There is a single sentence that tells an experienced examiner whether a candidate has understood their own method: "themes emerged from the data." In reflexive thematic analysis, themes do not emerge. They are constructed by the researcher, actively, through interpretive engagement with the data, and the belief that they emerge is not a harmless turn of phrase. It signals a misunderstanding of the entire epistemological basis of the method, and it is the most common reason reflexive thematic analysis chapters are sent back or challenged in the viva. Virginia Braun and Victoria Clarke, who developed the approach, have spent more than a decade publicly correcting this and a cluster of related errors. They renamed the method "reflexive" thematic analysis and rewrote the names of its phases. Candidates who cite only their original 2006 paper and write as though nothing has changed since, walk into the exact trap the authors have been warning against. Our dissertation chapter support is built to catch these errors before an examiner does.
This guide sets out what reflexive thematic analysis actually requires, the six phases in their current form, the specific mistakes that fail examiners, and the exact language to use and to avoid when you write it up.
Quick Answer:
Reflexive thematic analysis (RTA) is Braun and Clarke's approach to identifying patterns of shared meaning across a qualitative dataset, in which the researcher's subjectivity is treated as a resource rather than a bias to be eliminated. The errors that most often fail examiners are: writing that themes "emerged" (they are actively constructed); building themes as topic or domain summaries rather than as patterns of shared meaning organized around a central organizing concept; mixing RTA with incompatible practices such as inter-rater reliability and Cohen's kappa; failing to engage reflexivity; and citing only the 2006 paper while ignoring the authors' later corrections. The six phases, in their current wording, are familiarizing, coding, generating initial themes, developing and reviewing themes, refining and defining and naming themes, and writing up, and they are recursive, not linear.
What Reflexive Thematic Analysis Actually Is
Thematic analysis is not one method but a family of methods, and treating them as interchangeable is where much of the trouble begins. Braun and Clarke distinguish three broad schools. Coding reliability approaches, associated with authors such as Boyatzis and Guest, use structured codebooks, multiple independent coders, and measures of inter-rater reliability such as Cohen's kappa, and they treat coding as something to be verified for accuracy. Codebook approaches, such as framework analysis and template analysis, use a structured codebook but retain a qualitative philosophy. Reflexive thematic analysis, their own approach, sits apart from both: it uses no codebook, treats a single engaged coder as entirely appropriate, and understands the researcher's interpretive subjectivity not as a source of error but as the very instrument of analysis.
This last point is the heart of the method and the source of most misunderstandings. In reflexive thematic analysis, there is no assumption that a "correct" set of themes lies waiting in the data for any competent coder to find. Meaning is generated through the researcher's active, situated interpretation, which is why the method is called reflexive: the researcher is required to reflect on how their own position, assumptions, and choices shape the analysis. Importing practices from the coding reliability school, most commonly a second coder checking for agreement and a kappa statistic to prove it, is not a neutral addition of rigor. It is a philosophical contradiction, because it assumes the accuracy framework that reflexive thematic analysis explicitly rejects. Examiners who know the method spot this contradiction immediately, and it is one of the fastest ways to signal that the approach has been misunderstood.
The Six Phases, in Their Current Form
Braun and Clarke's method proceeds through six phases, and the names matter because the authors deliberately changed several of them to correct misreadings. In their current, 2022 wording, the phases are: familiarizing yourself with the dataset; coding; generating initial themes; developing and reviewing themes; refining, defining, and naming themes; and writing up.
The most important change is the third phase. In the original 2006 paper, it was called "searching for themes," a phrase the authors came to regret because "searching" implied that themes were already present in the data, waiting to be found. The current wording, "generating initial themes," makes the constructive nature of the work explicit: you generate themes, you do not locate them. Phase one, familiarization, means immersing yourself in the data through repeated reading and noting early analytic observations. Phase two, coding, means applying codes that capture features relevant to your research question, working through the entire dataset with equal attention rather than seizing on a few vivid extracts. Phase three, generating initial themes, means constructing candidate themes as patterns of shared meaning from clusters of codes. Phase four, developing and reviewing themes, means checking those candidate themes against the coded data and the full dataset, and reshaping them. Phase five, refining, defining, and naming themes, means determining the essence of each theme and its central organizing concept and giving it a name that captures it. Phase six, writing up, means producing the analytic narrative, woven together with data extracts, that makes the argument of the analysis.
