Braun and Clark 2022 Chapter 7 Thematic Analysis

Going deeper into interpretation and thematic analysis.

What should interpretation be embedded in?

Interpretation should be embedded into the analytic process in reflexive TA, rather than being a separate optional add-on.

What is interpretation not?

At the most basic, interpretation is simply the act of making sense of something. That said, it is sometimes (incorrectly) assumed to be a process that people implicitly understand, or it is just not explained well. Theory always grounds and delimits data interpretation. Interpretation is not some mystical process, but instead involves “doing what human being do” (J.A. Smith, 2019, p. 171).

How has ‘reflexive TA’ been misinterpreted?

Reflexive TA has sometimes been misinterpreted as a purely descriptive method, as if it offered a way of simply conveying information from point A (data collection) to point B (report), like a delivery drone who picks up an order from an Amazon warehouse, and drops it at your home. Far from it! Even more-descriptive accounts of data, which stay close to participant or text meanings, both require and reflect interpretative work. Qualitative analysis is always an interpretive activity. You are not a magician; meaning is not self-evident or just sitting in data, waiting for your flourishing reveal. If you start to think about interpretation as integral to doing (thematic) analysis as breathing is to living - and that’s how we believe it is useful to conceptualize it - then you start to appreciate how hard is it to talk about interpretation as a separate process and practice.

What does the quote interpretation is not some mystical process, but instead involves “doing what human beings do” by (J. A. Smith, 2019, p. 171) mean in qualitative research?

This quotation about interpretation as an essentially human practice comes from the developer of interpretative phenomenological analysis (IPA), Johnathan Smith, talking about IPA processes.

What does is mean by ‘reflexive TA’ being used as a descriptive method?

Reflexive TA has also been used as a descriptive method, which Janice Morse recently described as a form of “incomplete inquiry” (2020, p. 4). Morse warned against (such) “weak research that ‘signifies nothing’, simply theming for the purpose of theming” (2020, p. 4).

How has and can language cause confusion in interpretation within qualitative research?

Language might also contribute to confusion about the role of interpretation in qualitative research, as there are many different ways words like ‘analysis’ and ‘interpretation’ get used. Although early qualitative scholars framed a qualitative paradigm - in contrast to the empiricist tradition - as an interpretative one (for instance, in discussing quality, Yvonna Lincoln [1995] described “the entire field of interpretative or qualitative inquiry”, p. 275), this isn’t necessarily reflected in language used in qualitative reporting. Carla Willig (2017) suggested that qualitative researchers may have taken up the term analysis in preference to interpretation to describe what we do, to better align with the empiricist orientation that dominates the social sciences, as a way of rhetorically claiming validity. Within a quantitative positivist-empiricist tradition, a differentiation between doing statistical analyses and then interpreting the meaning of the results of those probably makes sense. But, in qualitative analysis, we feel analysis and interpretation are better understood as felted together, as impossible-to-separate strands.

How should you conceptualize your analytic task in qualitative research during interpretation?

It’s easy to conceptualize research and analysis as revealing the truth about something. Even if you understand truth-telling as your purpose with qualitative analysis, truth-telling framing can be unhelpful in developing an interpretative orientation for analysis. Our task is not to stand up in court and ‘tell the truth, the whole truth, and nothing but the truth’. Instead, we like Michael Quinn Patton’s (1999, p. 1205) description that the analytic “task is to do one’s best to make sense out of things”. Instead of conceptualizing your analytic task as one of discovering, distilling and revealing the essence of the data, we suggest it’s better to imagine you’re telling a story in a way that aims to make sense of what’s going on. A story that gives the audience (a reader, a listener) a clear take-home message - one that includes an indication of why they should care about the story you’ve just told them. This means you have to be clear what the take-home messages you want the audience to leave with are, and you have to be clear on why you think those particular take-home messages, those interpretations, are valid and important. This is why analysis cannot simply be a compilation of quotations of data, the meaning of which is treated as self-evident. Analysis needs a strong authorial narrative, which takes the reader beyond the data. Data extracts provide the reader with the tools to evaluate your analytic narrative and interpretative claims.

When does Interpretation start?

