Themes do not emerge. An editor’s reflections on the use of Braun and Clarke’s thematic analysis

I have great admiration and respect for the authors Virginia Braun and Victoria Clarke. Since their pivotal 2006 article, Using thematic analysis in psychology, they have written extensively to help researchers understand, undertake, and write about thematic analysis. In fact, I’ve lost count of the chapters, articles and books the two have published, both as a pair and with numerous collaborators. We were keen to review their most recent book, Thematic Analysis: A Practical Guide (2022), and we are delighted with Sally Nieman’s (2023) comprehensive review. Braun and Clarke are also very generous, sharing materials such as datasets, examples, and recordings, again to help researchers to use thematic analysis. In fact, I feel I know them via their recordings as I’ve watched them present online many times. They are also very funny – we will come on to that later.

Despite these vast resources, my heart sinks when an article is submitted to QSW which states that they have used Braun and Clarke’s thematic analysis, but I can see that there is a sole reference to their 2006 article. This is not a rare occurrence. Unfortunately, basing a study on such a dated reference tends to invoke a desk rejection as it signals that the authors are unaware of the evolution of Braun and Clarke’s understanding of thematic analysis formed over the past two decades. Thus, I hope this Editorial will encourage researchers to read beyond the 2006 article. In the spirit of transparency, I will highlight some of the elements that I’m looking for as an Editor when authors state they have used Braun and Clarke’s thematic analysis. Disclaimer: I am not an expert on thematic analysis which means I will be quoting extensively from Braun and Clarke’s work. For a comprehensive understanding, I highly recommend their latest book (2022).

In their 2019 article, Reflecting on reflexive thematic analysis, Braun and Clarke describe how they met in the late 1990s as PhD students in a ‘large and lively community’ (p. 590) in the Department of Social Sciences at Loughborough University. Braun and Clarke (2019) acknowledge that the ‘educational privilege’ and the ‘doctoral training we received’ (p. 592) in the Department at Loughborough, ‘(unintentionally of course) turned us into intellectual terrors’ (p. 590). Thus, this is the context in which they wrote the 2006 article. However, 13 years later, Braun and Clarke say that that are ‘(hopefully) not ‘terrors’ anymore’ (p. 592). Braun and Clarke (2019) reflect that ‘dual frustrations’ led them to write their 2006 paper. First of all, they had observed at conferences and when reading published research, that researchers claiming to use thematic analysis used theories, methods and techniques which did not align in terms of ontology and epistemology. Second, the researchers didn’t make clear what processes they undertook which led to the production of the themes discussed in their paper. Moreover, in writing their 2006 paper, Braun and Clarke wished to put forward a ‘fully qualitative’ approach to thematic analysis which would ‘reflect our view of qualitative research as creative, reflexive and subjective, with researcher subjectivity understood as a resource’ (p. 591). Thus, the 2006 paper aimed to address these two issues and foster a fully qualitative approach to thematic analysis. Their approach to thematic analysis required ‘reflexivity, theoretical knowingness and transparency’ (p. 592) from researchers in relation to making clear both how their own assumptions informed their approach to thematic analysis and how the analysis was enacted.

Braun and Clarke (2019) reflect that they didn’t anticipate the ways in which their initial approach to thematic analysis (as outlined in the 2006 paper) would be misunderstood. While they assumed most researchers would understand the approach, Braun and Clarke (2019) have identified ‘countless examples of researchers… explicitly claiming to follow our process and procedures, but outlining and doing something different, either partially or fairly comprehensively’ (p. 592). They acknowledge that these misunderstandings are - in some part - a reflection of their own assumptions and commitment to qualitative values not being fully articulated in the original paper. Unfortunately, some of these misunderstandings have continued. For example, in their latest book, Braun and Clarke (2022) highlight that ‘Only a minority of published TA research demonstrates a use of theory that is knowing, reflexive, and clearly articulated’ (p. 189).

In their more recent work, these assumptions and the distinctive nature of Braun and Clarke’s approach as reflexive thematic analysis are clearly articulated (see Braun et al., 2019). However, it is important to note that their reflexive approach is always developing. Thus, one of Braun and Clarke’s (2019: 592) ‘key appeals to researchers drawing on our approach is to read some of our more recent writing.’ Reflexivity is now at the heart of Braun and Clarke’s approach to thematic analysis. Interestingly, the word reflexive is only mentioned once in the 2006 article, another reason why basing a study solely on this version does not reflect Braun and Clarke’s current approach to thematic analysis. Braun and Clarke (2022) ‘view researcher subjectivity, and the aligned practice of reflexivity, as the key to successful reflexive TA’ (p. 12) as ‘reflexive TA is not a neutral activity, but a values-based, situated practice’ (p. 22). As social workers, reflexivity is core to our education and our practice so this is an area of the analytic process in which we should excel. However, we do not often see this element within articles submitted to the journal.

