The funny thing about publication ethics is that it always seems to be somebody else's problem. If you ask journal editors whether problems such as plagiarism, data fabrication, redundant publication or false authorship occur, they usually agree, wring their hands and say these are terrible, and probably worsening, issues. Yet, if you ask them if they face such problems at their own journals, they will often deny it1. Similarly, but perhaps less surprisingly, few researchers admit to committing misconduct themselves, but a much higher proportion believe that their colleagues have misbehaved2.

In this article, I will argue that publication ethics is everybody's business and that all sectors involved with scholarly publishing should examine their roles and the effects their actions can have.

Publishers: villains or experts?

The relationship between publishers and journal editors is often complex but the Committee on Publication Ethics (COPE) emphasizes that it ‘should be based firmly on the principle of editorial independence’3. However, a World Association of Medical Editors' Policy Statement notes that ‘The limits of editorial freedom are difficult to define in the general case’4. Similarly, Richard Smith, a former editor of the BMJ, has noted: ‘editorial freedom … cannot be total. I couldn't turn the BMJ into a soccer magazine because I'd got bored with medicine. Freedom must be accompanied by accountability’5. Smith has also written ‘everybody supports editorial independence in principle, although it sometimes feels to editors as if the deal is “you can have it so long as you don't use it”.’6

“… publication ethics is everybody's business …”

However, while it appears that ‘everybody’ supports the concept of editorial independence, it is less often appreciated that most editors receive little or no training to help them exercise their role responsibly. This lack of training means that editors often look to publishers for guidance. Also, since cases of serious misconduct are relatively rare, most academic editors will face only a few such cases during their editorial career. Therefore, professionals working for the publisher usually have more experience of handling cases of suspected serious misconduct than editors. Also, publishers can share experience between titles, learn from this experience, and develop sound policies. So, rather than the publisher being solely responsible for the financial and technical aspects of the journal, they often play an important role when ethical questions arise.

COPE was initially established by a group of editors to enable them to discuss difficult cases relating to publication ethics. Since 1997, it has grown into an international and interdisciplinary organization with over 7,000 member journals whose subscriptions (which fund COPE) are mostly paid by publishers7. Several major publishers have signed up all their journals so they can benefit from COPE's advice. During my period as Chair of COPE (2009–2012), I discovered that responsibility for handling cases often rested with the publisher, although, of course, they worked closely with the editor.

Occasionally, publishers, rather than editors, have brought cases to the COPE Forum and these have sometimes involved publishers seeking to rein in over-enthusiastic editors. When faced with allegations of serious misconduct or cases of disputed authorship, COPE generally advises that editors should not attempt to investigate the case themselves, but should refer the matter to the researchers' institution. However, editors sometimes rush in and try to arbitrate in such cases. This can have serious consequences and one journal was almost bankrupted by legal fees following an inappropriate editorial decision (on an authorship dispute) which was taken without consulting the publisher8.

“In many cases, publishers are instrumental in both setting and upholding sound policies …”

As Richard Smith has noted, the ‘pure editor concerned with science and quality and a grasping publisher bothered purely with revenue and profit’9 are, like most stereotypes, an over-simplification. In many cases, publishers are instrumental in both setting and upholding sound policies and, while allowing the editor to be independent, may also be an important source of advice on ethical matters.

Academic societies versus editors

The role of academic societies in scholarly publishing is often viewed as a positive and benign one. Yet, societies have occasionally interfered with editorial decisions to such an extent that editors have resigned or been dismissed. In 1999, the editor of the Journal of the American Medical Association was fired over a disagreement with the Association about an editorial decision.10 In 2006, two senior editors of the Canadian Medical Association Journal were dismissed following a dispute over editorial freedom11. Two former editors of the Annals of Internal Medicine have written that ‘there is an inherent friction between the society's journal editor and its executive officer. The mindset and mission of editors are frequently at odds with the understandable wish of the executive to control the society's affairs and realize as much income as possible for other activities’12.