Interpretation starts during familiarization though it’s quite likely you will have already started interpreting the data, if you’ve been engaged in data collection. During familiarization, or data collection, you often make observations or have ideas about things that are going on - the sorts of things you record in your familiarization notes. Interpretation at that stage is tentative and should be recognized as such - don’t cling tightly to early interpretations, assuming you’ve noticed everything and made sense of the data in the best way.

How does the interpretative process for TA need to operate?

The interpretative process for TA operates most strongly as you move from coding into theme generation, development and refinement, and is honed in and through writing. Interpretation needs to operate in concert with your research question - even if the particulars of that question aren’t yet refined. Just as there are endless patterns we could focus on, it is the ones that matter most in addressing our research question that become part of our developing analysis, so interpretation needs to be anchored to the research question(s) and purpose of a project.

What do our research questions provide us with?

Our research question(s) and purpose, alongside our dataset, provide the foundation for our interpretation; wider contextual elements provide the scope for it. A key practice point is to stay oriented to your research question - oriented but not shackled, as it is not fixed. Asking questions in relation to the patterns you’re developing, and the implications of them, can be useful in helping to develop an interpretative analytic orientation, one that moves beyond describing semantic content. Asking questions of yourself and your analysis is a useful tool for ‘going deeper’ into interpretation.

What is interpretation?

Interpretation is the meaning-making we engage in, and it does not occur in a vacuum, it’s not fixed, and it’s not self-evident from a stimulus itself.

What are some general key points about interpretations

• Interpretation depends on us

• Interpretation depends on contexts

• Interpretation brings together all our knowledge related to the subject or object at hand

• Not all interpretation holds up to scrutiny - it needs to be defensible

In relation to analysis, what is your interpretation based on?

In relation to analysis, your interpretation is never just based on the dataset. While the dataset grounds it, scholarly knowledge, theory, ideology, politics and all sorts of other factors can come together in how you make sense of meaning - and even the meaning patterns you notice. Your argument for what those patterns mean needs to be discussed in relation to the aspects that lead you to a certain interpretation. Good analytic practice requires that we explain our interpretative tools and contexts for understandings to convince the reader of our interpretation. The question is, why should we be believed.

What is the key take-away for us as qualitative researchers?

The key take-away is that for us as qualitative researchers, interpretation is inevitably subjective and there is no absolute singular, correct interpretation. Our framework for reflexive TA - with phases and processes for rigor, including revisiting the whole dataset, and the requirement for being an active thinking researcher who makes choices - provides tools to facilitate a thorough and data-connected interpretation.

How should qualitative researchers be guided when in comes to interpretation?

When it comes to interpretation, qualitative researchers should try to not be guided by the question ‘am I doing this right?’ so much as by the questions ‘are there good grounds for what I’m claiming?’ and ‘am I ignoring some inconvenient “truths”?’ If you can answer yes to the second, and no to the last, you’re well on your way to a defensible interpretation.

What does it mean by ‘interpretation needs to be defensible?’

As interpretation goes beyond the data, the challenge, or balancing act, nicely captured by UK-based critical psychology scholars Carla Willig and Wendy Stainton Roger, is: “to go beyond what presents itself, to reveal dimensions of a phenomenon which are concealed or hidden, whilst at the same time taking care not to impose meaning upon the phenomenon, not to squeeze it into pre-conceived categories or theoretical formulations, not to reduce it to an underlying cause” (2008, p. 9). As qualitative researchers, we need to keep asking whether our interpretation has moved too far from the data. Finality, permanence and correctness are not hallmarks of reflexive TA, nor of the entire field of Big Q qualitative researching. So you do have to try to settle in to some degree of uncertainty to conduct qualitative research effectively. Make sure that you’re not shaping the data to tell your story, or smoothing over complexity to (mis)represent the stories in the data. Don’t (mis)use data to tell your story, rather tell your story of the data. Don’t try to make the data fit your narrative.

What are qualitative researchers aiming for with interpretation?

We as researchers need to do the work to show how our interpretation is defensible, in light of the intersection of: (a) our topic, existing knowledge and the wider context that surrounds it; (b) the dataset that we’re working with; (c) the theoretical frameworks we’re working with; (d) our individual position(s); and (e) the processes we’ve engaged in to develop that interpretation - sometimes individually, sometimes in a research team, sometimes with the community the research connects with. In doing so, we have to convince our audience that we haven’t come to the data with our interpretation preformed, and that it has developed through our analytic process.