In another example of their generosity to researchers, Braun and Clarke (2019: 593) list the key developments in their conceptualisation of thematic analysis when compared to the 2006 article. For example, in the 2006 article, thematic analysis was described as theoretically flexible. More recently, it is recognised that there are constraints dependent on the paradigmatic and epistemological assumptions implicit in what type of thematic analysis is used, of which Braun and Clarke’s reflexive thematic analysis is one approach. For Braun and Clarke (2022), approaches to thematic analysis ‘sit on a spectrum from Big Q [Qualitative] reflexive to small q postpositivist’ (p. 228), with their approach being crafted as a fully reflexive Big Q approach’ (p. 5 footnote). We can see this fully qualitative approach stems from their PhD education at Loughborough. However, Braun and Clarke (2022) note that there isn’t:

…just one way to do reflexive TA…. What TA produces all depends on the particular version you use, and exactly how you use it… researcher’s subjectivity and skill are at the centre of good reflexive TA. (Braun and Clarke, 2022: 9)

Thus, it is crucial for researchers to be clear about which type of thematic analysis they are using in a particular study. And equally crucial to specify this in the paper. Each approach has different philosophical assumptions which will underpin the process of undertaking and writing up the analysis. This includes being explicit about the theoretical approach taken. Theory is key to thematic analysis, yet we often see that there is no mention of theory in papers submitted to QSW.

Braun and Clarke’s (2022) six phases are an analytic process, not a rigid method, and are grounded by qualitative values, with creativity and reflexivity at the core. The process is ‘not strictly linear… [it’s] a progressive but recursive process’ (p. 36) with ‘unexpected twists and turns’ (p. 76) meaning that revisions and changes are possible throughout. In the 2006 version, Phase 3 of Braun and Clarke’s stages of thematic analysis was called Searching for themes. This changed to Generating (initial) themes’ (2019) and is now Generating initial themes in the 2022 book. The change reflects Braun and Clarke’s approach to thematic analysis, making clear that themes do not emerge from the data, a sentence we often see in articles submitted to our journal. Instead, the researcher actively constructs each of the themes through their engagement with the material. Thus, themes (and codes) are analytic outputs which describe broader, shared meanings. These are essential developments which researchers who just cite the 2006 version are missing and thus, their analyses remain highly descriptive and unrefined.

Additionally, there are changes to the names of three other phases of reflexive thematic analysis. In 2006, phase 4 was called ‘reviewing themes’; it is now Developing and reviewing themes (2022). Likewise, in 2006, phase 5 was called ‘defining and naming themes’; it is now Refining, defining and naming themes. Whereas the original seemed to convey staged end points in the analysis, the subtle name changes in phase 4 and 5 convey the iterative nature of thematic development in which the researcher is intricately involved. In 2006, phase 6 was called ‘producing the report’; it is now called Writing up (2022). These nuanced differences reflect the development of Braun and Clarke’s reflexive approach.

Unsurprisingly, themes are a central element of thematic analysis (the clue is in the name after all). In their later work (unlike the original 2006 article), Braun and Clarke discuss in depth the differences in their conception of themes as patterns of shared meaning, underpinned by a central meaning-based concept, compared to themes as domain summaries. The latter are themes where everything relating to a specific topic is summarised, and so can contain conflicting and diverse material. They are often given a one-word name which doesn’t give the reader a sense of the content. In contrast, in reflexive thematic analysis, a theme has coherence, and a strong shared concept or essence, with a concise name which captures the core patterning of meaning. Thus, in reflexive thematic analysis, ‘themes can be contradictory to each other, just not internally contradictory’ (Braun and Clarke, 2022: 107). It is vital for researchers to grasp this distinction from the very start. Furthermore, Braun and Clarke (2022) advise researchers to be judicious in the use of subthemes in reflexive thematic analysis as these can lead to a thinner analysis. They ‘generally recommend somewhere between two and six themes (including subthemes) for an 8000-word report’ (p. 89). As QSW has a word limit of 7000 words, we would not expect to see anything above this number in articles using Braun and Clarke’s reflexive thematic analysis.

We now move to another key part of preparing a paper for a conference or publication: Analysis and interpretation of extracts from the data. It is vital to understand that data ‘do not speak for themselves.’ We often see articles which simply state the theme and then list several extracts from the data (such as direct quotations from participants) as if the meaning is obvious and with no analysis or interpretation from the author(s). Braun and Clarke (2022) explain that thematic analysis ‘goes beyond data summary (data ‘reduction’) to interpret, and to explain the significance of the data, in relation to the research question’ (p. 264). Thus, there needs to be a balance between ‘data extracts and analytic narrative – the precise proportion depends on the type of analysis undertaken, but generally the analysis will consist of at least 50% analytic narrative’ (p. 265). Furthermore, the analysis must be connected to ‘to the scholarly fields your work is situated within’ (p. 129). Therefore, ensure that your paper reflects these guidelines.

The final points I wish to make relate to terms we often see mentioned in articles submitted to the journal: data saturation, sample size, and consensus on coding/themes through multiple coders adhering to a coding framework. Spoiler alert: Braun and Clarke are not fans of any of these in their approach to reflexive thematic analysis. In their latest book (2022), they ‘recommend avoiding claims of ‘saturation’’ (p. 28) and ‘now try to avoid using the language of ‘sample’ and ‘sample size’’ (p. 28 Footnote). In terms of coding, unlike some other forms of thematic analysis, there ‘is no codebook or framework that guides the process of coding in reflexive TA’ (p. 54 Footnote). For Braun and Clarke, having a single coder is ‘good practice, for reflexive TA’ (p. 55) and they ‘do not recommend using multiple coders as a way to guarantee a ‘true’ or ‘accurate’ analysis’ (p. 55 Footnote). If there are multiple coders, the aim is not to reach a ‘consensus’ about a code, but to develop ‘richer and more complex insights into the data’ (p. 55). Thus, we wouldn’t expect to see these terms mentioned in an article claiming to use Braun and Clarke’s reflexive thematic analysis.