Original Article 'Mad scientists' emerge in study Researchers admit scholarly misdeeds Rick Weiss Washington Post Jun. 9, 2005 12:00 AM Few scientists fabricate results or flatly plagiarize the work of others, but a surprising number engage in troubling degrees of fact-bending or deceit, according to a survey of scientific misbehavior. More than 5 percent of scientists answering a confidential questionnaire admitted to having tossed out data because it contradicted their previous research or said they had circumvented some human-research protections. Ten percent admitted they had inappropriately included their names or those of others as authors on published research reports. And more than 15 percent admitted they had changed a study's design or results to satisfy a sponsor or ignored observations because they had a "gut feeling" they were inaccurate. None of those failings qualifies as outright scientific misconduct under the strict definition used by federal regulators. But they could take at least as large a toll on science as the rare cases of clear-cut falsification, said Brian Martinson, an investigator with HealthPartners Research Foundation in Minneapolis, who led the study appearing in today's issue of Nature. "The fraud cases are explosive and can be very damaging to public trust," Martinson said. "But these other kinds of things can be more corrosive to science, especially since they're so common." The new survey also hints that much scientific misconduct is the result of frustrations and injustices built into the modern system of scientific rewards. That could have profound implications for efforts to reduce misconduct, demanding more focus on fixing systemic problems and less on identifying and weeding out "bad apple" scientists. The prevalence of research misconduct has been uncertain, however, in part because the definitions of acceptable behavior have shifted. For scientists working with federal grant money, that issue got settled five years ago when the Office of Research Integrity drafted a definition: "fabrication, falsification or plagiarism in proposing, performing or reviewing research or in reporting research results." Martinson and two colleagues sent a survey to thousands of scientists funded by the National Institutes of Health and tallied the replies from 3,247 responses. Just 0.3 percent admitted to faking research data, and 1.4 percent admitted to plagiarism. But lesser violations were far more common, including 4.7 percent who admitted to publishing the same data in two or more publications to beef up resumes and 13.5 percent who used research designs they knew would not give accurate results.