Comments on Definitions of Research Misconduct
Below are some previous definitions of research misconduct that bear on the development of the Federal Policy, along with comments from Dr. Donald Buzzelli, then of the NSF Office of the Inspector General, on the NSF definitions.
Previous Definitions
Three categories of activity are commonly held to count as misconduct.
- Falsification
- Misrepresentation of results.
- Fabrication
- Reporting on experiments never performed.
- Plagiarism
- Taking the writings or ideas of another and representing them as one's own. (Copyright is a legal notion that does not cover ideas.)
-They are held to jointly define misconduct, or they are held to be good examples of the serious deviations from accepted practice that constitute misconduct.
Controversial Clause
"Other Serious Deviations from Accepted Research Practices."
This clause is included in definitions used by the National Science Foundation (NSF) and by the U.S. Department of Health and Human Services (HHS), which includes the National Institutes of Health (NIH). This clause is not included in the definitions from the National Academy of Sciences (NAS) and from MIT. NAS and MIT regard falsification, fabrication, and plagiarism as definingmisconduct, rather than as examples of it.
Comments from Donald Buzzelli
These views of Dr. Buzzelli do not necessarily represent those of the National Science Foundation or its Office of Inspector General.
NSF's definition says, "fabrication, falsification, plagiarism, or other serious deviation from accepted practices. . ." This means that practices that seriously deviate from the ethical standards accepted in the relevant community of scientists are misconduct in science. Fabrication, falsification, and plagiarism (ffp) are three broad classes of practices that illustrate such serious ethical deviations. Therefore, "serious deviation from accepted practices" expresses the fundamental notion of what misconduct in science is. This is a standard that is used daily in OIG to make judgments about whether something is misconduct in science, and NSF's Deputy Director uses it in deciding whether to make a misconduct finding. Universities are also asked to use it in dealing with NSF-related cases, and they are able to do so if the idea is explained to them.
This standard has several advantages:
- It provides a rule for interpreting "fabrication," "falsification," "plagiarism," or whatever else is listed in the definition. A rule of interpretation is needed. The definition never meant to say that anything whatever that can be described as fabrication, falsification, or plagiarism is misconduct in science. In fact, any one of these terms covers some practices that are "not really meant" because they are trivial, irrelevant, or whatever. (I expect that any other list of terms would present the same problem.) Serious deviation from accepted practices is a standard for deciding what actions falsification, fabrication, plagiarism, and similar terms should be understood to cover.
- The "serious deviation" standard makes the scientific community the ultimate arbiter as to what is misconduct in science. Adjudications of misconduct cases should be based on the consensus of scientists in a given field as to what are and are not acceptable practices. An alternative approach is to let the agency decide what it means by falsification, fabrication, plagiarism, or whatever. However, I don't think the agency should enforce a standard of its own creation rather than the ethical standards of the scientific community. For one thing, the agency will be tempted to stretch the meaning of ffp to fit cases where bad things have been done that ffp does not cover but that the agency wants to deal with, and to shrink the meaning of ffp to exclude cases that ffp literally covers but that the agency prefers not to deal with.
- The "serious deviation" standard restricts misconduct in science to deviations that are "serious." Any kind of unethical practice includes some instances that are so inconsequential that they should not be treated as misconduct in science. A definition consisting of "ffp" and nothing else would not allow for this. For example, if taken literally, an "ffp only" definition would require every plagiarism case, however trivial, to be treated as a misconduct case that required formal investigation and adjudication. This kind of abuse can be avoided by not taking the definition literally, but the NSF definition does not make it necessary to evade the literal meaning of the definition in that way. The standard for what is "serious," again, is the ethical standards held by the relevant community of scientists.
- The NSF definition allows the agency to deal with practices like misrepresentation of credentials in proposals or tampering with colleagues' experiments (there are many other examples) that it has to take action against as a responsible steward of government funds. Without the "other serious deviation" provision, either such practices must be fit under an excessively broad interpretation of ffp, or else they cannot be treated as misconduct in science. In that case, the agency has to create another regulation under which to place them (requiring another definition, I assume) or not deal with them at all. Using a longer list than ffp in the definition would address this problem to some extent, but there will always be the possibility that someone will commit a serious offense that is not on the list.
In summary, the "serious deviation" part of the NSF definition is sometimes taken as a meaningless "etc." that is appended to ffp and that can be interpreted in any way the agency likes. On the contrary, experience has shown that it meaningfully shapes OIG's recommendations and NSF's decisions about misconduct in science. The "serious deviation" standard limits the interpretation of "ffp," and in that sense it actually makes the definition more narrow and precise. Basically, however, NSF's definition is neither broader nor narrower than a bare "ffp" definition. It is an entirely different kind of definition from "ffp only" or any other definition that consists of a simple list of terms, because it gives serious treatment to the problem of interpretation.
Further Commentary
- A 1996 article from the AAAS Professional Ethics Report
- HHS Commission on Research Integrity Chair Kenneth J.Ryan discusses the definition drafted by the commission in 1995.