Generally speaking, grounded theory and its attendant methods is a way of structuring and making sense of open-ended data. These data might be generated from survey responses, interviews, focus groups, or in-field notes taken by a researcher. As the name suggest—and in contrast to other approaches—grounded theory's methods begin in the data themselves; in short, the data ground the analysis (and any subsequent insight, theory, or recommendations). Grounded theory practitioners are committed to the context, frame, and scenarios in which the participant provides the data, believing it to be imperative to making "analytic sense" of the data (Charmaz, 2006).
A set of research questions or hypotheses may guide the data collection, and the data themselves are the primary driver of the conclusions derived; a researcher is challenged to explore the data with a clear, reflexive mind, not one hunting for supportive evidence. In practice, a human-centered thinker may enter a space, begin an interview, or launch an open-ended survey with a few animating questions (e.g., "How do customers understand our brand?" or "What happens between account setup and first purchase?").
These questions would offer boundary condition on where to begin, what to ask, and what to note. The usefulness of grounded theory approaches surfaces in the curious, perplexing, and unexpected findings of the data. Very often, UXRs applying grounded theory methods include a section in their shareout or deliverable on "unexpected findings" or "future directions," a la a peer reviewed report. The opportunity for surfacing emerging needs spaces, friction points with experience mapping, or even broader strategic plays is a positive unintended consequence of grounded theory approaches.
The "code"—or short labels appended to data segments—is the heartbeat of apply grounded theory. As one is steeped in the data, themes within boundary conditions (e.g., answering the "why do our users slow down?" question) begin to present themselves. Coding is imperative to grounded theory. As luminary Kathy Charmaz describes, "Coding distills data, sorts them, and gives us a handle for making comparisons with other segments of data" (2006, p. 3).
A code is a thematic label applied to a segment of data, typically a section of a sentence or utterance. Codes should be mutually exclusive: that is, they are distinct from one another. Codes are compared to create analytic "memos" or notes about the ideas surfacing from the data. Codes are created for new ideas, nuances, angles, or potential "answers" to driving questions. One should strive for completeness in code amount, but should also be mindful not to rely on dozens and dozens of codes, which may dilute recommendations and conclusions.
Applying codes can be completed with and by collaborators, so long as adequate reliability—two people interpreting the same segment of data in the same way—is achieved. The combinations of codes serves to create interpretive mechanisms (theory) for what's going on in and with the data (Charmaz, 2006). Grounded theorists are constantly checking their codes, memos, and emerging interpretations against the data, which are the source of truth.
Grounded theory's methods mix with other approaches of user research, both quantitative and qualitative. Pairing grounded theory's ability to surface meaning, context, and value with quantitative data's scale creates a compelling triagulatory view of whatever is under investigation. Grounded theory's codes and frameworks can inform workshops, interview guides, or in-home explorations. Grounded theory methods are flexible, adaptable, and iterative: there is no wrong way to apply them, so long as one begins in and with the data.
Questions for Dr. Rachel Ceasar
Suggestions for advocating to use grounded theory in a quant-minded org?
I believe all research is subjective in nature--quant and qual. You’re telling the numbers or text to do something, to be organized in a way that’s based on a decision that’s based on your knowledge of the subject (or lack there of!). Having a method or standard in qual or quant to guide that thought process is needed. One way to make that thought process explicit is by writing down those processes and sharing them with team members. So in qual research, we use “memos” to make explicit our thought processes on a given transcript/data.
Ways of surfacing those unexpected insights that weren't part of the primary project?
One side thing is to create a depository of research organized so that the next person, your team, and/or your org can pick that work up and work from it--make recommendations or strategies that are data-driven. In terms of weaving in insights that don’t fit with your research goals/objectives. You can create a section--“new things we learned” or “unexpected discoveries” to share those things that don’t fit or were surprising. In the analysis process, this might be tracked from the start with a code “good quotes” or “unexpected discoveries”
Are there types of kinds of projects best suited for grounded theory methods?
Grounded theory can be applied to any type of project and data type. We only got into transcripts from one-on-one interviews, but you can also apply it to reports, images, and other artifacts. In terms of types of projects, I find it particularly helpful for discovery or exploratory projects--those projects like you mentioned above have a clear research questions/goal, but may also be interested in exploring and tracking unknowns.
How have you managed client or stakeholder interest in the data analysis process? Do you ever invite folks to participate in coding?
I do! I purposely use the codebook as a document to get not just formal sign off from stakeholders/my boss/client, but also for them to give their input so we track what’s important to them.
Sometimes these codes are thematic, and sometimes they can be addressed as a search item (like I want to track every time someone says the word “innovation”) or as the direct response to a particular question (I want to track the answer to question #5)
I’ve also been thinking a lot about my own contributions to white supremacy culture and how my research and methods add to this. The more I can adhere to a community-based participatory research model, the better, and someday, to really use CBPR methods. This means including community-based participants, as well as stakeholders and partners, in research throughout the research process and that their contributions aren’t just formal but meaningful as well.
Is there a particularly impactful way to package and present the findings grounded theory produces?
I find it’s important to make sure to advocate for people’s voices--and remind stakeholders/bosses/clients in the presentation that these aren’t just users, but that these are real people--showing photos of who you talked to and really storytell and give context to people’s decisions, needs, and aspirations.
Ben is the product evangelist at dscout, where he spreads the “good news” of contextual research, helps customers understand how to get the most from dscout, and impersonates everyone in the office. He has a doctorate in communication studies from Arizona State University, studying “nonverbal courtship signals”, a.k.a. flirting. No, he doesn’t have dating advice for you.