Editorial: Misconduct in Scientific Publications

Abstract

Several recent high visibility cases of fraud in scientific papers have attracted the attention of the public in the United States, as indicated by an editorial in the New York Times on 1 June 2015 (“What Happens When Scientists Cheat”). These include two cases of publishing apparently fabricated data in papers in Science and in the Journal of Clinical Investigation (JCI) and one case in which the lead author of a paper in Environmental Science & Technology (ES&T) failed to disclose a financial conflict of interest. Science and JCI have now retracted those papers, and ES&T published a correction that states the conflict of interest: the author had been supported as a consultant to perform the research in question by a corporation that had a major interest in the outcome of that research. The New York Times editorial stated that “cheating in scientific and academic papers is a long-standing problem, but it is hard to read recent headlines and not conclude that it has gotten worse.” In fact, reliable data on trends in publishing fraud are scarce, but we should all be concerned. Retractions of scientific papers in the biomedical and life sciences have increased 10-fold since 1975, with most retractions due to deliberate fraud (Fang et al. 2012), although whether the increase is due to greater malfeasance or better journal practices is not clear (Scott-Lichter et al. 2012 2012). And a 2009 meta-analysis of survey data indicates that approximately 2% of scientists admit to having fabricated or “modified” data and over 30% indicate they have participated in some “questionable practices” (Fanelli 2009). My personal belief is that fraud is relatively rare in our fields of limnology and oceanography. Many of the problems in the biomedical literature can perhaps be traced to the strong influence of funding from the biomedical industry. Our disciplines receive less funding from industries that could lead to bias, although this may be a growing concern as climate change and environmental pollution become more politically charged issues, and as public funding continues to fall. As a community, we should remain vigilant: any fraud is serious, and the science community and scientific journals should have a zero-tolerance policy for fraud. The Times editorial focuses on the role of data accessibility for minimizing fraud, stating that “the scientific community clearly needs to build a better safety net” and concluding that this could “start by ensuring that scientists, especially peer reviewers, are allowed to see the underlying data of a paper, which researchers are typically reluctant to share.” I applaud the Times for highlighting the problem with fraud, but find their suggested solution to be lacking. It would not, for instance, have prevented the failure by the lead author of the ES&T paper to disclosure a conflict of interest. Further, the term “underlying data” is vague. If the Times means the raw data upon which analyses, tables, and figures are based, this would in many cases – at least in our fields of limnology and oceanography – be unworkable. To examine and begin to understand the raw data collected in another lab is a daunting task, and to ask a peer reviewer to independently verify calculations and analyses based on such data is beyond what can reasonably be expected. In the words of the Council of Science Editors, “peer review is usually a gift of uncompensated time from scientists to whom time is a precious commodity,” and “therefore, it is important … . to implement processes that streamline the peer review process as much as possible” (Scott-Lichter et al. 2012). Increasing the burden on peer reviewers is counter-productive, since they are such a critical part of quality publication, including capturing fraud. Of course, good journal practices have always stressed the need for peer reviewers to have sufficient access to data, and peer reviewers must feel confident that they are being given enough information to make an informed decision on the reliability of the data in a submitted manuscript. If the information presented in the manuscript is not adequate for this, reviewers should (and often do) ask for more information, either to be presented in the paper by means of additional figures or tables or to be available as on-line supporting information. At Limnology & Oceanography, we fully support such requests from reviewers and would not publish a submission where the author is unwilling to share the data in this manner. What can be done, beyond continuing to make data available to reviewers, to reduce scientific fraud? Minimizing fraud is the responsibility of the entire scientific community. The first line of defense is for the community to embrace a value system that is intolerant of fraud at any level, for us to do our best as individual scientists to avoid even the perception of fraud, for us to expect the same from our colleagues, and for us to provide clear ethical guidance to students and other young scientists. Limnologists and oceanographers tend to do well in meeting these goals, I believe, but we should not be complacent and we should strive to continually do even better. We need to encourage open debate and discussion on the issue of fraud. Another line of defense, and one that is perhaps not emphasized enough, is the role of co-authors. Not all papers have multiple authors, of course, but most papers in Limnology & Oceanography have several authors. These co-authors are responsible for the content of the paper, and should not be co-authors unless they are entirely confident in the quality and reliability of data being published, and in their interpretation. They also bear responsibility for ensuring that all authors fully and truthfully disclose conflicts of interest, that material is in no way plagiarized, and that due credit to the work of others is fully acknowledged. The guidelines for Limnology & Oceanography specify that all authors of a paper “should have contributed substantially to the study's conception, data acquisition, or analysis; contributed substantially to drafting the manuscript; and approved the final submitted manuscript.” The submitting author for each submission, including revisions, certifies that all authors meet these criteria and have been fully engaged in the submission. This requirement for shared responsibility and active engagement by all authors is likely to lead to a higher quality publication, but it also reduces the likelihood of fraud. In the rush to publish a paper, carelessness can sometimes creep in. The co-authors can and should play an important role in ensuring that ethical corners not be cut. For any journal, the ultimate defense against fraud lies with the editors and peer reviewers. At Limnology & Oceanography, we have a large and diverse editorial board of Associate Editors, and we do our best to assign submitted manuscripts to Associate Editors who are truly expert in the subject matter of the submission. The Associate Editors, in turn, select peer reviewers who are very close to the subject matter. These individuals are then well equipped to evaluate the data behind a manuscript, to observe unusual aspects which may potentially indicate fraud, and to point out these potential problems to the journal editors (in addition to performing their primary job as reviewers, which is to find ways to improve a submission and to help us determine which papers are most appropriate for the journal). When journals fall into trouble with publishing fraudulent materials, I suspect it is often because the peer reviewers are simply not familiar enough with the area of the paper, and this in turn can easily result when editors are not sufficiently expert in the specialty and so choose inappropriate reviewers. Insufficient effort by a reviewer can contribute to this, and becomes more likely when the subject of a submission is rather distant from their areas of expertise. I am fairly confident that fraud is rare in our journal, thanks to the strong sense of ethics shared by our community, and to our very talented and dedicated Associate Editors and peer reviewers. Nonetheless, we should remain attentive. We have guidelines in place for addressing suspected fraud, generally following the guidance of the Council of Scientific Editors and the Committee on Publication Ethics. And I welcome suggestions as to how we can improve Limnology & Oceanography, including ways to even further reduce potential fraud. Please do not hesitate to contact me or any of our Associate Editors if you see ways we can do better. Robert W. Howarth Ithaca, NY

Ähnliche Arbeiten