How would you like your salami? A guide to slicing

A variety of conversations of late have led me to believe that medical education researchers are thinking about salami more now than ever before. More specifically, it has become clear that judgements of ‘salami slicing’ in various journals have cut fingers in ways that leave many well-intentioned authors uncertain about how to safeguard themselves and many editors frustrated with the need to play the role of ‘food inspector’.1 I'll now try to finish this editorial without over-processing the metaphor. When I say ‘salami slicing’, I mean the act of dividing a single story (usually derived from a single study) into multiple papers. This definition prioritises the division of a ‘story’ rather than defining salami slicing through the division of a ‘research project’ because what constitutes a project is so variable in our diverse field that one set of data can tell multiple stories and an interpretable story may require multiple studies. Think of this in terms of other genres. A story might require no more than a short film for it to be told in a compelling and complete fashion. If, however, core scenes that provide plot-crucial information are left out or are split across multiple movies, the story will not be coherent and the audience will be frustrated. Salami slicing is considered unethical in the publishing world because it wastes resources (reviewers’ time, journal space, etc.), creates a disjointed literature (forcing readers to go on scavenger hunts to find essential information), and dilutes an important currency of academia (inflating the curriculum vitaes of some and risking the possibility that rewards will flow to those who ‘play the game’ rather than to those who contribute most effectively).2 There are many explanations for salami slicing, some of which are more charitable than others, but I will not go into those because this editorial is not another admonition against unsavoury practice. Rather, this editorial is an attempt to clarify what is meant by ‘salami slicing’ and to offer concrete guidance regarding ways in which it commonly presents in an effort to help authors judge when they might be slicing too thinly. As with anything that requires judgement, a number of confusable concepts arise when considering salami slicing, as does a certain amount of ‘I know it when I see it’. There are a variety of features (spread across every section of a manuscript), however, that commonly raise suspicions. The introduction: if two papers require the same build-up, summarise the same literature and lead to very similar research questions, then it can be difficult to determine why a second paper is required. Sometimes such overlap reflects efforts to conduct programmatic research or replication. Both of these activities are to be encouraged, but such labels can rarely be maintained when both the data and the story arise from a single research study. In any case, avoiding charges of salami slicing requires clarity with regard to how the current work builds substantively upon what has come before it. Having ‘a lot of data’ to present is insufficient grounds for splitting studies into multiple papers because it may not be necessary to share everything to tell the story compellingly. Also, although many journals in our field have removed their strict word limits to enable stories to be told properly keep in mind that more can be less given that key messages can be lost in a sea of text.3 Methods: the statement that ‘these data are derived from a larger research project’ is guaranteed to generate questions about salami slicing that may yield positive or negative answers. I certainly would advocate for authors to transparently reflect the context within which their data were collected, but the inclusion of such a statement demands an effort to clearly explain how, when and why different stories emanated from the same project. Further, there should be enough detail in the methods section of the second paper to allow the paper to be read and evaluated in isolation without requiring the reader to look elsewhere to consider key methodological decisions. Results: the most egregious forms of salami slicing I have come across have been apparent in results sections, whereby some authors present one analysis (or a subset of the data collected) in one paper and a different or expanded analysis aimed at exploring the same research question in a second. This is especially troubling when it is clear that analyses of relevance to the research question are simply withheld so they can be ‘revealed’ in a second publication, as was the case in a paper sent to us outlining a regression analysis performed on data that had been published previously with the conclusion that ‘a regression analysis would be beneficial’. Discussion: if interpretation of the data presented in two papers requires knowledge of the data presented in the other paper, then the data should probably be presented together rather than simply cross-referenced. I appreciate that it requires hard work and creativity to write concisely when there are many things that could be said, but such effort generally results in greater influence. In fact, ‘the whole is greater than the sum of the parts’ is likely to be the side on which one should err as a single strong paper will usually be more impactful than a series of weaker ones. I trust that it goes without saying that this list of red flags is not meant to be a guideline for how to cheat the system. Nor are they offered to publicly shame anyone who might have engaged in salami slicing. Rather, they are relayed for two reasons: (i) as an effort to help authors understand the observations that have led editors to become far too aware of the practice of salami slicing and explain why editors must remain vigilant although only a small minority of authors engage in such behaviour deliberately, and (ii) because the people who are best positioned to consider whether or not a story might be sliced too thinly are the very co-authors, who sit in the well-intentioned majority. There is much to be gained from debating publication strategy with co-investigators as such decisions can be as important to the quality and impact of a work as any of the methodological choices we adopt. In this editorial I hope to have illustrated the many close cousins of salami slicing that are not only acceptable, but should be encouraged. If ‘diagnostic uncertainty’ exists with regard to the category into which one's efforts fit, the best advice anyone can offer is to be transparent. Alert editors to articles, published or otherwise, that might be construed as a slice of salami so that you make your intentions clear and enable a genuine discussion to take place. Authors, reviewers and editors need to work together, after all, to ensure that our plates remain palatable despite my extensive butchering of the salami metaphor.