The full extent of fraud in the scientific and medical technical literature (including false or exaggerated research results, misleading claims, improper citations, and plagiarism) is difficult to assess. However, the extent of plagiarism, one aspect of scientific fraud, is easier to quantify, because the text of technical articles is open to full review.
Software has been developed that enables one to investigate the similarity between two technical articles, in terms of the text, figures, references, and other features of the two articles. This software has been used to identify possible cases of plagiarism.
However, there is little information available on how the authors of the manuscripts and the editors of the technical journals respond to possible evidence of plagiarism. Harold Garner (University of Texas Southwestern Medical Center) and coworkers have delved into this issue, finding that a disturbingly high percentage of the editors seemed to take no action at all.
The articles in question.
The scientists have previously developed a process to identify potentially plagiarized articles listed in MEDLINE, an online database of health and medical technical journals. They found 212 pairs of suspect articles, one in each pair that is possibly plagiarized by the other article.
The average similarity in text between the two articles is 86%. The average shared references percentage is 73%.
Seventy-one percent of the article pairs share a figure or table that is highly similar. Forty-two percent of the article pairs also share such features as inconsistent data and erroneous calculations.
Given that these 212 pairs of articles are so similar, does the subsequent article so much as cite the original article? No; only 22% of the original articles are cited by its alleged duplicate.
Responses to plagiarism.
It seems clear that most of these are examples of plagiarism. How do the authors and the editors of the manuscripts in question respond to the allegations?
The scientists forwarded the results of their investigations to all involved parties (authors and editors), and asked that they all fill out a confidential questionnaire. They received an 88% response rate.
Ninety-three percent of the orginal authors were unaware of the possible plagiarism. This uniform response contrasts with that of the alleged duplicate authors.
In this group, 28% denied doing anything wrong, while 35% admitted their actions. Furthermore, 22% claimed that they were not involved in drafting the manuscript, and 17% claimed to be unaware that they were listed as an author on the manuscript.
These last two claims should not be viewed with skepticism. Manuscript authors do not always make a serious attempt to consult with all of those whom they list as coauthors.
What was the response of the editors? Among the editors who responded to the questionnaire, the scientists' findings led to internal investigations 58% of the time, and article retractions 32% of the time.
This says that 42% of the time, clear and substantiated allegations of plagiarism seems to have gone entirely uninvestigated by the editors of the technical journals. This seems to be a serious dereliction of duty on their part.
How to act on these findings.
The alleged duplicate articles were only cited 7% as often as the original articles. While some may say that this means the alleged duplicates have had little impact, plagiarism should not be tolerated on the basis of expediency or poor public relations.
It would be extremely unfair to expect technical journal editors to uncover all scientific dishonesty in their articles. After all, one who is intent on fraud will often find a way to commit it.
However, once provided with evidence of fraud, editors have a responsibility to fairly investigate the matter, and take any necessary corrective action. Tolerating scientific fraud only encourages some scientists to engage in this unethical practice.
The most effective method to combat fraud in the technical medical (and science) literature would seem to be imparting a strong sense of ethics among scientists and medical professionals, and reducing the pressures put upon some of them that may encourage fraud, such as the demand for research results or grant money. In the meantime, editors of all technical journals, whatever their perceived prestige, should take action on the matter when it arises in their journals.
for more information:
Long, T. C.; Errami, M.; George, A. C.; Sun, Z.; Garner, H. R.
Responding to possible plagiarism.
Science 2009, 323, 1293-1294.