Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative research in psychology, 3(2), 77-101.
This is my first attempt at what I will call a live discourse analysis (LDA) challenge. An LDA challenge is a flexible exercise that follows the following steps:
- Choose a piece of discourse to dialogue with. This can be an article, something in the news, a podcast episode, a YouTube video, etc.
- Choose three short quotes. You can choose these intentionally, or select them randomly if you want more of a challenge.
- Arrange those quotes into an order that tells a story you want to tell, and tell a story or make an argument based on the limits (and limitations) of your own living story.
This time I chose Braun & Clarke (2006) — arguably one of the most misunderstood papers in qualitative research methods. At least, it seems that way with the amount of digital ink Braun and Clarke have spilled since its publication to redress reader misinterpretations – a 20-year battle with misconstrual. Thus-
(1) “[Thematic analysis] can be seen as a very poorly ‘branded’ method, in that it does not appear to exist as a ‘named’ analysis in the same way that other methods do” (Braun & Clarke, 2006, p. 78).
This sentence hits different in 2026. Thematic analysis is a named analysis now. In applied linguistics, it’s the most named qualitative method, particularly for studies related to identity. And I don’t think applied linguistics is alone in this. The paper has 50,000 citations at minimum. What happened? Did thematic analysis become a well-branded method? Or does flexibility negate branding?
If I were to paraphrase the brand of thematic analysis…Thematic analysis calls for analysts to tell a story through the language of themes. What is a theme? A theme is a shared story across cases. And we live in a world saturated in story. Themes are everywhere.
So how do we notice or generate these themes? It’s up to you. But it’s also up to us. It’s up to us to document our shared patterns of research activity as we code our data for shared patterns of meaning. Braun and Clarke politely advocate for shared yet flexible methodological assumptions in a qualitative landscape adrift in implicit epistemological assumptions. But they also give bonus points to researchers who document systematic methodological moves. This emphasis on the systematic and its links to logical, scientific indexicalities outlines an enduring epistemological divide within qualitative research: are themes out there in the cultural-material world? Or are themes confined to our heads and tongues? Answers vary.
(2) “If themes ‘reside’ anywhere, they reside in our heads from our thinking about our data and creating links as we understand them” (Braun & Clarke, 2006, p. 80).
Braun and Clarke preached for reflexivity before the rise of the positionality statement. Conversations on positionality morphed to conversations on reflexivity. Yet a most important reflexive move, Braun and Clarke note, is to recognize that our methodological moves are interpretive. All analysis strives to be interpretive and grounded. Yet everything else — from study designing to question asking to semi-structured interviewing to broad or close transcription to analysis is interpretive. This asks us: Who is the interpreter?
For many, the interpreter is a researcher committed to science before storytelling if forced to choose. Yet even within a methodology as flexible thematic analysis, it is difficult to do justice to both science and storytelling at once. A commitment to science requires robust, novel, and, ideally, applicable findings. Robust findings require a systematic and trustworthy method of data collection, analysis, and representation as research reports. A commitment to storytelling requires, at times, suspend readers’ disbelief. It prefers occasional abductive leaps connecting far flung fragments of data over predictable, systematic inductive and deductive processes of coding and theming.
In applied linguistics, we have a lot of teachers. And teacher-scientists drawn to evidence-based practices want robust and generalizable findings. Such commitments to logic and science also calls for a predictable repetition of forms and structure in the representation of data and findings. Some call this the IMRAD structure — Introduction-Methods-Results-Analysis-Discussion. Storytellers are loyal to different conventions. Storytellers are not free from convention, but good stories also know when to skirt convention for rhetorical and argumentative effect. Good writing is risky writing — writing that takes leaps beyond stratified convention and expectation. Yet within these conventions, we also sediment basement-level epistemological convictions. It is, therefore, here in writing in reading, that epistemological commitments clash. Yes, thematic analtsis is flexible, but as 20 years of publications by Braun and Clarke show, it is not flexible enough to survive the clash. They have spilled gallons of ink to rein in what they see as misconceptions and misapplications of thematic analysis. It is flexible and limited — limited by Braun and Clarke’s inscription of six steps and the centrality of the theme.
