Story URL:
Story Retrieval Date: 11/23/2014 11:14:18 PM CST

Top Stories

Scandalous science: Scientists cheating on data

by Daniel Peake
Feb 17, 2010

Science retractions

Daniel Peake/Medill

 Peer-reviewed science journal retractions have increased tenfold over the past two decades. The total number of retractions has significantly outpaced the total number of science journal articles, particularly in the past five years. Still, retraction rates remain relatively small (less than one percent), and the bigger issue may lie in the unknown number of fraudulent papers that should be retracted but never are.


Daniel Peake/Medill

 The frequency of scientific misbehavior may be underestimated. A recent analysis, which compiled data from various surveys that asked scientists to report scientific misconduct about themselves and their colleagues, offers an idea of how often misconduct occurs.

Despite increasing science journal retraction rates in recent years, scientific misconduct in research and publishing persists – particularly in pharmacological and medical research.

“The fact that scientists sometimes might cheat and can fail to be objective is something that is very well known,” said Daniele Fanelli, a scientific fraud researcher at the University of Edinburgh. “Research is full of conflicts of interest.”

The British journal Lancet's recent retraction of a discredited study linking vaccination to autism exemplifies Fanelli's point. (Read about the reasons behind research fraud and journal retractions in the main story found in Related Links.)

Peer-reviewed science journal retractions have increased tenfold over the past two decades, according to an exclusive Thomson Reuters data analysis for the London-based Times Higher Education in August 2009. This suggests that “people are more willing to report problems in their research,” Fanelli said, calling the trend a good sign.

“The high profile journals that have been burned with recent scandals have been putting in place new policies [to prevent fraud],” Fanelli said. “Something is definitely happening.”

Dr. Jerome Kassirer, editor in chief of the New England Journal of Medicine from 1991 to 1999, said that science journals have changed a lot in the past 10 to 15 years: “Today medical journals are much more careful about identifying financial conflicts of interest. There’s a lot of disclosure now. But disclosure doesn’t tell you everything, so it is only a partial solution to a conflict of interest.”

Still, retraction rates remain relatively small (fractions of a percent), and the bigger issue may lie in the unknown number of fraudulent papers that should be retracted but never are. This number, according to Fanelli’s recent analysis of surveys asking scientists to anonymously report their research transgressions, could be staggering.

Nearly 2 percent of scientists admitted to at least one instance of serious misconduct, which comprises fabrication (making up data or studies), falsification (purposely distorting data or results) and plagiarism (copying ideas or data without crediting the source). “The attention to deceive is a key element,” stated Fanelli’s May 2009 report, published in the open-access journal PLoS ONE.

More than a third of scientists admitted to at least one instance of engaging in "questionable research practices" - basically scientific misdemeanors that include “dropping data points based on a gut feeling” and “changing the design, methodology or results of a study in response to pressures from a funding source,” the report stated.

When asked the same questions about their colleagues, the numbers jumped dramatically: More than 14 percent of scientists had observed their colleagues engaging in serious misconduct at least once, and about three out of every four scientists had witnessed questionable research practices by their colleagues.

In the report, medical and pharmacological researchers reported a higher frequency of misconduct in their fields. This finding supports “growing fears that the large financial interests that often drive medical research are severely biasing it,” the report stated.

The analysis uniquely aggregated and distilled recent data to address the question – “how many scientists fabricate and falsify research?” – that is as old as modern science. “I figured out a way to extract information from these surveys that allows you to compare all of them as one,” Fanelli said.

Starting with thousands of surveys (from 1986 to 2005) that included questions about misconduct in research and publishing, he pruned them down to 18 surveys that matched a specific set of criteria that ensured scientific soundness.

“My findings were interestingly very consistent,” Fanelli said. “Nobody was really surprised, particularly scientists. Nobody really questioned it.”

But these numbers are likely “a conservative estimate of the true prevalence of scientific misconduct,” the report concluded, noting that historically people undervalue their questionable ethics on surveys.

“Science is made by human beings, and they’re not as objective as we’d always like,” said Fanelli, who was a former evolutionary biologist. The purpose of his research is to understand an important issue in the field he loves - "not to criticize or undermine science," he said.

Fanelli is currently working on two papers that extend his original analysis and “try to overcome the limitations of surveys.”