The 2012 Committee on Publication Ethics (COPE) North American Seminar, held on 19 October at the Bechtel Conference Center in Reston, Virginia, focused on the importance of correcting the scientific literature. Speakers and members of the COPE Council discussed ways to handle expressions of concern, corrections, and retractions. The seminar was attended by more than 50 science, technical, and medical editors in various roles and disciplines.
Retraction of Scientific Papers: The Science Experience
COPE defines retraction as a mechanism for correcting the literature and alerting readers to publications that contain flawed or erroneous data. Barbara Jasny, deputy editor of Science, shared her journal’s best practices in dealing with retractions. Her first piece of advice: Rapid online retraction can do harm. Jasny recommended that journals exercise caution and ensure that a fair process has been implemented before issuing retraction notices and editorial expressions of concern. She emphasized that each case is different; science editors must evaluate each scenario carefully to ensure that appropriate steps are taken. For example, if a retraction request were prompted by an anonymous whistle-blower, it is important to ask about the whistle-blower’s motivations. According to Jasny, when Science receives a credible report from an anonymous whistle-blower, it requests identification of the person making the report with the understanding that he or she will remain anonymous throughout the retraction process.
Jasny presented three kinds of retractions: good—retractions are initiated or agreed upon by the original authors; bad—one or more authors refuse to sign the retraction; and ugly—authors refuse to accept findings of an institutional investigation.
The retraction process takes staff time and journals’ resources. What can journals do to help prevent retractions? At Science, all authors are required to be responsible for the submitted work. According to Jasny, the requirement comes at a time of increased multidisciplinary papers, which makes it more difficult for journals to verify author responsibility. As a result, Science asks for specific author activity on the basis of the level of participation measured by percentages.1 In addition, senior authors are asked to affirm that they have personally checked all original data. Science’s authorship policy follows authorship requirements presented in On Being a Scientist: A Guide to Responsible Conduct in Research, third edition, published by the National Academy of Sciences.2
The Hwang et al papers that Science published in 2004 and 2005 and later retracted in 20063 prompted the journal to revisit some of its policies.4 Since the Hwang et al retraction, Science has implemented the following procedures: Alerts all coauthors when an author submits a paper with their names on it; requires all authors of accepted manuscripts to affirm and explain their contribution to the manuscript and any conflict-of-interest disclosures; sends submitted figures through a digital image check; and asks authors to ensure that all data necessary to understand and extend conclusions of the manuscript are included so that they are available to readers. Science does not allow references to unpublished data; all references and data must be available at the time of publication.
Jasny attributes the rise in retractions to large data sets, the possibility that supplemental material is not as rigorously reviewed as the main article, an increase in interdisciplinary papers, miscommunication, and pressures on scientists associated with grants, prestige, and public attention. Jasny identified risk factors that can help journals to flag papers that should receive a higher level of scrutiny: the multidisciplinary nature of the work, results that seem too good to be true, involvement of multiple laboratories in different countries, and fast turnaround in data preparation.
Jasny mentioned an article by Casadevall et al in the Proceedings of the National Academy of Sciences of the United States of America that reported that of the 2,047 biomedical and life-science research articles indexed by PubMed as having been retracted on 3 May 2012, only 21.3% were attributable to error. Meanwhile, 67.4% of retractions were attributable to misconduct (43.4%), duplicate publication (14.2%), or plagiarism (9.8%).5
Jasny acknowledged the complications associated with retractions but emphasized that problem papers are in the minority.
CrossMark: There Is No Final Version
CrossRef’s Carol Meyer provided updates on how CrossMark is helping journals to deal with the increasing incidence of retractions.6 Readers need to know when scientific literature changes, said Meyer. She showed five samples of actual journals and asked attendees to try to identify quickly whether any of the articles were corrected or retracted. Most of the samples were clearly marked; a couple required a few extra seconds to identify whether they were linked to a correction or retraction.
Meyer noted that although the Internet helps to disseminate scientific literature, search results can include republished PDFs, outdated press releases, and other unidentified sources, which leave readers in the dark as to whether a published work has been corrected or retracted.
To help in identifying corrected and retracted literature, a CrossMark ribbon logo is added to documents published by participating CrossRef members. By clicking on the CrossMark ribbon, a reader can identify whether the literature is current or is linked to a published correction or retraction.
CrossMark is available to all CrossRef members, and participation is optional. Participants must maintain their content, keep CrossMark metadata up to date, and adhere to CrossMark logo display guidelines.
Legal Issues in Corrections, Retractions, and Expressions of Concern
Mark Seeley, senior vice president and general counsel of Elsevier, offered advice on how journals should handle legal correspondence, which he said often deals with patent issues, defamation, or allegation of ethics violations. For someone to take the time to have a lawyer write to a journal is superfluous, said Seeley, in that anyone can contact a journal or institution directly to air concerns.
According to Seeley, patent issues are not for journals to solve. Inventor–authors have a short window to file their patents, so they are not always mindful of the legal requirements. Publication of a paper is considered public disclosure.
A journal’s primary defense against defamation is clearly stated policies, Seeley said. In defamation cases, a researcher’s or academician’s reputation may be at stake. A journal’s objectivity and due process are also being questioned. He emphasized the need for consistent, well-communicated journal policies.
Legal correspondence that involves allegations or refutations is typically related to fraud, plagiarism, unreported conflict of interest, or submission-process misconduct. Seeley advised treating such legal correspondence like any other complaint: evaluate the alleged act to determine whether it violates publication ethics and whether it matches the journal’s policy, Seeley added. He suggested reaching out to trusted experts to help in evaluating the allegations.
Seeley also noted that before editors of journals consider reporting the violation to an author’s institution, they should provide an opportunity for an alleged wrong-doer to explain and defend against allegations.
The courts are not unsympathetic but are reluctant to intervene in these matters because they recognize that there are other forums for them. “The courts seem to respect the scientific process and are not overawed by legalistic complaints,” he said.
Closing remarks by COPE Council members Geri Pearson and Charon Pierson underscored the value of well-written retraction statements. They recommended making sure that statements are linked to manuscripts, clearly identified, published promptly, and freely available.
A retraction statement should clearly note who issued the retraction and the reason for it and should avoid defamatory language, Pearson said. In handling retractions, it is important for journal editors to move cautiously but decisively. The retraction process can put a strain on smaller specialty journals that have small staff and limited resources, Pierson said. Editors should turn to their trusted experts and not make decisions in isolation.
- Science/AAAS authorship form and statement of conflicts of interest. Science. 2012; http://www.sciencemag.org/site/feature/contribinfo/prep/coi.pdf.
- Committee on Science, Engineering, and Public Policy. On being a scientist: a guide to responsible conduct in research. The National Academies Press. 2009; http://www.nap.edu/catalog.php?record_id=12192.
- Special online collection: Hwang et al. controversy–committee report, response, and background. Science. http://www.sciencemag.org/site/feature/misc/webfeat/hwang2005/
- Vogel G. Picking up the pieces after Hwang. Science. 2006;312(5773):516–517. http://www.sciencemag.org/content/312/5773/516.full?sid=6d792ee0-2c02-44bc-b746-07ac52812a35.
- Fang FC, Steen G, Casadevall A. Misconduct accounts for the majority of retracted scientific publications. Proc Natl Acad Sci. 2012 (published ahead of print Oct 1, 2012); http://www.pnas.org/content/early/2012/09/27/1212247109.full.pdf+html.
- Van Noorden R. Science publishing: the trouble with retractions. 2011 Nature. ;478:26ñ28. http://www.nature.com/news/2011/111005/full/478026a.html.