by David Gutierrez
One in seven scientists report that they have known colleagues to falsify or slant the findings of their research, according to a study conducted by researchers from the University of Edinburgh, Scotland, and published in the journal PLoS One.
A number of scientific data falsification scandals have emerged in recent years, such as the case of a South Korean researcher who invented data on stem cell research. At the same time, increasing controversy over close industry ties to medical research has called into question whether researchers who take money from drug companies might be induced to falsify their data.
“Increasing evidence suggests that known frauds are just the tip of the iceberg and that many cases are never discovered,” said researcher Daniele Fanelli.
The researchers reviewed the results of 21 different scientific misconduct surveys that had been performed between 1985 and 2005. All respondents were asked whether they or anyone they knew of had taken part in either fabrication (outright invention of data) or “questionable practices.”
Questionable practices were any improper procedure short of fabrication, including failing to publish results contradicting one’s prior research, modifying data based on a “gut feeling,” changing conclusions after pressure from a funder or selectively choosing which data to include in an analysis.
One in seven scientists said that they were aware of colleagues who had engaged in fabrication, while nearly half — 46 percent — admitted to knowing of colleagues who had used questionable practices. Only two percent, however, admitted to fabricating results themselves.
While two percent is higher than previous estimates of the prevalence of data fabrication, researchers believe that the number is still too low. In all likelihood, it reflects both a reluctance by researchers to admit to serious misconduct and a tendency to interpret one’s behavior as favorably as possible — questionable instead of fabrication, or acceptable rather than questionable.
Researchers in the medical and pharmacalogical fields were the most likely to admit to misconduct than researchers in other fields.