Example of defensible interpretation

To give an example of a tendency for a gendered analysis: as tempting as it is, we don’t stomp around shouting ‘that’s gendered!’ at everything we encounter, without giving it deeper thought. Without asking ourselves what that might mean, and if, and how, gender play out, and plays out in particularly classed and/or raced ways. Instead, we use our gendered lenses as a kind of signal to look at and think about something, more deeply and critically, in a particular way; to ask questions that locate gender as a key component to consider. We strive to use it to open up, rather than close down, meaning-making.

What does it mean by ‘what is defensible depends on context’?

What is defensible depends on context, and your task as analyst is to show how and why you believe your interpretation has validity. To come back to our text message example: imagine it is from your unreliable friend, and your reaction is to shout and be tempted to throw your phone across the room. Someone who witnesses your response tells you that you’re overreacting (which annoys you more!), but your response, in context, seems justifiable to you. It reflects the history that precedes the text, the rich context that grounds your response. The person judging your reaction doesn’t know that, unless you explain to them. To make your reaction defensible to the person you are with, you have to explain to them why that particular reading makes the most sense to you, and why it matters, effectively providing you a rationale for your reaction. If you explained that to the person with you, and they then agreed your reaction wasn’t an overreaction, you’ve shown your interpretation is defensible. That’s not to say that your reaction is the only one possible! It is not about saying you are right or no other reading is possible. It is about convincing someone else that the reading you’ve offered makes sense and matters.

How do we go from more descriptive to more interpretative modes of analysis?

All reflexive TA needs to involve interpretation. But your analyses can be situated somewhere along a continuum from primarily offering description (e.g. of an experience, such as of what it is like to be child-free, or of a concept, such as pronatalism) to interrogating, unpacking, and even theorizing (e.g. the assumptions underpinning the ways people describe being childfree); the more interpretative end of the continuum. Even within a more descriptive mode of interpretation, we’d still characterize the interpretation involved as “coming from the researcher and informed by insights from theory and other related research” (Chamberlain, 2011, p. 50), and an understanding of the wider context. Analysis always involves interpretation but it can be primarily descriptive, in the sense of ‘staying close’ to the data, to participants’ sense-making, or primarily interpretative, in the sense of bringing in the researcher’s conceptually informed lenses to interrogate the ideas expressed.

Example of going from a more descriptive to a more interpretative mode of analysis

In the worked example analysis of the childfree dataset, the contradictory theme good and bad parents offers an example that is fairly descriptive but also deeply interpretative. The theme describes a pattern that is evident at a quite semantic level (even if the idea of a contradictory theme is somewhat conceptual). In so doing, it provides a descriptive account, and one that would be easily recognizable to someone reading the dataset for themselves. However, our analysis is interpretative because we don’t stop at the point of identification of this pattern - instead, we ask what are the implications of this pattern? We also located our analytic interpretation of this pattern within the wider societal context to consider such implications. The worked example analysis developed within an overarching theme of choice matters offers an example from the interpretative end of the ‘more descriptive to more conceptually interpretative’ spectrum. This end of the spectrum often aligns with more critical, constructionist and latent approaches within reflexive TA. The analysis was developed through the interrogation of the logics of choice within the dataset. Instead of stopping at reporting that ‘choice’ was important in how people made sense of being childfree, our analysis brought a range of conceptual/theoretical tools to the interpretative process, as the authors sought to parse out the different logics of choice within the dataset. In doing so, they identified that there were quite different frameworks or understandings of choice at play, associated with bigger Anglo-western systems of meaning. The analysis went far beyond a direct reporting on recognizable patterned dataset content, to ask questions about what might be at stake when we talk about parenting within choice frameworks. Why might this meaning and interpretation matter?

How can ‘interpretative analyses’ matter?

More interpretative analyses can matter in a range of ways. They can matter through the contribution they make to theoretical discussions or understandings. They can matter through providing deeper or more complex understandings of what is at stake for some kind of practice or policy development. For example, Ginny’s analysis of ‘national identity’ explanations of sexual health risk in Aotearoa New Zealand provided access to meaning-making with direct consequences for how sexual health promotion should be framed, to increase uptake of the message (Braun, 2008). They can matter through providing a nuanced explanation for what might be going on, when more descriptive accounts just don’t quite feel like they’re getting to the heart of things.

