Jan. 22, 2013 at 2:07 PM ET
By Stephanie Pappas
Men are more likely than women to commit scientific fraud, a new analysis of misconduct convictions reveals. And the urge to cheat spans the entire range of academic careers, from students to seasoned professors.
For the new study, published Tuesday in the journal mBio, scientists examined 228 cases of misconduct in the records of the United States Office of Research Integrity (ORI), a government agency that oversees research funded by federal, public health-related agencies. Part of the ORI's mission is to monitor investigations of charges such as fabrication of data and plagiarism.
"The big picture is not that most scientists are dishonest, it's the opposite," said study researcher Ferric Fang, a microbiologist at the University of Washington School of Medicine. "But on the other hand, a few scientists being dishonest is a very bad thing, because it casts doubt on the whole enterprise."
Fraud in science
As of May 2012, at least 2,047 biomedical and life science studies had been retracted by the journals that published them, meaning that the studies contained errors or fabrications that rendered their results meaningless.
Fang, along with Arturo Casadevall, a professor of medicine at the Albert Einstein College of Medicine of Yeshiva University in New York, and colleagues analyzed these studies and found, to their surprise, that 67.4 percent were retracted because of fraud, duplicate publication (essentially, researchers "double-dipping" to get a paper published twice) or plagiarism. [ Understanding the 10 Most Destructive Human Behaviors ]
There are no firm numbers about how much misconduct goes on in science, but Fang, Casadevall and their colleagues turned to the most complete database on the subject, which is run by the ORI. It's the best database in the world, Casadevall said, because the cases have been thoroughly investigated and documented.
Between 1994 and the present, the ORI investigated 228 cases of alleged misconduct. Of these, 215 were found to involve wrongdoing. In 40 percent of these cases, the guilty party was a trainee (a student or postdoctoral researcher). In 32 percent of cases, it was a faculty member, and in 28 percent of cases, the fraud was committed by technicians, study coordinators or other lab staff.
"We originally thought that misconduct was going to be a problem primarily of trainees or people starting out," Casadevall told LiveScience. "We were surprised to find that, in fact, a lot of them were quite established."
Another key finding was the gender schism in fraud. Even given that men outnumber women in the upper echelons of science, males committed more of the fraud than would be expected. The gap appeared on every rung of the career ladder given the relative proportion of men and women at each step.
Among research staff, 43 percent of those committing misconduct were male. Among students, men made up 58 percent of transgressors. That number rose to 69 percent among postdoctoral researchers and to 88 percent of faculty. [ Oops! 5 Retracted Science Results of 2012 ]
Among the 72 faculty members who committed fraud, only nine were female, the researchers found. That's one-third of what would be expected if the genders were committing fraud at the same rates.
It's not clear why the gender gap exists, Casadevall said. Men are generally known to take more risks than women, which could play a role. Additionally, the researchers can't rule out the possibility that women commit misconduct as frequently as men, but don't get caught.
The researchers did find, however, that the proportion of men and women investigated for fraud was similar to the proportion found guilty, Fang said. So the investigation process itself does not appear gender-biased.
Stiff competition for research funding, jobs and scientific awards is likely behind the urge to cheat, Fang said. In the 1960s, 60 percent of researchers who applied for a standard federal research grant won that grant. Today, the chance of success is only 18 percent.
"It's become extraordinarily competitive," Fang told LiveScience.
That doesn't mean that cheating scientists are off the hook ethically, he said, but the environment of science likely contributes to the problem. Among faculty, almost all misconduct recorded by the ORI involved grants or papers, while among trainees and lab staff, the motivations appear to involve working in the "pressure cooker" of a lab where results are expected. The pattern suggests that principal investigators in charge of labs need to take heed of the climate they're creating, Fang said.
"Even without being a crook, you can be a principal investigator who, under pressure, may be creating pressures on your people to generate certain results," he said.