Settling on a design for data collection and analysis is daunting for any qualitative researcher, especially those new to the field. There are books and other resources available to learn about qualitative research, but they often include jargon or assume background knowledge that can make it challenging to parse the pros and cons of different approaches. This post outlines some considerations that may be useful to those new to qualitative research who are working on research design. I focus primarily on examples and considerations for academic medicine, but these insights may be useful to researchers in other settings as well.
As a kid, some of my favorite books were from the Choose Your Own Adventure genre. These books included key decision-points where I could choose what the main character would do. Each choice led to a new version of the story. At any point, I could also backtrack and explore what it would've been like to have made a different plot decision. In addition to being a fun way to read, this format represents a very good platform for learning. So here is a choose-your-own-adventure for qualitative research design.
The crucial decision point to navigate is What is the research objective? Are you looking to describe respondents’ lived experiences of a setting, phenomenon, or condition? OR Are you interested in identifying or characterizing factors of a setting, phenomenon or context, e.g. barriers, facilitators, challenges, opportunities, qualities? OR Do you want to characterize or explain processes of a setting such as how something works or why something does or does not happen?
To be clear, these types of research objectives can overlap conceptually; they can even overlap within a given publication. Scholars may, for example, examine lived experiences in order to understand how a given process unfolds. But I discuss the objectives separately here for clarity. Identifying the main research objective helps the researcher understand what path they will need to traverse regarding data collection and analysis.
(a) Describing lived experience
In some cases, deeply understanding how a given condition, process, or setting feels to respondents is the main contribution researchers wish to make. This objective is most closely aligned with epistemological and analytic approaches such as grounded theory and phenomenology, but other traditions may be used as well. Examples of inquiries into lived experience include: an exploration of patient experiences of temporomandibular disorders; of nonmedical sources of suffering among chronically ill individuals; of “heavy users” of emergency departments; of “moral distress” among nurses in the UK; or of the first clinical placements of Irish nurses. To develop deep, empathic, vivid descriptions, researchers typically conduct substantial data collection using methods such as participant observation and longitudinal in-depth interviews. To accommodate the intensity of data collection, sample sizes are often, but not always, small. Analytic approaches are often inductive and exploratory, focusing on an experience or setting about which researchers know little. Though this type of research is deeply important, I see it less frequently in academic medicine than the other two objectives I describe here.
(b) Identifying and characterizing factors
An important contribution of qualitative research in academic medicine is documenting, naming, and describing factors that characterize a phenomenon or setting of interest. This describes much of the applied and translational research on, for example, barriers and facilitators to the implementation of interventions; challenges patients face when trying to access care; stakeholder perspectives on an issue; and forward-looking, sometimes human centered design research such as qualities patients or providers desire in novel models of care. Because this path (b) seems to be taken more often in academic medicine than the others (a, c), it’s worth unpacking to a greater degree.
Using the schema outlined in Rendle et al 2019, the design of research aiming to identify or characterize factors may be exploratory if very little is known about the setting or phenomenon; confirmatory if the state of the science is more advanced; or comparative if the researchers are looking to identify similarities and differences across settings or populations. The work of identifying and characterizing factors can be done with a variety of analytic approaches, selected based on the goals of the analysis. Combinations and variations of what this could look like are infinite, but I describe a couple common examples below.
An exploratory or confirmatory project might have an initial goal of identifying what kinds of concerns arise when participants consider a medical or self-care decision. For example, some of my past work in this style has examined what factors patients consider when deliberating about joining a clinical trial of a new cancer treatment. This might be detected via observations of patient-provider discussions, which are recorded in field notes by a fieldworker or transcribed from a recording of the interaction. Alternately, factors might be elicited in qualitative interviews where the interviewer asks about treatment decision-making; about barriers or facilitators to self-care, cancer screening, or pain management; or about unmet needs of people living with dementia and their caregivers. It could be from focus groups where individuals brainstorm or collectively identify challenges. Or the data could be collected in open-ended survey questions, in which patients describe e.g., barriers to seeking healthcare in the context of COVID-19. Regarding a comparative study design, an example could be a qualitative investigation of drivers of clinician moral distress in the United States vs. the United Kingdom. Each of these approaches have strengths and weaknesses that the researcher must weigh with regard to their objectives, resources, and abilities.
Once data are in hand, many academic medicine researchers employ thematic analysis to iteratively identify, categorize, and/or characterize the factors under study. (The term "thematic analysis” has been applied to a range of analytical approaches and techniques. Braun and Clarke's 2019 review is an invaluable and clarifying overview.) Continuing with an example from above, a researcher may want to identify concerns that candidates have about a medical procedure. They may wish to describe what those concerns look like or, in comparative work, the goal may be to determine what factors exist or are more relevant in some groups/settings vs. others. Accordingly, the analytic designs are often a hybrid of deductive and inductive analysis, or even primarily deductive in the comparison example. A primarily inductive approach may be appropriate if the study’s focus is a truly novel phenomenon or setting. Examples of the end results of thematic analysis include the identification of relevant topics or domains for further study; a presentation of themes that reveal core concepts or stories from the data; or the development of a conceptual model based on the thematic insights, which begins to get at how things work in the focal setting.
(c) Finally, researchers may want to characterize or explain a process.
In contrast to (a) deeply describing something or (b) identifying and characterizing factors, research on processes is focused on how something happens or works. This might be a standalone research objective or it might be paired with, for example, the earlier identification and characterization of factors in one or more sites/populations. To convincingly describe how a process works, researchers benefit from having data from multiple viewpoints and/or positions in a setting. This might come from ethnographic observations in e.g., a clinic, where the field worker would learn about both patient-facing and behind-the-scenes processes; longitudinal in-depth interviews with individuals as they navigate the process under study; and/or analysis of documents that could reveal resources and/or policies that shape the process. Understanding and describing dynamic processes typically takes more data and/or more intensive analysis of data than analysis focused on identification of factors. This type of work is common in fields such as sociology and anthropology, but it has an important role in academic medicine as well. Analytic methods may include more intensive and iterative applications of thematic analysis or other approaches with robust inductive and deductive resources e.g., matrix analysis, framework analysis; or those designed to specifically generate more complex and abstract insights, such as the constant comparative method from grounded theory. Examples of this type of work include studies investigating how late stage cancer patients move through the early phase clinical trial pipeline; how leaders of a surgical training program navigate challenges to implementing duty hour reforms, how a safety-net emergency department manages the day-to-day provision of indigent care; how individuals make meaning of “moderate risk” genetic results; and how policies at single-room occupancy (SRO) residences dynamically shape the behaviors and mental health of unstably housed women.
There is a wealth of work that qualitative researchers can do in academic medicine to can help us to understand patient experience, inform intervention development and implementation, and illuminate complex and consequential processes in healthcare. All of these require different types of data and different analytic methods. For novice researchers, being aware of the different paths they can travel, what they require, and where they lead may be helpful starting point. We will describe these in greater depth in future blog posts.
h they were a patient, or if you have provided professional