had turned to digital methods and the study of virtual worlds (Snodgrass 2015). Anthropologists had also contested the notion of “the field” as a bounded, unitary, or even physical place (Bonilla and Rosa 2015; Marcus 1995) and argued for participatory and decolonizing modes of research (Harrison 2010). COVID-19 intensified all these trends. We do not yet know what the long-term effects of this moment will be for anthropological research, but I imagine we will look back on it as a rupture.
Whatever changes are to come, however, will be shaped by established traditions of research. My goal in this chapter is to review the basic elements of research design and provide a framework for matching methods to questions across different research traditions. Medical anthropologists come from a wonderful array of paradigms – positivist, critical, constructivist, interpretive, evolutionary, ecological, and more. It’s true that certain methods are associated with certain traditions, but no one tradition can lay claim to any particular method (Pelto and Pelto 1996, p. 294). As Bernard (2018, p. 2) puts it, “Methods belong to all of us.” The COVID-19 crisis accentuated this point, as fluency in a broad range of methods proved essential for adapting to novel circumstances, developing successful collaborations, and designing research that matters.
RESEARCH DESIGN
Research design is about posing good questions and finding empirical answers. The hallmark of well-designed research is that it justifies the claim that your particular answer is better than the alternatives. The goal is not to claim perfect knowledge – that goal is unattainable – but rather to generate systematic evidence that minimizes the errors of everyday reasoning and casual observation. Good research design thus requires researchers to be explicit about the methods and logic we use to connect theory and data, so that others can evaluate the validity of our claims.
Whole books have been written about research design, and several extended treatments discuss applications to anthropology in particular (Bernard 2018; Brim and Spain 1974; Johnson 1998; LeCompte and Schensul 2010; Pelto and Pelto 1996). This work is essential reading for medical anthropologists. Here I outline some basic ideas for connecting data and theory through research design.
Qualitative, Quantitative
Medical anthropology, like the social sciences generally, is often described in terms of a dichotomy between “qualitative” and “quantitative” methods of social research. However, a growing number of methodologists across the social sciences advocate “taking the ‘Q’ out of research” (Onwuegbuzie and Leech 2005; Sobo 2009).
There are at least two reasons why the qualitative–quantitative distinction is usually counterproductive. First, the collection and analysis of both qualitative and quantitative data are compatible with the same logic of inquiry (Keohane et al. 2021; Teddlie and Tashakkori 2009). From this perspective, researchers should use whichever methods work best for a particular research question. Second, the qualitative–quantitative distinction conflates data collection and data analysis. Bernard (1996) identified this problem by noting the ambiguity of the phrase “qualitative data analysis.” From the syntax alone, we cannot tell whether the phrase means the analysis of qualitative data or the qualitative analysis of data. We can avoid this ambiguity by using “qualitative” and “quantitative” to modify specific types of data and types of analysis – not types of research.
Figure 4.1 illustrates the point (Bernard 1996). The stereotypes of qualitative and quantitative research are depicted in cells A and D, respectively. Cell A captures interpretive approaches to text, including traditions such as grounded theory (Charmaz 2014) and discourse analysis (Farnell and Graham 2015). Cell D captures the statistical analysis of numerical data, such as from closed-ended survey research. But these combinations don’t exhaust the possibilities. In cell B, the qualitative analysis of quantitative data refers to the act of extracting meaning from the results of statistical analysis or mathematical processing. All so-called quantitative research involves this interpretive act; without it, there would be little point in running a regression model or producing a scatterplot. Last, methods in cell C generally involve turning words – or images or photos or artifacts – into numbers to look for patterns. That’s the essence of classic content analysis (Krippendorff 2018). We can also place methods for cultural domain analysis like free listing and pile sorting in this cell (Dengah et al. 2021; Weller and Romney 1988).
Figure 4.1 Qualitative and quantitative data and analysis (adapted from Bernard 1996).
The point is that medical anthropologists, like all social scientists, have access to many tools for data collection and analysis, and we ought to use the right ones for a given research question. Dividing the toolkit of social science into qualitative and quantitative methods tends to obscure that point.
Exploratory–Confirmatory Questions
A better starting point for thinking about research design is to recognize a continuum of research objectives, ranging from exploratory to confirmatory research questions. Exploratory questions seek to understand how and why things work as they do; confirmatory questions seek to test hypotheses based on new or existing theory. These different types of questions imply different types of methods along a parallel continuum of relatively unstructured to structured methods of data collection and analysis (Figure 4.2). This framework is useful because it helps to ensure that decisions about research design flow from the research questions.
Figure 4.2 A continuum of research questions and methods of data collection and analysis.
Exploratory research questions are common in medical anthropology. For example, Chavez et al. (1995) studied beliefs about breast and cervical cancer in Orange County, California. They asked: “‘Do Latinas, Anglo women, and physicians have cultural models of breast and cervical cancer risk factors? If so, how similar or different are their models?’ Another way of asking this question is, ‘Do they agree on the relative importance of risk factors?’” (p. 42). Here researchers began with limited expectations about what they would find and sought to detect patterns that would help to generate theory. This approach is appropriate whenever there is insufficient existing theory or evidence to establish expectations.
Exploratory questions are also apt for centering people’s expertise about their own lives, which can challenge dominant narratives, existing theory, or researchers’ preconceptions. For example, Reese (2019) begins her ethnography of racialized food apartheid and Black self-reliance in Washington, DC, by recounting a “conversation on Mr. Johnson’s front porch.” Reese chose this starting point because of the way it and other conversations “changed what I was listening for” (p. 2). Her initial concern was the influence of the built environment, a theoretical orientation “heavily influenced by anthropology, food studies, and sociology.” But Mr. Johnson had other stories to tell, and by “listening to him more than doing much talking” (p. 1), Reese left with a new set of questions that framed the rest of the work.
The flow of his storytelling revealed what Zora Neale Hurston wrote about in Dust Tracks on the Road: that research was the blessing through which I could formalize the curiosities that emerged on Mr. Johnson’s porch, and that if I got out of the way, Black people would tell their stories how and when they wanted. It was not my job to dictate which stories should be told, but if I let them, Black storytelling would lead me places that I had not planned to go. (Reese 2019, pp. 2–3)
Reese exemplifies the power of listening to generate
