Commentary—On the “Indistinguishability Thesis”: A Response to Morgan

This commentary is a response to an article by David L. Morgan titled “Living Within Blurry Boundaries: The Value of Distinguishing Between Qualitative and Quantitative Research.” In it he argues against what he calls “the indistinguishability thesis,” the claim that “it is impossible to distinguish between qualitative and quantitative research.” He claims that a number of writers, including myself, have put forward this thesis. I deny this charge and try to provide a clearer outline of what is wrong with the quantitative–qualitative distinction and the sort of alternative conceptualization that is required.

In this issue of JMMR, David Morgan challenges what he refers to as the “indistinguishability thesis”: The claim that it is impossible to distinguish between qualitative and quantitative research (Morgan, 2018), in his words “that it is not actually possible to distinguish between what are traditionally known as QUAL and QUANT” (pp. 1-2). He also formulates the thesis as stating that “the differences between QUAL and QUANT research are overstated, causing more harm than good” (p. 1) and that “there are no meaningful distinctions between QUAL and QUANT research” (p. 2). He lists a number of proponents of this thesis, including me, though he focuses primarily on an article by Sandelowski (2014).

Morgan argues that, while the boundaries between quantitative and qualitative are fuzzy—there is no list of necessary and jointly sufficient conditions identifying them—each of these categories can be characterized in terms of a set of features that tend to go together. As a result, qualitative methods share family resemblances, despite their differences, and the same is true of quantitative methods. He rejects the argument that, because no set of essential criteria can be identified, there is no significant difference between quantitative and qualitative methods. In this, he is particularly concerned to challenge Sandelowski’s argument that the mixed methods literature tends to preserve a false dichotomy between qualitative and quantitative, to mistakenly require that good research should combine the two, and to neglect the significant choices that often have to be made within the qualitative or quantitative realms.

There is an immediate problem in assessing Morgan’s argument. This is that his formulation of the indistinguishability thesis is itself fuzzy: One can believe that the differences between quantitative and qualitative research are often overstated, and may do more harm than good, without assuming that “it is not actually possible to distinguish” between quantitative and qualitative methods. And one can believe that there are meaningful differences between what would normally be classed as quantitative and qualitative studies, while also insisting that a global distinction between the two types of research is misleading. However, the thrust of Morgan’s argument, as I have indicated, concerns whether the two sets of methods can be distinguished. Thus, he does not enter into discussions about whether the differences between them have been overstated, or whether this dichotomy has had, or could have, negative consequences—aside from arguing that qualitative and quantitative methods have different strengths in the context of particular studies, so that combining them can be fruitful, a claim that is not in dispute. ¹

This leads to a second problem, which is that neither Sandelowski nor I have claimed that quantitative and qualitative methods are “indistinguishable.” And I believe the same is true for the other “proponents” he lists. On my reading, Sandelowski’s discussion emphasizes differences between methods, and her point is that these do not line up into a single binary distinction. In her conclusion, Sandelowski (2014) writes, “In my effort to unmix mixed-methods research here, I am not positioning myself against mixed-methods research qua research, or proposing that theories and methodologies cannot or should not be distinguished from each other” (p. 6). Given this, the label “indistinguishability thesis” is a misnomer for her argument, as it is for mine.

My argument in the pieces that Morgan cites (Hammersley, 1992, 1996) was that the terms “qualitative” and “quantitative” have varied in what they have been taken to imply, that each covers a heterogeneous range of kinds of work, and that there is some overlap in the features of some of the research within the two categories. By contrast, “indistinguishability” would imply complete overlap, which I certainly did not claim. My main point in the work he cites, and some later discussions (Hammersley, 2013, 2018), was that any dichotomy at the level of collections of methods is of very little, if not of negative, value. I suggested, instead, that we need to distinguish among the various decisions that researchers must make in the course of a research project, and the range of options available in relation to each of these. Moreover, on the basis of the pattern of options selected, we can identify different styles of research, such as social survey, experiment, ethnography, and so on (see Hammersley, 2018, Table 2). This way of dealing with the issue identifies similarities even between the styles of research that are typically labeled “quantitative” and “qualitative,” while yet also noting their differences and, at the same time, mapping variation within the qualitative and quantitative domains. My view is that we would be better off not using the global quantitative–qualitative distinction, but relying instead on more specific style and option distinctions. And, on my reading, Sandelowski (2014) adopts a similar position.

Morgan also attributes “essentialism” to proponents of the indistinguishability thesis. As I noted, he recommends instead of this that quantitative and qualitative research can each be identified in terms of a set of features that typically occur together, but do not always do so, appealing here to Wittgenstein’s notion of family-resemblance and Rosch’s empirical studies of categorization. He apparently thinks that anyone who argues that there is diversity within the quantitative and qualitative categories, and overlap between them, must be committed to essentialism. But, in fact, there would be nothing to stop someone arguing that qualitative and quantitative methods are indistinguishable from one another because their components lack sufficient family resemblance. While this claim may be implausible, and has not been put forward by the writers he criticizes, it highlights an important point: that it is incumbent on anyone who makes a distinction on the basis of family resemblance to show that the resemblance reaches some necessary threshold. However, Morgan does not address this issue. And I suggest that it is more difficult to identify what components research methods share, and do not share, than it is in the paradigm case of the concept of “game” (Wittgenstein, 1953).

If methodological discussions are to be fruitful, it is essential that opposing positions are characterized accurately. Falsely attributing an implausible thesis to opponents subverts discussion in unhelpful ways. I spent much of my book on What Is Qualitative Research? (Hammersley, 2013) outlining the distinctive features of this type of research, while explicitly recognizing that no set of necessary and jointly sufficient conditions can be specified for identifying it. And my argument has long been that the quantitative–qualitative distinction is too crude and potentially misleading, not that the two kinds of research are indistinguishable. I also believe, and Sandelowski seems to take the same view, that rather than the mixed methods movement providing a better typology of research strategies it has effectively institutionalized the old distinction, and has obscured the value of combining one qualitative method with another, or one quantitative method with another. And I share her concern about the simplistic contrasts that are typically drawn between quantitative and qualitative methods in the mixed methods literature, though these are also to be found elsewhere of course. None of the people that Morgan criticizes, as far as I know, denies the value of combining methods—indeed some of us believe that this is usually crucial, since triangulation provides a basis for checking validity as well as offering complementary data (Hammersley, 2008). But its value lies in comparing what is produced in relation to specific aspects of the research process (type of data, kind of analysis, etc.) rather than in the global terms of quantitative versus qualitative.

Perhaps the most disturbing feature of Morgan’s article is that he seems to be primarily concerned with defending the idea of mixed methods research against criticism. He appears to assume that the quantitative–qualitative distinction is essential to any commitment to combining methods, and therefore must be preserved. I suggest that this is to put the cart before the horse in precisely the ways that both qualitative and quantitative researchers sometimes did during the “paradigm wars” (Gage, 1989). While I do not share Bryman’s (2008) optimism that these wars are over, I do share his desire that such military maneuvers should cease.