Elizabeth Loder on making publication ethics matter to authors

Elizabeth LoderThese are heady times for the Committee on Publication Ethics (COPE). Its membership has broadened considerably in recent years. It now boasts over 6000 member journals and is no longer mainly biomedical.  The attendance roster for the second US COPE seminar, held in late November in Washington, D.C., reflected this diversity. It listed editors from such publications as the Journal of Bridge Engineering, the Journal of Research on Christian Education, Plant Physiology, and Police Crisis Negotiations, in addition to traditional medical publications such as Lancet and BMJ.

COPE chairman Liz Wager noted in her opening remarks that there was reason to think COPE’s focus might change with the influx of non-biomedical journals, but, “actually, the problems we see are the same… academic rivalry, pressure to publish, and the involvement of money” can cause problems in any area of science.  “Ethical editing is ethical editing,” Ms Wager said, and COPE’s aim remains the promulgation of international ethical standards for editing. There is, however, a new awareness of the need for those standards to have a “broad frame of reference” and relevance to all scientific publications. In the last year, COPE has published retraction guidelines and is preparing author and editor guidelines for release early next year. COPE also plans an e-learning module, a guide for new editors, forum podcasts, and discussion papers on thorny topics where consensus has not emerged (such as what constitutes “publication”).

The seminar got underway with a talk by Dr Harold Sox, former editor of the Annals of Internal Medicine. He spoke on plagiarism in the electronic age. He would begin, he said, by “trying to impress upon you just how serious plagiarism is,” operating on the premise that “the scientific record should be sacrosanct.” A corollary of this, he said, is that “fixing errors and removing fakery is an obligation of the scientific community…It’s not optional.” He recounted several cases of plagiarism and concluded that offenses that are “not very bad” should often trigger a more extensive investigation, since if someone would fake one thing they might fake others.

The electronic age, Dr Sox said, makes plagiarism easier to commit but also easier to identify. This led naturally to a presentation by Carol Meyer of CrossRef, a not-for-profit association of 128 publishers that does things that “individual publishers can’t do for themselves.” Participating publishers have the options of using a CrossRef service called CrossCheck, which uses Ithenticate software to scan submitted papers against a database of previous publications and the world wide web. CrossCheck generates a percentage figure that represents the proportion of the submitted material that matches content in these sources. It is then up to a human being to evaluate the result, since there are many innocent reasons for an apparent high degree of similarity. Ms Meyer’s slides from the meeting are posted here.

Lance Small, the editor of an algebra journal, spoke next about ethical problems in the world of mathematics. The mathematical world, he told us, is “simpler, but not much.” A longstanding convention in the field of mathematics is that authors are listed in alphabetical order.  It’s less clear, however, what constitutes plagiarism in a discipline where papers build closely on previous work and incorporate equations and theorems whose original authors and sources may not be clear. He did, however, point to one clear example of mathematical plagiarism (involving an exposition of the Poincare conjecture) that was interesting enough to merit coverage in a New Yorker article.

The meeting ended with a discussion of thorny cases. There was broad but not universal consensus about the ethical matters involved, but far less agreement about the appropriate response to transgressions. Referring the matter to the authors’ “institutions” was often mentioned, but academic institutions vary in their willingness and ability to pursue publication problems. It is also unclear whether they share the world view of journal editors and publication professionals about fundamental aspects of publication ethics. And then there is the problem that not all authors are affiliated with institutions. Even in this era, some investigators work alone or in arrangements that limit their connection with institutions.

I left this second COPE meeting full of admiration for what has been achieved by the group in a short time. We’ve come a long way in establishing standards for publication ethics. In the end, though, I was struck by the fact that mandates and standards get us only so far in the absence of reliable enforcement mechanisms. COPE’s good ideas are unlikely to produce desired changes without attention to a universal and workable enforcement process that will make them matter to prospective authors.

Elizabeth Loder is the BMJ’s US based clinical epidemiology editor.