I went to a panel discussion at the New York Academy of Sciences on the evening of April 30th that addressed the topic of various forms of scientific malfeasance, ranging from plagiarism to outright manipulation of data. A gripping and deeply unsettling topic, as it relates directly to the research studies that I pore over every day at my desk. After the session moderated by Scientific American Editor in Chief Mariette DiChristina ended, I thought that would be my fill of data finagling for at least maybe a week.
The next morning, though, there was a breakfast gathering of editors and other representatives from foreign editions of Scientific American. I asked a bunch of the editors there about coverage that might be of interest to us here stateside. Koen De Buck, commercial director of EOS, the group of Antwerp-based science magazines that publish the Dutch edition of Scientific American, mentioned a study by online news coordinator Reinout Verbeke on surprise, surprise—fraud of all things.
Verbeke, in collaboration with Dutch psychiatrist Joeri Tijdink, decided to explore the prevalence of fraudulent research practices in the Flanders region of Belgium, as shocks from the 2011 scandal involving Dutch social psychologist Diederik Stapel continued to reverberate. (See the New York Times Magazine feature on Stapel that appeared on April 28.)
Verbeke and Tijdink cast a wide net, with support they received from the Pascal Decroos Fund for Investigative Journalism. They contacted researchers from the medical science faculties of every university in Flanders, sending out more than 2,500 questionnaires and receiving 315 fully completed anonymous responses in return.
The answers startled. Four of the researchers who responded, or 1.3 percent, acknowledged that they had fabricated data at least once during the past three years, misdeeds that may still be unpunished. What's more, 23, or 7.3 percent, of those who sent back questionnaires had engaged in the quaint term “massaging”—in which data or results were removed to make their work true up with original hypotheses. The roughly 8 percent of fraudulent practices found at the universities in Flanders compared with an average of 2 percent of smelly stuff going on that turned up in a 2009 meta-analysis in PLoS ONE of studies from around the world.
EOS published the survey (English PDF) in its April issue and Tijdink and Verbeke are now preparing to submit an article to a scientific journal. “When you get four people fabricating data that's a lot,” Verbeke said in an interview, adding: "People were very willing to cooperate with this survey and admit that there is a problem. They had been unable to speak up and suddenly there was a chance to communicate about science fraud. That can explain the higher figures in my survey [compared to the PLoS ONE meta-analysis]... I'm happy now that there is a debate in Flanders and universities are taking action. They want to install courses in research integrity. They want researchers here to publish their raw data."
Respondents said the publish or die imperative was one of the main reasons for the infractions. The survey found that two thirds of the professors polled ran into excessive pressure to get their work into journals and nearly 70 percent of all of those surveyed had added the name of one author who had not participated in a study.
Ivan Oransky, a founder of Retraction Watch, and a former editor at Scientific American, as well as one of the panelists at the NYAS event, was quoted in the article as saying that these figures don’t surprise when it comes to medical research. “Cooperating with the pharmaceutical industry gains researchers financial rewards. That could pressurize scientists to cut corners," he remarked.
It's studies like this one that have prompted a new look at the way that science is carried out. The Nature journals have just published editorials calling for more transparent research. (Scientific American is part of the Nature Publishing Group.) And the Third World Conference on Research Integrity is scheduled to be held in Montreal from May 5 to 8.
EOS ended its article by calling for a rethink of the scientific enterprise. “Has publication pressure and the demand of specialist journals for positive and spectacular results become so high that it jeopardizes the scientific value of many studies? This survey can be a starting point to open the debate, and where possible, improve or thoroughly revise the scientific system.”
Image Sources: from an English-language version of "Science Fraud: The Hard Figures" published in EOS's April issue.