One of the responsibilities of the Inspector General's office at the National Science Foundation (NSF) is to deal with allegations of misconduct that NSF receives in connection with its proposals and awards. In my position within that office, I have had the opportunity to observe a wide variety of allegations. Plagiarism and related phenomena make up a large part of our case load. My conclusion is that the scientific community should review its standards and practices concerning plagiarism.

NSF's regulations on misconduct in science and engineering contain the following definition: "Misconduct means (1) fabrication, falsification, plagiarism, or other serious deviation from accepted practices in proposing, carrying out, or reporting results from activities funded by NSF; or (2) retaliation of any kind against a person who reported or provided information about suspected or alleged misconduct and who has not acted in bad faith." (45 CFR 689) Thus, plagiarism is explicitly defined as misconduct in science.

In spite of this, NSF does not have its own definition of plagiarism, and there are different views among scientists as to what offenses the term "plagiarism" should cover. Everyone probably agrees that copying another person's original text into one's own publication without acknowledging that it was copied and without giving a reference to the source is plagiarism.

Outside this central concept, however, many possible violations of intellectual property may or may not be counted as plagiarism. If the text that was copied is considered to be derivative itself, or to contain no novel ideas, the argument may be made that there was no plagiarism because no ideas were stolen. On the other hand, ideas may be stolen without copying any text. Sometimes the ideas taken were never written down but merely communicated orally. Some would call this plagiarism, while others might call it "intellectual theft" and place it under the "other serious deviation from accepted practices" part of the NSF definition. It may make a difference whether one perpetrator writes down and publishes the stolen ideas, as opposed to using them in the design of his or her own research. Some would argue that plagiarism occurs only if the stolen text or ideas were used in a published paper, as opposed to a grant proposal or an oral presentation. However, NSF has consistently rejected the position that, since a grant proposal is not a publication, copying in a proposal cannot be the basis for a plagiarism case.

The term “plagiarism” is sometimes raised in connection with simple priority disputes or other squabbles between scientists. For example, a complainant will occasionally charge plagiarism when he or she only means that someone published something in the complainant's field without citing the complainant's work to the extent that the complainant considers fitting. Collaborators on a research project may have a falling out or a simple failure of communication that leads to charges of plagiarism. One collaborator may publish material from the project without the knowledge or permission of the others. There is often disagreement as to how much each contributed to that material. A charge of plagiarism is especially likely to arise if one collaborator uses text or ideas in a publication while other collaborators believe they were the original source. Again, some would call such unauthorized publication plagiarism, while others would not.

There are special problems if some of the collaborators are graduate students or postdoctoral researchers. Such individuals are very much under the authority of the research director. Their intellectual property rights are not fully defined. A graduate student or postdoc may be asked to write a research proposal for the research director, and may then find that he or she is excluded from the project in the final version. The junior person may then feel that the text and ideas have been stolen. Disagreements also arise over the publication of the results of supervised research. The research director may publish the results without giving the junior researcher the credit to which he or she feels entitled.

Plagiarism may involve diagrams and charts as well as texts. It may also involve graphs or tables containing data. If a scientist plagiarizes the data of another, a charge of data fabrication may arise that is possibly more serious than plagiarism itself. A similar situation arises in connection with the peer review of proposals. The research proposals that NSF receives are sent out for review by scientists knowledgeable in the field of science involved. Since they may be doing research in the same field, these reviewers are specifically instructed not to use the proposal to their own advantage by copying, quoting, or otherwise using material from it. If a reviewer violates these instructions and plagiarizes a proposal, this violation of the confidentiality of peer review may be a greater offense than the plagiarism itself.

There are many other opportunities in the proposal process for allegations of plagiarism. One investigator may simply send in a proposal that was written by another. If this was done without the original writer's authorization, it is clearly plagiarism. However, even if the original writer consented, one may want to ask whether "plagiarism with permission" isn't still plagiarism. Certainly some kind of misrepresentation of authorship seems to be occurring. From NSF's viewpoint, it is important to know who wrote the proposals NSF receives. NSF revised its proposal cover page in April 1992 to include a certification that the proposal was written by the signatories or by individuals working under their supervision, unless otherwise indicated.

In a typical year, NSF receives about fifty allegations of misconduct in science. About half of these will be about an intellectual property violation of some kind. (Another ten percent will be mainly data related, and therefore will cover what NSF's definition calls falsification or fabrication. About fifteen percent will be other kinds of misconduct in science, and twenty-five percent will ultimately be judged not to be allegations about misconduct in science at all.) Many of the cases that have come in have not been closed, and many others have been closed with findings of no misconduct.

However, among the cases that have been closed with findings of misconduct, plagiarism cases are predominant. Specifically, the cases received at NSF since the beginning of 1989 have led to three individuals being found guilty of misconduct in science by their universities. All of these were major plagiarism cases. In four cases (including two of the cases just mentioned) NSF made its own findings of misconduct in science. Three of these were also plagiarism cases.

It may be comforting to scientists that these numbers are so small. Still, these major cases emerge from a great number of less serious intellectual property offenses and numerous squabbles and misunderstandings. The frequency and variety of intellectual property complaints indicate that the scientific community needs to clarify the rights that individuals have in different situations and to communicate these standards better to working scientists. I would be very happy to see fewer cases of plagiarism.

The views expressed here are mine. They are not necessarily shared by the Office of Inspector General, the National Science Board, or the National Science Foundation.