While it appears fashionable to criticize commercial publishers for their profit margins13, such critics often overlook the fact that journals usually generate considerable income for their societies. As with a commercial publisher, the exact nature of editorial independence is hard to define – for example, editors cannot expect to be allowed to publish defamatory material – but being published by a ‘non-commercial’ organization seems to be no guarantee that this freedom will be respected, and the risk of interference may actually be greater with society publications.

Competing interests of institutions

Journal editors are often the first to become aware of suspected misconduct. However, as mentioned above, journals generally should not attempt to investigate such suspicions since they are not equipped to do this and do not have the legal standing necessary either to access evidence or to discipline researchers. Therefore, COPE recommends that editors should refer cases to researchers' institutions and request that they conduct an investigation.

However, editors sometimes find that institutions are unco-operative or unresponsive to their requests. Just as the interests of societies and editors may diverge, institutions may be more concerned about preserving their good name than ensuring that fraudulent publications are retracted and miscreants disciplined. An analysis of cases brought to COPE between 2007 and 2011 in which editors requested an institutional investigation found that the institution's response was unsatisfactory in 12 of the 24 cases.14 These numbers are probably an overestimate of the true frequency of problems, since editors probably bring only their most difficult cases to COPE, so the committee tends not to hear about cases where institutions responded appropriately. Indeed, there have recently been examples of institutions thoroughly investigating misconduct and acting in an exemplary manner by publishing their findings and ensuring that the research record was promptly corrected15,16. Equally, there have been cases of editors failing to retract fraudulent papers despite being informed of the outcome of well-conducted and conclusive investigations17, so the problems do not lie solely with the institutions. However, based on our experience of the difficulties editors sometimes face, COPE has recently produced guidelines on co-operation between journals and institutions on cases of suspected misconduct18.

Institutional policies can have an important influence on researchers' behaviour. While good policies and a healthy research environment probably promote research integrity, poor policies, especially those that create pressure on researchers to publish, may actually encourage misconduct. If research productivity is measured by the number of articles published, this may provide incentives for ‘salami-sliced’ publications (i.e. generating as many publications as possible from a single data set) and gift authorship (e.g. when friends or colleagues who have contributed little or nothing to the research get listed as authors largely to enhance their CVs).

“A survey of instructions to authors from 234 journals found that 41% provided no guidance on authorship …”

Journal editors, distant from the research, generally have no way of distinguishing true from gift authors (or to detect when deserving authors have been omitted). Journals therefore rely on institutions to enforce sound authorship policies and to resolve disputes if they arise19.

While determining authorship usually rests with the local institution, educating researchers on ethical issues can be a joint responsibility with journals. COPE recommends that journals should ‘publish guidance to authors on everything that is expected of them’20. Unfortunately, not all journals do so. A survey of instructions to authors from 234 journals found that 41% provided no guidance on authorship and, of those that referred to the criteria of the International Committee of Medical Journal Editors (ICMJE), 35% cited an outdated version21.

Institutions can also play an important role in educating researchers about topics such as plagiarism and can also screen academic work to ensure it is not plagiarized. So, while the ultimate responsibility for ethical behaviour lies with authors themselves, journals, publishers and institutions can also contribute to preventing some ethical problems.

The academic research environment

Another important contribution that institutions can make to publication ethics is their influence on the research environment. Publications not only serve to communicate research findings but are used to measure the research productivity of individuals and departments. Being an author on a publication in a respected journal may be the key to getting a job, a promotion, or a grant. In some countries, researchers get direct financial rewards for publishing in international journals22. While pressure to publish cannot entirely explain misconduct, it may contribute to it. Even though most researchers would never fabricate results, the temptation to commit lesser offences, such as adding colleagues' names to papers or producing ‘salami’ publications, become harder to resist under pressure. Therefore, institutions and funders should consider the impacts their policies may have and work to reduce unintended consequences.

