a choice to respect people’s experience. And so she entered the field with a fundamentally exploratory question: What is going on here? Her decision to listen allowed people to define what matters on their own terms; her job, then, was to “formalize the curiosities that emerged.” Reese did so primarily using methods that prioritize discovery: participant observation, archival research, and semistructured interviews.
When prior theory and evidence warrant specific expectations, confirmatory research questions are appropriate. One distinctive feature of confirmatory research in medical anthropology is that it often builds on an exploratory phase in the same study. We see that progression in Reese’s work. Toward the end of her fieldwork, Reese and a community collaborator conducted a survey designed to place the ethnographic findings in context. Part of the purpose was strategic: “The hope was that this data would put some numbers behind the anecdotal experiences that we all knew were true but were not always heard by those in power” (p. 15). But there is also an implicit confirmatory question: To what extent do the stories of Mr. Johnson and others characterize a broader geographic and social context?
Similarly, in their exploratory work, Chavez et al. (1995) found that Latinas’ beliefs about cervical and breast cancer differed from biomedical models more than Anglo women’s beliefs did. “We were left wondering,” they later wrote (Chavez et al. 2001, p. 1114), “to what extent these patterns of belief were associated with behavior, specifically the use of Pap exams, a screening test for cervical cancer. In other words, to what degree do cultural beliefs matter in the use of medical services?” Chavez et al. (2001) combined ethnographic interviews and survey research to address this question and found that, under certain circumstances, beliefs matter a lot.
Dressler’s (2005, 2020) research on culture, stress, and health also illustrates the progression from exploratory to confirmatory research questions. The central thread of Dressler’s work is to identify how culturally meaningful aspects of social status shape the distribution of stress-related health outcomes. This objective entails (a) a set of exploratory research questions about how social status is culturally constructed in specific times and places and (b) confirmatory research questions about the relationship between locally meaningful aspects of social status and health. This example illustrates the interdependence of exploratory and confirmatory questions in medical anthropology. Dressler drew on existing theory to anticipate an association between social status and health. This theory – and substantial empirical evidence – justified confirmatory, hypothesis-testing research, but Dressler first adapted the general theory to specific ethnographic contexts through exploratory research. Dressler refers to this strategy, which depends on working across the exploratory–confirmatory continuum, as “the ethnographic critique of theory” (Dressler 1995, p. 45).
Unstructured–Structured Methods
The continuum of exploratory to confirmatory questions is useful because it informs the choice of methods for data collection and analysis. One approach is to strive for a fit between exploratory–confirmatory questions and unstructured–structured methods (Figure 4.2). By “structure,” I mean the amount of control researchers impose on data collection. The difference between a structured and unstructured interview, for example, is the likelihood that all participants respond to the same questions in the same order. The basic principle for matching methods and questions is that the less we know about any given phenomenon, the less structure we ought to impose, so that we remain open to discovery. As we learn more and begin to develop hunches about what’s going on, we often want to impose more structure to test our ideas (Weller 2015).
Note that “structure” does not mean “qualitative” or “quantitative”; qualitative and quantitative data and analysis cut across the continuum. For example, informal interviews conducted during participant observation – Reese’s (2019) “conversation with Mr. Johnson” – generate qualitative data and fall at the unstructured end of the continuum. But we could also obtain qualitative data from more structured methods, such as an open-ended survey that poses the same questions to each respondent in the same way, as Reese (2019, pp. 52–55) also did. The choice between these methods depends on the balance between exploratory and confirmatory objectives. Informal interviews would remain open to discovery, whereas an open-ended survey would permit systematic comparisons between respondents.
We can also place both qualitative and quantitative methods of data analysis along the entire spectrum. Both grounded theory (Charmaz 2014) and semantic network analysis (Doerfel 1998) could be described as unstructured methods, in the sense that researchers try not to impose a prior theoretical framework about the concepts and relations in a given corpus of text. Both approaches are appropriate for exploratory aims. Yet grounded theory relies only on words, whereas semantic network analysis relies on turning words into numbers and on mathematical processing.
You may have noticed that the exploratory–confirmatory and unstructured–structured framework mirrors another important continuum in research design: that between inductive and deductive logic. Inductive reasoning starts with empirical observations – data – and works toward empirical generalizations to develop theory. Deductive reasoning begins with theory and works toward specifying expectations, or hypotheses, that can be checked against empirical observation. As Figure 4.3 suggests, these modes of reasoning are inextricably linked in the logic of social research, which seeks to generate (inductive) and verify (deductive) theory about how the world works. Regardless of their epistemological perspective, most researchers engage in both types of reasoning at one point or another. Decisions about which methods to use at any point in time should be informed by consideration of where researchers are in the research cycle.
Figure 4.3 Inductive and deductive modes of reasoning in the cycle of research.
Ethnography is commonly considered an inductive approach, but it is compatible with deductive modes of reasoning, too. Cascio (2017) describes a process of operationalizing biopolitical theory for ethnographic research on autism in Italy. Using a meaning-centered approach, Cascio conducted participant observation and ethnographic interviews with staff at autism-specific service organizations, autism professionals, and people who attended autism services, as well as their parents. Cascio analyzed transcripts and notes using both inductive and deductive techniques, including a set of a priori codes derived from biopolitical theory (e.g., molecularization, somatic expertise, collective identities and socialities). This deductive approach positioned Cascio to ask how well contemporary discourses on autism in Italy resonate with theories about biosociality and biological citizenship.
Elements of Research Design
Good research design consists of an explicit, logical plan for connecting data and theory. In all types of research – participant observation, surveys, experiments – the major components of this plan are the same:
Formulating research questions (and hypotheses, if appropriate)
Selecting a research site where the research questions can be addressed
Developing a sampling strategy for selecting observations required to answer the research questions or test hypotheses
Choosing methods to collect data needed to answer research questions
Creating a plan for managing, documenting, and archiving data
Selecting methods for analyzing data to answer specific research questions and test hypotheses
This list makes clear how research questions ideally permeate all major components of research design – sampling, data collection,