What are experiential and critical orientations in interpretation of data patterns?

The more descriptive to more interpretative modes overlap with broad takes on the data whether your analysis is based in a more experiential or critical framework for qualitative researching (Bran & Clarke, 2013). An experiential orientation to TA grounds it within the meanings and life-worlds of the participants or within the meaning of the data. Critical orientations to TA tend to take more researcher-directed interpretative frames, or an interrogative approach, where interpretation is not entirely determined by data-based meaning (Braun & Clarke, 2013; Willig; 2017). Each position is based in different epistemological and ontological frameworks and allows us to make quite different claims, different types of interpretation.

What does it mean by shifting from experiential to more critical orientations in interpretation of data patterns?

Often, interpretation based around participant meaning might seem obvious or straightforward, and that focus might continue to be the basis on which your analysis develops and is ‘completed’. Or it might prove to be a step in the journey towards a quite different take on things. Turning to the author’s adventure analogy, imagine you’ve now returned home and are starting to write a story of your trip. You have loads of photos, social media posts, as well as a journal and other bits and pieces you picked up along the way. You start to write and find that as well as a description of what you did, you can identify and describe some key patterns across the diverse experiences. You’re pretty happy with this first attempt, and people are interested in it - you get it published in an online magazine you admire. But your curiosity has also been piqued, and you start to read other stories, and consider different information. In doing so, your position shifts. You read a critique of the colonial gaze and neocolonialism in and through travel and although it’s a bit uncomfortable, you wonder if you might inadvertently have reproduced these ideas in how you’ve written up your experience. You read around gender politics in the region, and reflect on how your sex/gender may have shaped not only how you’ve told your story, but the very experiences you had in the first place. You start to reflect and ask different questions about the ways you have understood your journey. Going back to all the evidence of your trip, you read and make sense again with new questions at the forefront of your mind. Questions such as: what sorts of things need to be in place for me to have this experience, and to interpret it this way? What sorts of assumptions does it rest on? What (problematic or not) ideas am I inadvertently reproducing in how I’ve told this story? These sorts of questions - and others - help you to develop a differently nuanced and more complex understanding, including an interrogation of how you are making meaning about your journey. And as a result of this, your account of the trip may change.

How is the adventure example of shifting from experiential to more critical orientations in interpretation of data patterns analogous of what can happen in reflexive TA?

This is analogous to what can happen in reflexive TA, as a researcher shifts from a more experiential to a more critical orientation within the scope of a developing analysis. Sometimes, you might develop one more experientially-oriented analysis, and then move on to a new, more critically-oriented one. Hence, your analytic orientation can shift from more experiential to more critical through your engagement with the data, resulting in your setting on one analysis. Or you can offer multiple interpretative ‘takes’ on the data.

What are some themes you want to develop when shifting from an experiential to a critical orientation to built analytic depth?

• Don’t get too attached to themes early on, as your analysis might radically change.

• You may shift across more inductive/experiential and more theoretically-informed approaches to TA.

• There is value in the reflexivity in analysis.

What is the value of TA in producing analyses?

The value of TA in producing analyses that illuminate our understanding of a particular issue, but which, instead of revealing the truth, can be regarded as critical interpretations arising from a context and a relationship. Between ourselves as researchers, our participants (if we have them), the data and our engagement with them, our theories around analysis, existing scholarship in this area, and any communities we might be working in/with.

What does it mean to use deductive orientation to work with existing theoretical concepts in doing interpretation?

To really get to grips with interpretation in reflexive TA, you need some understanding of the meta-theoretical positions - the ‘ologies’ that underpin and give validity to your interpretative practice. Theory is everywhere, and “even without great theoretical awareness, underlying theories will always be present, leading the researcher’s gaze”. To do good interpretative reflexive TA, you need to understand this, and be explicit around theory, and where and how theory informs the analysis. Big Theory, is theory at a philosophical or meta level - the ontological and epistemological positions we’re working with. We can understand data in quite different ways, and the ways we theorize them determine which sorts of interpretations hold validity, and which don’t. For instance, we can treat text as simply a conveyance for experience or opinion - in the childfree comments, we’d treat the views expressed as reflecting the opinions of the people who posted them. Alternatively, we can theorize our data as constructing rather than simply reporting realities and truths, and we might explore how the very concept of ‘being childfree’ is constructed in the dataset - what particular nature is ascribed to it.