“While pressure to publish cannot entirely explain misconduct, it may contribute to it.”

Editors behaving badly

While many editors are conscientious and strive to ensure their journals publish high-quality material and maintain ethical standards, occasionally editors abuse their position or fail to live up to the highest standards.

As COPE mainly receives cases from editors, its database probably under-represents such problems although it does sometimes provide advice to publishers about suspected editorial misconduct. In one such case, a publisher raised concerns about an editor who appeared to be abusing his position by allowing articles from his friends and relatives (including his wife) to be published without independent peer review. In another case, reported by several newspapers and journals, an editor published positive papers about a medical device without disclosing that he had received over $19 million in royalties from the manufacturer23,24. The publisher is reported to have explained that these papers were rigorously peer reviewed, but that misses the point that such a clear conflict of interest should have been disclosed, so perhaps the publisher was also at fault for not having a more stringent policy on this.

Sins of omission

Publication ethics is not only about wilful misbehaviour. Journal policies, be they established by editors, societies or publishers, can influence behaviours and therefore contribute (in either a negative or positive way) to the overall ethical ‘climate’. Marusicć et al showed that the design of forms used to elicit information about the contributions of authors influenced the truthfulness of the responses25. Journal policies may also help prevent publication bias (i.e. under-publication of negative findings and repetitive publication of positive findings). For example, public registration of clinical trial details at the start of studies and use of trial registration numbers in publications can highlight non-publication of negative findings and selective or repetitive publication of positive findings. Members of the ICMJE had a major influence on the number of trials that were registered when they adopted a policy of mandatory registration in 200426. However, recent surveys have shown that only around 20–30% of journals require registration27,28. It could be argued that journals are acting unethically by failing to take this opportunity to raise publication standards.

Screeners or trusters?

While computer software has made some types of misconduct, such as copy–paste plagiarism and image manipulation, much easier to commit, technology also provides tools to help journals detect such problems. Text-matching software, such as CrossCheck29, can be used to screen for plagiarism or redundant publication. Similarly, the same programs that can be used to ‘doctor’ digital images can also be used to detect the alterations30. Publishers have to decide how much time and money to invest in such systems and editors have to decide when to apply them. Journals that have implemented routine screening (i.e. of all articles, not just those in which misconduct is suspected) have often discovered a higher incidence of problems than they previously imagined.

Is it unethical for a publisher not to provide all possible resources for detecting misconduct to its editors (or editorial staff)? And, if a publisher provides such tools, is it unethical for editors not to use them? Another interesting question is whether journals have any ethical duties concerning manuscripts they intend to reject. For example, if a journal screens all submissions for plagiarism, what should it do if it detects signs of misconduct in a manuscript it plans to reject? The COPE Code of Conduct indicates that editors ‘should not simply reject papers that raise concerns about possible misconduct. They are ethically obliged to pursue alleged cases’31. Yet, editors may argue that they barely have enough time to deal with ethical issues concerning the papers they have published or plan to publish, let alone time to fix problems in other submissions that will end up with other journals.

“… if a journal screens all submissions for plagiarism, what should it do if it detects signs of misconduct in a manuscript it plans to reject?”

Conclusions

Most articles about publication ethics focus on misconduct by authors and peer reviewers (i.e. people who are not employed by journals or publishers). While such problems should not be overlooked, I hope I have demonstrated that all players have ethical responsibilities. While it is, of course, important to seek to prevent and detect author misconduct (such as plagiarism and data falsification), the ethical issues relating to publishers, academic societies, research institutions and journal editors cannot be ignored. Closer co-operation, for example between journals and institutions, and between editors and their publishers, could reap considerable benefits. On the other hand, complacency and attitudes that publication ethics is ‘somebody else's problem’ will mean that little progress will be made.

Competing interests

The author was Chair of COPE from 2009–2012, and developed several of the COPE guidelines. This was an unpaid position.