The Great Saturation Debate: What the “S Word” Means and Doesn’t Mean in Qualitative Research Reporting

Claiming that one has “achieved saturation” (whether theoretical, data or some other form) has become something of an unquestioned orthodoxy within research reporting circles (Saunders et al., 2017). The term appears in standardized qualitative research reporting guidelines (cf., Tong, Sainsbury, & Craig, 2007), the review guidelines for many nursing journals, and is ubiquitous within qualitative methods texts for the point at which one can legitimately stop data collection (Hennick, Kaiser, & Marconi, 2017). The term “theoretical saturation” does have an important meaning within the conventional grounded theory tradition. It is congruent with the underpinning epistemological position of that method, which is “to develop an explanatory theory of the social processes that are studied in the environments in which they have taken place” (O’Reilly & Parker, 2013, p. 192). It may also be quite relevant for the development of robust forms of theorizing within the social sciences on the basis of other methodological approaches (Saunders et al., 2017). However, its relevance is much more questionable in the context of the kinds of studies that most of us are doing within the applied and practice disciplines such as nursing (Malterud, Siersma, & Guassora, 2016; Thorne, 2016). Our business in nursing and health is rarely building a theory that is so coherent that it surpasses all other theories in explaining the fundamental nature of social processes. Rather, we often point our qualitative inquiry lenses toward expanding on existing understandings of clinical phenomena that matter, adding new richness, texture, and dimension to what is already known and understood about aspects that intrigue us. In many studies, the most important outcome is new questions about the phenomena we might not have thought to ask without such in-depth inductive investigation. But rarely, I would argue, are we intending to claim a finding that deserves to be attached to the idea of saturation. So in the context of our nursing applied qualitative studies, when we invoke the sacred language of “saturation,” what is it that we really mean? Generally, the term is used to suggest confidence that we have obtained sufficient richness of information to conclude that nothing new or surprising will arise from further data gathering and that any additional data gathering would be redundant (Hennick et al., 2017). As Morse (2015) explains, saturation is set out as an expectation to guard against acceptance of published qualitative reports based on thin data and/or incoherent conceptualization—criteria with which I happen to fully agree. Sometimes, researchers insert qualifying terms however to imply a different kind of saturation, such as code saturation to indicate that they have “heard it all” or meaning saturation to suggest that they “understand it all” Hennick et al., 2017). However, according to Saunders et al. (2017), the use of the saturation concept implies an “uncertain logic” in the form of “a predictive statement about the unobserved based on the observed” (p. 1903). It is frequently referenced as a matter of declaration rather than in the context of supporting evidence (Green & Thorogood, 2009; Hennick et al., 2017; Morse, 2015; O’Brien, Harris, Beckman, Reed, & Cook, 2014). Far too often, we see qualitative nursing research reports that provide no explanation to substantiate the enthusiastic assertion that a study has reached that pinnacle of saturation at which the authors claim to have arrived—often after precisely the same number of interviews or observations they had planned for in their study designs. Thus what began as a methodological tool for the explicit purpose of formal social science theorizing has deteriorated into the ubiquitous defense for deciding to conclude data collection across the spectrum of qualitative studies (Thorne & Darbyshire, 2005).