New Federal Policy Sharpens Definition of Scientific Misconduct
By JEFFREY BRAINARD
Washington
The Clinton administration has published a new policy that defines scientific misconduct and describes how universities should investigate it. The policy offers clarifications but no major changes from an earlier draft that many researchers had praised, but some had called inadequate.
The policy, published in Wednesday's Federal Register, is meant to deal with confusion and complaints voiced by researchers about the government's previous policies on scientific misconduct. The National Institutes of Health, the National Science Foundation, and other federal agencies that finance research all have used slightly different definitions of scientific misconduct. The new standard is meant to provide a uniform rule.
The new language is somewhat narrower than some of the existing misconduct policies. Like the others, it includes plagiarism and fabrication or fudging of data in the definition of misconduct. But the final policy dropped additional language, adopted in 1989 by the National Institutes of Health, that defined as misconduct practices that "seriously deviate from those that are commonly accepted in the scientific community."
Researchers have long chafed under that standard, complaining that it could include sloppy research practices or authorship disputes that were less serious than scientific fraud.
The new policy will take effect in one year. Until then, a committee representing the federal agencies that finance research will work to make their various regulations on research misconduct consistent with the policy. The goal of all the policies will be to protect the integrity of federally financed research.
The president's Office of Science and Technology Policy received more than 230 written comments in response to a draft it issued in October 1999.
Some commenters suggested significant changes. For example, critics urged the government to take a more-direct role in investigating misconduct at institutions that receive federal research funds, since the existing model gives the institutions themselves primary responsibility, subject to review by the government. But the final policy maintains the status quo. In a statement accompanying the final policy, officials of the science-policy office wrote that that "would have involved a substantial new federal bureaucracy, which is not thought desirable."
In addition, the policy does not encourage universities to adopt tougher in-house rules to supplement the federal policy, either, said C. Kristina Gunsalus, associate provost at the University of Illinois at Urbana-Champaign, who has served on several national committees on scientific integrity.
"I think it would have been really in the interest of the entire system if they had made a much stronger statement, in a positive way," to encourage such policies, she said. Nevertheless, Ms. Gunsalus added, greater uniformity of misconduct rules across federal agencies would represent "a major step forward."
University officials have also said that the federal government should provide immunity for university officials who are required by the policy to investigate allegations of misconduct. Some institutions -- and individual investigators -- have been sued by faculty members whom they accused of misconduct, leading in some cases to protracted litigation and large settlements.
In a statement accompanying the final policy, officials of the science-policy office wrote that providing such immunity "would require significant statutory or regulatory initiatives which will be explored separately from this policy."
In addition, the policy requires charges of misconduct to be proved by a preponderance of the evidence -- not by the tougher standard of "clear and convincing evidence," as some commenters requested.
Among the clarifications offered by the final version, researchers may be committing fabrication if they misrepresent their qualifications or ability to perform the research in grant applications to federal agencies.
The policy also notes that the rules cover federally financed research not only in science and math, but also in economics, education, linguistics, medicine, psychology, social sciences, and statistics.
Background articles from The Chronicle: