted studies continued to contaminate the field.
A lot of scientific research is publicly funded. In a small number of cases, when researchers are found guilty of misconduct, the public ends up hearing about it because the federal Office of Research Integrity gets involved and announces its findings. Calls for a similar agency in Britain have grown louder after a survey of doctors by the high-profile journal BMJ found that about 1 in 8 said they had firsthand knowledge of “inappropriately adjusting, excluding, altering, or fabricating data.” But all too often, investigations of such misconduct are left to institutions with an interest in keeping their reputations as pristine as possible.
That’s why it’s reassuring when some journal editors push institutions farther than they would likely go on their own. Steven Shafer, editor in chief of Anesthesia & Analgesia, has been the driver behind several recent high-profile misconduct cases, starting in 2009 with Scott Reuben, the Baystate Medical Center researcher in Springfield who was forced to retract more than 20 papers and ended up spending time in prison for fraud. Shafer also forced concerns about research by German anesthesiologist Joachim Boldt into the limelight, and gathered a coalition of more than a dozen journal editors who eventually retracted some 90 papers. Not surprisingly, Shafer is also behind the 193 potential retractions by Japanese anesthesiologist Yoshitaki Fujii.
Science needs more Shafers, at least as long as institutions insist on hiding facts from the public that’s paying for its research. Yes, we can trust science, but only when it gives us reason to.