Table 1: The Six Phases of Reflexive Thematic Analysis (Current vs Original Wording)
Phase | Current Name (2022) | Original Name (2006) | What It Involves |
|---|---|---|---|
1 | Familiarizing yourself with the dataset | Familiarizing yourself with your data | Immersion through repeated reading and early notes |
2 | Coding | Generating initial codes | Coding features relevant to the question, whole dataset |
3 | Generating initial themes | Searching for themes | Constructing candidate themes from clusters of codes |
4 | Developing and reviewing themes | Reviewing themes | Checking themes against coded data and full dataset |
5 | Refining, defining, and naming themes | Defining and naming themes | Determining each theme's central organizing concept |
6 | Writing up | Producing the report | Building the analytic narrative with data extracts |
The single most important thing to understand about these phases is that they are recursive, not linear. You do not complete one and move cleanly to the next. You move back and forth, returning to coding when a theme will not hold, revisiting the data when a definition shifts. A write-up that presents the phases as a tidy linear sequence, each completed once in order, misrepresents how the method works and tells an examiner the candidate followed a recipe rather than conducted an analysis. Analysis of qualitative data is one of the sections examiners probe hardest, a pattern our guide on the ten most common dissertation mistakes covers across the whole thesis.
The Emergence Fallacy: Why "Themes Emerged" Fails
The language of emergence is worth dwelling on, because it is both the most common error and the most revealing. When a candidate writes that themes "emerged from the data," they are, usually without realizing it, making a claim about the nature of knowledge: that the themes existed independently in the data and the researcher simply reported them. This is precisely the position reflexive thematic analysis rejects. Braun and Clarke are emphatic that themes are generated, created, or constructed by the researcher, not identified, found, or discovered, and certainly not spontaneously emergent. They have described the image of themes emerging as if they were a fully formed figure rising from the sea, and they treat it as a marker of a fundamental misunderstanding.
The reason this matters beyond terminology is that the emergence framing lets the researcher disappear from their own analysis. If themes simply emerged, there was no interpretive work to account for, no positionality to reflect on, no choices to justify. That is the opposite of what the method demands. Reflexive thematic analysis requires the researcher to be visible and accountable as the author of the analysis. This is why the fix for the emergence problem is not merely swapping one verb for another; it is engaging seriously with reflexivity, showing how your position and decisions shaped the themes you built. A chapter that claims themes emerged has, in a single word, undone the reflexivity that gives the method its name.
Themes Are Not Topic Summaries
The second structural error is subtler and just as damaging: building themes that are really topic summaries. A theme, in reflexive thematic analysis, is a pattern of shared meaning organized around a central organizing concept, a single idea that unifies the disparate extracts gathered under it. A topic summary, by contrast, simply collects everything participants said about a given subject. The two look similar on the page and are completely different analytically.
The tell is usually the theme's name. Themes named after the topic, or worse, after the interview questions, are almost always domain summaries. If your themes are called "views on supervision," "experiences of the ward," and "barriers to access," they are buckets for content, not analytic claims. Each will contain internally contradictory material, because it is organized by subject rather than by meaning. A genuine theme has a point of view; it says something. A useful diagnostic is to ask whether each theme has a central organizing concept that could be stated in a sentence. If it cannot, if the theme is just a heading under which related quotes sit, it is a topic summary and needs to be developed into an actual theme or dissolved. Themes can contradict one another across an analysis, but a single theme must be internally coherent around its organizing idea. This distinction is frequently where a promising analysis falls apart, and it is one of the things we look at first in dissertation support.
Mixing Incompatible Methods
The third error is methodological incoherence, and it usually takes a specific form: a candidate declares reflexive thematic analysis, then reports that two coders independently coded the data and reached an inter-rater reliability of some value, or a Cohen's kappa. This is meant to demonstrate rigor. It demonstrates the opposite. Inter-rater reliability belongs to the coding reliability school, which assumes there is an accurate coding to be agreed upon. That assumption is incompatible with reflexive thematic analysis, where coding is interpretive. A second coder's different reading would be a resource for reflection, not a threat to reliability. Reporting kappa alongside reflexive thematic analysis tells an examiner that the candidate has assembled procedures from different, contradictory traditions without understanding what each assumes.