What is another way that theory comes into play in doing TA?

Another way theory comes into play in doing TA is through theory-driven analysis, where interpretation utilizes existing theory to guide the developing analysis. This captures analysis that is strongly informed by existing theoretical constructs (e.g. heterosexism and heteronormativity) or a wholesale theory (e.g. Foucault’s theory of sexual ethics or identity process theory). In such strongly deductive or theory driven reflexive TA, you, as researcher, deliberately seek to explore, or develop your analysis in relation to, one or more pre-existing ideas or frameworks. You do not have to know this in advance! Sometimes early in the interpretative development, it may become clear that a more theoretically-informed analysis can tell a richer, more complex, and more useful story.

What does the whole process of interpretation rely on? What does explanatory theory provide?

The whole process of interpretation - sense-making - relies on enmeshed layers of theory. When working in a theoretically-driven way in TA, when working with explanatory theory, Big- or meta-theoretical positions provide the all-encompassing theoretical framework for interpretation. Big Theory permeates it all, like the air we breathe: mostly we don’t notice it, and we’re often unaware of it, but it’s constantly there.

What do other types or levels of theory include?

• The theory that informs our interpretation practices in broad conceptual ways, such as the ones related to phenomenology (Langdridge, 2017), discourse (Wiggins, 2017), or affect (Moreno-Gabriel & Johnson, 2020; Wetherell, 2015).

• Explicitly socio-politically inflected theories that you draw on to make sense of the possibilities and boundaries for understanding and experience within the material, symbolic, and power organization of our worlds - our immediate and broader contexts. Such frameworks include various feminisms (Collins & Bilge, 2016; McCann & Kim, 2013), postcolonial (Said, 1994) and decolonization theories (G. Adams & Estrada-Villalta, 2017; L. T. Smith, 2013), crip theory (McRuer, 2006), and many more.

• Theories that are more specific - you might call them lower-level theories - theories that focus on exploring or explaining a specific topic, mechanism or process. Within the mainstream of our discipline of psychology, this includes popular theoretical frameworks like social cognition (Carlston, 2013), and identity process theory (Jaspal & Breakwell, 2014). In mainstream psychology, this level of theory is often equated with theory, full stop; ontological and epistemological assumptions are rarely discussed.

What can we do to help us recognize that a theory is operating even though we may not notice it?

Theory is harder to notice if our explanatory frameworks are effectively the common-sense ones, the dominant or normative ones, because they are then shared or assumed by many, and frequently are invisible to us - just like the air we breathe. The trick is to recognize that theory is operating in all sorts of ways, even if it’s not explicated. Our best practice guidance is to try to be as clear as we can about the assumptions that inform, and validate, the interpretation we do in reflexive TA - and to use reflexivity as a tool for striving to recognize and interrogate these.

How do researchers normally go about theory-directed analysis?

Sometimes you will have a sense in advance (or early on) that you want to develop a theory-directed analysis. What you want to do with your interpretative work might be predominantly located and framed - at least initially - by ideas derived not from the data themselves, but from ideas already at play in the wider social or scholarly context. These ideas will direct your interpretive engagement with your data. In some sort of theory-directed interpretation, your conceptual and theoretical ideas, then, explicitly guide the sorts of ways you engage with the data (sometimes even how you collect the data), and the interpretations you develop in relation to this.

How is theory-directed reflexive TA different from quantitative hypothesis testing?

Theory-directed reflexive TA is not aiming to prove (or disprove) a theory or hypothesis. You don’t apply a theory to the data to test if the data evidence it. ‘Deductive’ reflexive TA offers a theoretical exploration of qualitative data, and remains embedded within a framework of openness and situated meaning connected to Big Q research values. This approach is not about massaging your data to fit into your preconceived notions, or telling a partial story of the data that fits with a pre-existing theory or concept. Rather, this analytic orientation recognizes and acknowledges the conceptual ideas we (always) come to data and a project with, and gives them greater analytic priority in our interpretative processes than in inductive-oriented analyses.

When working with data in a more theoretically-informed or directed way, what are some guidelines that will help to avoid (inadvertently) shoe-horning the data to fit an existing idea, and to demonstrate how to avoid this pitfall?

