© 2002 American Thoracic Society
AJRCCM's Policy on Duplicate PublicationInfrequently Asked QuestionsEditor The policy of AJRCCM on duplicate publication is clearly stated in our Instructions for Contributors (viewable online at http://www.thoracic.org/publications/ajrccm/bluecont2a.asp ). Yet authors continue to send us manuscripts without telling us about a closely related manuscript they had submitted to another journal or had in press. We have had ten transgressions, varying in degree, of our policy over the past two years; these authors also failed to list the second manuscript in the references of the submitted manuscript. What is a duplicate publication? A duplicate publication is an article that overlaps substantially with another manuscript, although the texts are rarely identical (1–3). Other names that describe the practice include dual, repetitive, fragmented, and disaggregated publication (1, 4, 5). "Salami science" refers to the slicing of a dataset into several pieces, called least publishable units. Generating more than one manuscript from a single dataset can be legitimate, provided the manuscripts are cross-referenced and each addresses distinct and important questions (3, 6). If authors believe this to be the case, it is still imperative that they state this point in a cover letter. AJRCCM has only two exemptions to its policy on duplicate publications. One is an abstract containing no more than 400 words, such as those submitted to the International Conference of the American Thoracic Society. A second is a press report made by a journalist attending a public meeting. If, however, the author gives written material to the journalist, the exemption no longer holds. Disseminating material on a website server is not exempted (7), nor is the publication of research findings in a letter to the editor. Our policy applies not only to original research articles, but also to overlap between review articles. What do authors need to do? Our Instructions for Contributors state that when submitting a manuscript to AJRCCM, "all other data submitted, in press, or published, which is potentially overlapping or related to the actual data presented, must be attached. These documents can be uploaded online for an online submission, or mailed to the Peer Review Office for a manuscript submitted in hard-copy format." On the Mandatory Submission Form, authors must state whether or not the submitted manuscript overlaps with another manuscript. Authors must also include a cross-reference in the submitted manuscript to overlapping work. Some authors see that two manuscripts overlap, but they consider the overlap insufficient to constitute duplicate publication. At AJRCCM, we wish to make up our own mind as to whether the overlap is significant or not. No matter how small the overlap, authors need to inform us about related work. If any of the control data in a manuscript are also included among the control data in another manuscript, we wish to be informed. A statement about overlap is best made in the methods section of a manuscript, although informing us in a cover letter will also suffice. That a manuscript contains data that overlap with data in another manuscript will not in itself cause a manuscript to be rejected. In the case of a review article, we wish to be informed about other review articles that the authors have published on a related topic. When authors are in any doubt as to what material to enclose with a submitted manuscript, they should err on the side of including too much rather than too little. If authors still have doubts about whether a submitted manuscript overlaps with another, they can write to me (mtobin2{at}lumc.edu). The best tool for preventing problems related to ethics and policy in scientific publishing is disclosure made by authors to editors. If, at any stage of the publication process, authors have doubts about duplication of publication, AJRCCM wants to be informed. Why do duplicate publications occur? When confronted with an allegation of duplicate publication, authors offer the following answers: "We did not read the instructions," "We wanted to reach a different audience," "Our failure to cross reference the other article was a simple oversight," "We perceive the overlap to be much less than the reviewer or editor thinks," and "We now see that we broke the rules, but this was never our intent." One answer is never given: "We thought this would be a good way of lengthening our curriculum vitae." Like it or not, publish or perish is a reality of academic life. Productivity, measured as the number of papers published, is heavily valued by promotion committees and grant-giving agencies. But productivity and funding should not be confused with contribution to the research enterprise. The whole purpose of research is to advance understanding. And dollars received provide not even a hint of a researcher's contribution to better understanding. Making the fundamental distinction between contribution to better understanding versus productivity would eliminate the incentive for duplicate publication. Editors may also contribute to the problem of duplicate publication (8). During peer-review, editors sometimes tell authors to delete sections of a manuscript. Authors may later decide to build a new manuscript around the deleted data. Editors also push authors to write concisely. Some authors misinterpret this advice, and believe a journal wants to publish only brief reports. Let me emphasize that AJRCCM prefers detailed manuscripts, often incorporating several experiments that address a single research question. We place no limit on the number of words in a research paper. That said, we also want authors to avoid long-windedness and redundancy. AJRCCM is changing its policy to allow the designation of more than a single first author when two authors have made equal contributions to a study, in the hope that this will decrease one motivation for duplicate submissions. Why should duplicate publications be avoided? One, before turning to a research report, a reader wants assurance that it contains new and important information (9). It is this very assurance that determines a journal's reputation. If reviewers and editors are unaware that a submitted manuscript overlaps with another, they may incorrectly conclude that it contains unique information. Publishing rehashes harms a journal's reputation, and will decrease its ability to attract good papers. Two, if the results from a randomized controlled trial are reported in two articles, they may be entered twice in a meta-analysis (4). Likewise, reporting the results of laboratory experiments in two articles overemphasizes their significance (10). Because clinical and scientific judgments need to be based on evidence, duplicate publications may ultimately result in decisions that hurt patients (1). Three, duplicate publications distort the whole purpose of journals (4, 11). Scientific journals exist to help researchers communicate new information so that others can build on it. At the end of the day, the only meaningful legacy of researchers is their contribution to the store of knowledge, not their fame. Duplicative publication colludes with the notion that publishing for the sake of publishing is meritorious in itself. Instead, authors should heed Scott Fitzgerald's maxim: write because you have something to say, not because you want to say something. How do allegations of duplicate publication arise? Most allegations arise during peer review of manuscripts. For any specialized area, the pool of expert reviewers is small. When authors send related manuscripts to two journals, there is a considerable likelihood that the two editors will invite the same reviewer to review both manuscripts. Diligent reviewers also enter an author's name into PubMed and learn of undisclosed related work. After an article has been published, readers may alert editors to undisclosed overlapping work in another journal. What happens when an allegation of duplicate publication is made? As soon as an allegation is made, I contact the editor of the other journal to learn whether she or he had been informed about the overlap. I then inform the authors of the allegation, and give them an opportunity to defend their action (provision of "due process" to the authors). Depending on the details, the manuscript may be rejected, a notice of duplicate publication may be posted, and the authors may be banned from submitting manuscripts for a period of time. The implications differ depending on whether the journal is informed by the authors or by a third party (12). When the authors make the disclosure, the issue is one of judgment and policy. When the journal is informed by a third party, the problem is one of ethics. If any authors believe that a paper of theirs in AJRCCM represents a duplicate publication, we wish to be informed. By connecting notices to previously published papers, MEDLINE alerts readers to the duplication of research reports. Editors do not wish to be regarded as prior-publication police (13) because it is important that authors can look on editors as colleagues. Ultimately, it is authors who guarantee the integrity of the scientific literature. What is integrity of the literature? All professionals rely on the literature of their discipline, and have a vested interest in its integrity. The highest responsibility of editors is to ensure that every item published in their journal satisfies the highest standards of scientific integrity (12). Progress in research is a community endeavor, and the community is formed around the sharing of information. Authors want the greatest possible visibility for their work, but self-centered desire must be subsumed to the needs of the larger community. The ultimate responsibility for the integrity of the literature rests with the individual author (14); and the rarity of aberrations is a major factor in the success of science. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
