Publication Overlap: Building an Academic House with Salami Shingles

Salami slicing, shingling, and meat extending are all fanciful terms for a thorny and often overlooked issue. Publication overlap—the presentation of redundant ideas or data in multiple papers by the same authors—is a practice that warrants serious discussion. When there is little to no publication overlap, the findings in one paper lead to new ideas and the collection of new data, producing a paper trail of scientific progress that moves the field forward (Fig. 1). At the other extreme, authors present the same text and data in two papers, which is usually a clear violation of journal copyright policies (“dual publication,” Merrill 2015). In between these two extremes are different degrees and forms of overlap that may result in multiple low-impact papers. For example, authors may ask the same question with different datasets, or they may ask different questions with the same dataset. In these cases, authors have to critically evaluate whether publishing multiple separate but overlapping papers is in the best interest of the science community or whether it is self-serving. We aim to spur conversation about the consequences of publication overlap in the ecology and evolutionary biology (EEB) community, where discussions of this issue are lacking despite the pressure to “publish or perish.” We offer generic examples of three forms of publication overlap that may commonly occur in EEB and other fields. Authors offer several justifications for producing overlapping publications. These include having a large project with multiple endpoints, quickly disseminating initial results without waiting for additional data to be collected, publishing in multiple languages, reaching different audiences in different journals, publishing in high profile journals with strict page limits, proving to funding sources that progress is being made on a project, and addressing significantly different questions that would not result in a single cohesive paper (Buddemeir 1981, Refinetti 1990, Susser and Yankauer 1993, Schein and Paladugu 2001, Tobin 2002, Statzner and Resh 2010). Scientists may also be unaware of the concept of publication overlap and may receive little to no training on this issue (Lazaridou et al. 2017). However, we suspect that publication overlap is often the deliberate result of scientists wanting to inflate their research productivity and increase their publication number. In a publish-or-perish world, researchers may be willing to sacrifice research quality and impact for publication count. Combining overlapping publications into a single paper has negative consequences for readers, journals, and authors, and the ethics of publication overlap have been called into question (Alfonso et al. 2005, Benos et al. 2005). With an explosion in publication number (Grossman 2014), readers must spend their time sorting through overlapping publications, identifying novel results, and comparing papers by the same authors that answer similar questions (Berk 1996). Publication overlap can also lead to practical problems in meta-analyses, and excluding duplicate results and identifying non-overlapping data is an essential step to prevent bias in meta-analyses (Tramer et al. 1997). Furthermore, when a large dataset is split apart to answer similar questions, the separate but overlapping publications fail to offer a cohesive picture. These issues are compounded when authors fail to cite their previous overlapping publications, an unethical and stagnating practice called covert duplication (Jefferson 1998, Smart 2017) that occurs surprisingly frequently in some disciplines (Schein and Paladugu 2001). Overlapping publications also contribute to a backlog of papers needing to be reviewed and requires the time of editors and journal resources (Statzner and Resh 2010, Grossman 2014). Finally, authors waste time and resources finding new ways to present similar questions or data in multiple publications. Severe forms of publication overlap risk retraction, which can be damaging to the reputations of all authors involved. In a published survey of science editors, publication overlap was the top issue of concern across disciplines (Wager et al. 2009). However, during our initial research on publication overlap, we were intrigued by how frequently the issue is discussed in the medical field relative to other disciplines. We searched the Web of Science Core Collection using a query that encompassed many synonyms and forms of publication overlap (7 March 2018; see Appendix S1). This query returned 912 articles. We excluded correspondences about retracted articles, publications not discussing publication overlap, and duplicate publications (e.g., the same editorial printed in multiple journals). Of the 459 remaining articles, 346 (75%) were in the medical field (Fig. 2). Seven articles (1.5%), published between 1981 and 2015, were specific to EEB (see Appendix S1). It is possible that the medical field dominates the discussion on publication overlap because this issue is a bigger problem in medicine than in EEB and other fields. To investigate this possibility, we searched the Retraction Watch Retraction Database (Retraction Watch 2018) on 9 March 2018 for retracted peer-reviewed publications in medicine and EEB. We defined EEB publications as either those tagged with “ecology” or those having both the tags “env” (i.e., environment) and either “zoology” or “plant biology/botany.” We tabulated the number of articles retracted because of duplication of the article, data, images, or text, “euphemisms for duplication,” or salami slicing. The medical field had about 20 times more retracted articles than EEB. However, EEB and medicine had similar proportions of articles retracted due to publication overlap (Fig. 3). Given that many cases of publication overlap may be unethical but not retraction-worthy, these numbers most likely highly underestimate the prevalence of potentially problematic overlap in both fields. For example, in a systematic screening of medical publications, 14% were suspected of some form of overlap, especially salami slicing (Schein and Paladugu 2001). Our results suggest that publication overlap is a problem in both fields, and that the EEB community, though smaller than the medical field, should be vigilant of its occurrence and negative effects. Authors should critically consider whether publishing a single comprehensive paper or multiple overlapping papers would be more appropriate to present the information concisely and cohesively. If multiple papers are needed, authors should take care to self-cite to avoid covert duplication. Though the onus for preventing publication overlap is primarily on authors, journals can help by defining publication overlap and making their policies explicit. To investigate current EEB journal policies, we selected the top 20 journals listed by Scimago Journal Rankings in the “Ecology” category and the “Ecology, Evolution, Behavior and Systematics” category (see Appendix S1). We excluded journals that only publish reviews, leaving 32 journals that were listed in one or both categories. We surveyed each journal's instructions to authors, editorial processes, and permissions web-pages to determine what policies on publication overlap were available to authors before submitting manuscripts. Of the journals surveyed, 31% (n = 10) asked authors to include descriptions of any overlapping publications in the cover letter. Interestingly, 12.5% (n = 4) required authors to submit several similar publications. Though the rationale for identifying similar publications is to ground the work in a body of literature and aid in the selection of reviewers, it could also help editors identify possible cases of overlap. Only 9% (n = 3) of the journals had statements explicitly discouraging salami slicing. Some resources for investigating possible cases of publication overlap exist. For example, Alfonso et al. (2005) provide a list of criteria to help editors identify publication overlap. Similarly, many journals belong to the Committee on Publication Ethics (COPE), which provides guidelines, action flowcharts, and resources for journal editors to use when investigating possible ethical issues, including publication overlap. Seventy-eight percent (n = 25) of the journals we surveyed were members of COPE. However, of the members, only 60% (n = 15) referred to COPE's guidelines in their instructions for authors, making the resources provided by COPE that define publication overlap unlikely to be accessed by authors before submission. We suggest that discussions of the definition and consequences of publication overlap be incorporated into university ethics training, graduate student journal clubs, academic reading groups, online forums, and laboratory meetings. We hope that these conversations will help scientists recognize and evaluate problematic cases of publication overlap that are a disservice to science. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.