The same incoherence appears in three other habits: applying the language of saturation to reflexive thematic analysis, framing reflexivity as an effort to "eliminate bias," or describing the goal as capturing the themes "accurately." Each of these smuggles in a positivist assumption that the method rejects. Coherence between your stated method, your epistemology, and your actual practice is what examiners check, and mixing schools breaks it. If you truly want a coding reliability approach, with multiple coders and kappa, that is a legitimate choice, but then you should cite and follow that school, not Braun and Clarke's reflexive approach, and say so. What you cannot do is claim one and practice another.
Stating Your Theoretical Assumptions
The fourth error is silence about the assumptions underlying the analysis. Reflexive thematic analysis is a flexible method that can be conducted from different theoretical positions, and that flexibility carries an obligation: you must state where you stand. A rigorous write-up specifies four things. First, its epistemology: essentialist or realist on one hand, or constructionist on the other. Second, its orientation: experiential, staying close to participants' own meanings, or critical, interrogating them. Third, whether coding and theme development were primarily inductive, driven by the data, or deductive, guided by existing theory. Fourth, whether the analysis works at a semantic level of explicit surface meanings, or a latent level of underlying ideas and assumptions. These are not decorative labels. They determine how the analysis should be read and judged, and an examiner cannot assess whether your analysis is coherent without knowing which choices you made. Inductive analysis, in particular, does not mean analysis conducted in a theoretical vacuum; it means themes were built from the data rather than from a prior framework, which is a different claim.
Worried your thematic analysis will not survive the viva? |
|---|
Send us your analysis chapter and your dataset. A qualitative methodologist will check whether your themes are patterns of shared meaning or topic summaries, whether your method and epistemology cohere, and whether your write-up uses the language an examiner expects. Request a qualitative analysis review and receive an itemized quote within 2 to 4 business hours, no obligation. |
The Language: What to Write and What to Delete
Because reflexive thematic analysis is judged partly on whether your language reveals a correct understanding of the method, the specific words you choose in the write-up carry real weight. This is not stylistic fussiness; examiners read the vocabulary as evidence of comprehension.
Use the language of construction. Write that themes were generated, constructed, developed, or created. Name the method reflexive thematic analysis, and cite beyond the 2006 paper to the authors' later work. Head your analysis section "Analysis" rather than "Findings" or "Results," because "findings" implies you found something pre-existing and "results" belongs to statistical work. Describe your active role and your reflexivity explicitly. State your epistemology and your semantic or latent, inductive or deductive choices.
Delete the language of discovery. Do not write that themes "emerged," "were found," "were identified," or "were discovered." Do not cite only Braun and Clarke 2006 as though the method had not developed since. Do not report inter-rater reliability, Cohen's kappa, or consensus coding as evidence of rigor. Do not claim saturation. Do not describe reflexivity as a way to remove bias. Do not name your themes after your interview questions. Each of these is a specific flag that experienced examiners look for, and removing them is among the simplest high-value edits you can make to a reflexive thematic analysis chapter.
Table 2: Reflexive Thematic Analysis Write-Up: Language to Use vs Avoid
Use This | Not This | Why It Matters |
|---|---|---|
Themes were generated / constructed / developed | Themes emerged / were found / were discovered | Themes are actively built, not pre-existing in the data |
Reflexive thematic analysis (cite 2006 + later work) | Thematic analysis, citing only 2006 | Shows you know how the method has developed |
"Analysis" as the section heading | "Findings" or "Results" | "Findings" implies discovery; "Results" implies statistics |
The researcher's active, reflexive role | Reflexivity used to "eliminate bias" | Subjectivity is a resource, not an error to remove |
A single engaged coder is appropriate | Inter-rater reliability / Cohen's kappa / saturation | These belong to a different, incompatible TA school |
What Good Reflexive Thematic Analysis Looks Like
Set against these errors, a strong reflexive thematic analysis has a recognizable shape. It names the method precisely and cites the current literature, not just the 2006 paper. It states its epistemology and its analytic choices openly. It engages reflexivity substantively, showing how the researcher's position shaped the analysis rather than asserting neutrality. Its themes are genuine patterns of shared meaning, each with a statable central organizing concept, named for what they claim rather than the topic they cover. It presents the phases as the recursive process they are. It carries an analytic narrative that interprets rather than merely paraphrases the data, with extracts that illustrate the analytic points rather than standing in for them. And its number of themes is proportionate, typically a small number of well-developed themes rather than a long list of thin ones, because too many underdeveloped themes is itself a recognized weakness. A candidate whose chapter has these features can defend it in the viva, because the method, the epistemology, and the practice all line up, which is exactly what our dissertation defense preparation guide helps candidates rehearse.
