Editorial: Multiple outputs from single studies: acceptable division of findings vs. ‘salami’ slicing


In this information age, transgressions in publishing ethics can readily occur and many people are concerned that these behaviours are on the rise. The term ‘salami slicing’ is considered to be a publication transgression, carrying connotations of inappropriate practice and referring to publishing an excessive number of papers from a single study. Salami slicing describes ‘artificially segmented articles in which related aspects of the same study were published separately’ (Bailey 2012, p. 212). As implied by this definition, the term suggests that each paper is so thin (akin to slices of salami) and that the whole purpose of multiple outputs is to bolster author CVs, perceived performance and scholarly standing rather than disseminate research findings with integrity. Indeed, the practice is said to be driven primarily by the ambition of authors, particularly from a ‘publish or perish’ culture, with pressure for staff to publish in academic journals for tenure, promotion and other career progression opportunities.

Notwithstanding the above issues, there seems to be confusion in some quarters about salami slicing and the legitimate division of research findings for dissemination purposes; sometimes, too, there is conflation between salami slicing and ‘duplicate publication’. This confusion could stem from salami slicing having received less attention over the years than more blatant forms of publishing misconduct, such as duplicate publication, publication of fraudulent research and plagiarism, and so, many authors may be less aware of the reasons why the practice is problematic. Despite a relative lack of attention, multiple articles – often editorials, letters and other content – discuss the negative implications and specific concerns about salami slicing, and these tend to focus on several key objections: the practice exaggerates research findings, potentially threatens and skews the evidence base for care (e.g. systematic reviews), takes up valuable journal space and makes further demands upon and wastes the resources of editorial teams, peer reviewers, readers and libraries (Walter & Bloch 2001, Norman & Griffiths 2008).

There are different ways to split papers or ‘slice the salami’. For example, authors may submit several papers to the same journal or to different journals, using the same research cohort yet neither disclosing this to the reader in the text of the manuscript nor advising the Editor in the cove-ring letter accompanying submission (Karlsson & Beaufils 2013). Wittingly or unwittingly, this practice can deceive editors.

There can clearly be tensions for researchers in adequately reporting findings and avoiding inappropriate splitting of research findings. There can be quite legitimate reasons for division of research findings. While there are ethical issues associated with salami slicing, there are also ethical issues with not appropriately disseminating research findings in their entirety. Most research participants involve themselves in research not to enhance the careers of researchers, but because they seek to contribute to knowledge through the sharing of their story, experiences and other aspects of their personal (often intimate) information. It is contingent on researchers to disseminate findings with integrity so as to facilitate optimal benefit from research findings, but also to do justice to research participants/subjects who have taken the time to help to generate the data.

An important practical aspect is that it is not always possible to convey all findings in one article within the word restrictions of a journal. Some qualitative and mixed-methods studies, for example, generate reams of valuable data that cannot be effectively reported in a single paper. Furthermore, different journals target different sets of readers and papers may be fashioned to meet the needs of a particular audience. In addition, some of the most exciting and cutting edge research is driven by doctoral students enrolled in doctoral programmes that promote or even require a series of publications.

In the context of duplicate publications, changes to authorship – ordering and names – have been framed as a deceitful strategy aimed at avoiding detection of author misconduct (Scanes 2009). However, there can be many legitimate reasons for differing authorship of papers arising from a single study. It is also necessary to acknowledge that membership of a research team does not denote automatic authorship (Cleary et al. 2013), and so, changes to author order or inclusion or exclusion, can simply denote a greater or lesser contribution or no contribution at all (Cleary et al. 2012). Furthermore, research team membership can change over the life of a project, with staff changes or where unanticipated findings necessitate the co-option of a specialist to work with data and provide crucial input into publications.

Writing in 1999, Walter identified five strategies to guide authors and editors in relation to possible salami slicing. These strategies are that authors should present results in as few papers as possible; ensure as little similarity as possible between papers; provide explanations as to why multiple papers are needed; disclose to editors information about previously published or future planned publications; and that editors should consider the potential contribution of the paper to the journal as well as the wider discipline area (Walter 1999).

A good rule of thumb is if in any doubt at all consult the Editor and be prepared to provide copies of previously published papers (including those ‘in press’) related to the study. In addition, it is wise to cross-reference previous papers highlighting the original contribution of each paper. Slicing papers is considered appropriate provided that each publication is cross-referenced and that each ‘slice’ addresses a distinct and important research question (Durani 2006). If the primary publication is cited/cross-referenced in these successive manuscripts, then this is transparent and the editor can determine whether subsequent manuscripts are worthy of publication (Karlsson & Beaufils 2013).