Reading Braun and Clark (2006) and subsequent publications, tensions between systematicity, reflexivity, and transparency arise. If an analyst’s approach is systematic and linear, there is no conflict. Writing transparent, systematic methods is simple. But thematic analysis is far from a linear process.
(3) “Analysis is not a linear process of simply moving from one phase to the next. Instead, it is more recursive process, where movement is back and forth as needed, throughout the phases” (Braun & Clarke, 2006, p. 86).
Braun and Clarke created six steps (I almost said a six-step list). These six steps, they argue, provide a portal through which new qualitative researchers start doing thematic analysis. But many people miss the word “start.” Once we start, our paths diverge. We jump between data sources, following our analytical intuitions into a story of themes. Analysis becomes nonlinear, recursive, and far from replicable in a scientific sense. Following steps as if they were a recipe that will yield compelling results as predictably as cookies baked at 385 for 10-12 minutes. A recipe, however, cannot sustain within the flexibility of thematic analysis unless analysts are willing to improvise.
I used to be very critical of thematic analysis, inheriting critiques of Martin Packard and David Boje. Both, in their own ways, resist the ways thematic coding decontextualizes excerpts of data for the purpose of comparing statements across cases — a methodological movement that abstracts extracts away from the situated, interactional events that produced them. Braun and Clarke actually acknowledge this, saying TA is a great methodological fit for comparing patterns of meaning across cases, but not for unraveling the tensions and contradictions latent within a single case. To this they defer to discourse or narrative analytic approaches. This is striking considering thematic analysis is the most common analytical method among narrative-focused studies in applied linguistics — an all-too-common strident epistemological soup.
I encourage everyone to read Braun and Clarke (2006) if you do qualitative research. It is almost prophetic in its present-day problems. It is a poorly branded method that became wildly popular. And in its popularity it took on a brand that Braun and Clarke have spend the last 20 years trying to recast. Braun and Clark’s reflexive thematic analysis recognizes the tension between transparency and opacity; between systematicity and flexibility. Others cite Braun and Clarke but misread transparency as validity and misread systematicity as the defining feature of rigor. The question I have is: What does a transparent, non-systematic analysis look like? Braun and Clarke repeat TA’s flexibility, yet at times read as rigid when describing underdeveloped themes as having too much overlap or not accounting for a majority of the data.
Yet what really fuels my departure from thematic analysis is not something greater transparency or systematicity can resolve. Rather, it’s that Braun and Clarke are enamored by the potentiality of the theme and I am not. They seem to think through themes. But since reading David Boje, I have become enamored by thinking through antenarrative — fragments of story discourse that cohere with other circulating stories to produce narratives. In some ways, this is TA by a new name. Both advocate for stories and to connect patterns of meaning across disparate sources of data. Thinking through antenarratives does not arrive at themes orbiting a central concept. It culminates in movements between story threads rife with tensions, contradictions, and surprises. It honors data as a co-constructed interactional event and locates tensions not only within language, but within that co-construction itself. TA asks for analysts to generate themes. Antenarrative analysts generate story lines. Both ground analyses in data. Yet both also assume different analytical names and, by extension, different methodological movements with different final products.
Because as TA’s admitted poor branding shows, names matter. And I argue that names are only so flexible in the field of social discourse. Different names unlock different indexical fields — different storytelling fields — and therefore different worlds of discourse through which the analyst tells stories. I therefore do not advocate against TA. Some people live through themes and ought to let their freak flag fly.
Rather, I ask analysts: What is that unique unit of analysis through which you think? It might be themes. Or it might be something else. That reflexive self-discovery requires not accepting or rejecting methods, but for cultivating our own gardens of terms and connections — to tell the stories only we are situated to tell in a voice that is distinctly ours. It also requires us to negotiate a core tension in writing — to write through our own analytic process while making that process legible to readers. Interpreters also assume the role of translators. This is not branding. It is honing and owning our own analytical fingerprint — not for profit but for creativity’s sake. Yet analysis also remains forever unresolved. Manuscripts are only resolutions for now. Some find this energizing; others find it unsettling. Regardless, it makes qualitative research feel alive.
Leave a Reply