Choosing the Method Without Self-Deception
One final point, because it prevents the deepest kind of error. Reflexive thematic analysis should be chosen because it fits your research question and your epistemological position, not because it has a reputation for being accessible. Its flexibility is real, but flexibility is not the same as ease, and the method makes genuine demands: interpretive depth, reflexive honesty, and theoretical clarity. If your question calls for a structured, codebook-driven analysis with multiple coders, a different approach is the right one, and choosing it openly is far stronger than forcing reflexive thematic analysis to behave like something it is not. The decision between qualitative approaches and between qualitative and quantitative designs in the first place is the foundation on which everything else rests. Our guide on whether your dissertation should be qualitative or quantitative works through that choice, alongside our guide on choosing a dissertation topic that works.
Frequently Asked Questions
Do themes emerge in reflexive thematic analysis?
No. In reflexive thematic analysis, themes are actively constructed by the researcher through interpretive engagement with the data, not discovered lying within it. Writing that themes "emerged" is the single most common error that fails examiners, because it denies the researcher's interpretive role and undercuts the reflexivity the method is built on. Use "generated," "constructed," or "developed" instead, and show how your choices shaped the themes.
What are the six phases of reflexive thematic analysis?
In their current wording, the six phases are: familiarizing yourself with the dataset, coding, generating initial themes, developing and reviewing themes, refining and defining and naming themes, and writing up. The third phase was renamed from the original "searching for themes" because searching implied that themes pre-existed in the data. The phases are recursive, not linear: you move back and forth between them rather than completing each once in sequence.
What is the difference between a theme and a topic summary?
A theme is a pattern of shared meaning organized around a central organizing concept, a single unifying idea, and it makes an analytic claim. A topic summary simply collects everything participants said about a subject and is usually named after that subject or an interview question. Topic summaries contain internally contradictory material because they are organized by topic rather than by meaning. If a theme has no statable central organizing concept, it is a topic summary and needs to be developed.
Can I use inter-rater reliability with reflexive thematic analysis?
No. Inter-rater reliability and Cohen's kappa belong to the coding reliability school of thematic analysis, which assumes there is an accurate coding to be agreed upon. Reflexive thematic analysis rejects that assumption: coding is interpretive, and the researcher's subjectivity is a resource, not a threat. Reporting kappa alongside reflexive thematic analysis is a philosophical contradiction that signals to examiners the method has been misunderstood. If you want a reliability-based approach, cite and follow that school instead.
Why can't I just cite Braun and Clarke's 2006 paper?
Because the method has developed substantially since, and the authors have publicly corrected common misreadings, renamed it reflexive thematic analysis, and rewritten several phase names. Citing only the 2006 paper signals that you have not engaged with these developments, and some editors treat it as grounds for rejection. Cite the 2006 foundation alongside the later work, including the 2019 and 2021 papers and the 2022 book, to show you understand the method as it now stands.
What theoretical assumptions do I need to state?
State your epistemology (essentialist or realist versus constructionist), your orientation (experiential versus critical), whether your analysis is inductive or deductive, and whether you code at a semantic or a latent level. Reflexive thematic analysis is flexible and can be conducted from different positions, so an examiner cannot judge the coherence of your analysis without knowing which one you chose. Inductive analysis does not mean a theoretical vacuum; it means themes were built from the data rather than from a prior framework.
How many themes should a reflexive thematic analysis have?
There is no fixed number, but a small set of well-developed themes is stronger than a long list of thin ones. Too many themes, or themes that are underdeveloped or that simply restate the interview questions, are a recognized weakness. Each theme should be a fully realized pattern of shared meaning with its own central organizing concept, developed with enough analytic depth to say something substantial rather than merely label a topic.
You May Also Find Useful
Writing an Analysis That Holds
The reflexive thematic analyses that pass are the ones where every part lines up. The method is named correctly and cited currently, the epistemology is stated, and the reflexivity is real. The themes are built as patterns of shared meaning rather than topic summaries. And the language is free of the emergence and reliability framings that betray a misunderstanding. Construct your themes rather than claiming to discover them, keep your method and your philosophy coherent, and make your own interpretive role visible on the page. Do that, and your analysis chapter stops being the place an examiner probes for weakness and becomes the demonstration that you understand the method you used.
If you want a qualitative methodologist to check whether your themes hold and your method coheres before your viva, send us your analysis chapter. You will have an itemized quote within 2 to 4 business hours, with no obligation.