• Work from a curious, open and questioning position when engaging with data more theoretically. For instance, understand your task as exploring how a theory or concept is evidenced - and not evidenced - within the dataset, rather than one of merely identifying it in the dataset.

• Always keep your interpretations tentative. Don’t only seek affirming evidence, solidifying your analysis early on.

• Keep asking which data aren’t fitting with the developing interpretation, and, importantly, in what ways they aren’t fitting. This is about ensuring that the interpretative frames you’re bringing to the data aren’t obscuring a different story, a fuller story, or one that may be more important to the topic. Is your conceptual framework limiting what you are able to say about the data, in ways that provide either an impoverished analysis, or one that only gives a partial view? Reflect on the gains and losses connected to your particular theoretical lens.

• Always be wary of imagining your reading as the right or the only one possible from the data. Because interpretation is inherently subjective there is always potential - in theory at least - for multiple readings of data.

How can researchers be explicit and reflexive about how theory informs analysis?

The take-home message from Braun and Clark is that theory is unavoidable, when it comes to interpretation and analysis in TA, and it need not be scary. Recognizing that theory informs the whole endeavor is a useful starting point for being explicit and reflexive about how theory informs the analysis - at all levels. Knowing to avoid simply laying theoretical constructs over the top of the data patterns, and always, always being tentative in taking a deductive orientation, will set you up well for doing more theoretically-driven TA.

Why should researchers locate data within the wider context of their study?

It is good practice in reflexive TA to locate your interpretation within the wider context. This is not about providing a factual summary of the context of data collection, so much as treating your data and your participants (if you have them) as embedded in contexts that have inflected the data. Much as your interpretation is similarly inflected by your positions. We often describe a topic as if it happened in any place but it was really just the US or the UK. We also read about the experiences of a participant group as if they were any people or a mythical people but actually they were just White people from high consumption westernized countries. These experiences or perspectives are written about as if they’re universal and don’t need to be situated, don’t need to be considered as partial and contextual - but they inevitably are that. This ties to an important analytic practice of locating your interpretation in the wider context. By that, we mean that the particulars of the participant group/dataset, and the context of the research, are described or considered in interpretating the data, and making claims about them. Not to do so is to produce not only a poorer analysis, but to also reduce the reader’s ability to consider and evaluate things like the transferability of your study, the applicability of your analysis to their context (see Yardley, 2015).

What are some ways that interpretation can help researchers to locate the data at different levels, and orientate to various aspects of the wider context?

Interpretation can locate the data at different levels, and orientate to various aspects of the wider context. These various aspects, none of which is entirely separable from each other include, but are not limited to:

• Ideological aspects - such analysis would interpret the data in light of prevailing broad meaning-making frameworks - ideologies - that form the dominant common-sense of society. An example of an ideologically-located analysis would be consideration of neoliberal ideology, and how it shapes what is experienced, what is desired, and what is imagined (e.g. Chen, 2013; Gill, 2008; Schariff, 2016).

• Political aspects - such analysis would interpret the data in light of contemporary and/or historical political arrangement and governance. Political here could be big-P, formal government-related politics and practice. Or it could encompass small-p politics, related to the structuring, organization, and operation of power within society, leading to differently-organized potentialities and marginalizations for different groups.

• Historical aspects - such analysis would interpret the data with a longer-term view of meaning politics and ideology in society. Such analyses seek to recognize continuances from the past in terms of available meaning, as well as departures and disjunctures.

• Material aspects - such analysis would interpret the data in light of material conditions in which people’s lives are embedded.

• Policy aspects - such analysis would bring policy, contemporary and/or historical, to bear in the interpretation of data. Note that it is particularly important to watch out for the trap of ‘arguing with the data’ when doing policy-inflected interpretation, because you might easily be drawn into comparing policy and data, and reporting on ignorance of policy or ‘facts’ (such as demographic data about rates of childlessness), or misunderstandings of these.

• Discourse aspects - such analysis might explore the discursive ‘conditions of possibility’ (Gavey, 2018) in and through which the data need to be interpreted, to make sense. Discursive locating meanings exploring what discursive formulations of the topic, and of society more generally, are at play in the dataset, and how understanding these might add to our analysis.

Which contexts are relevant to a specific analysis will be shaped by the purpose of the research and the research questions.

What is ‘representational ethics’?

Discussing interpretation, Carla Willig noted that “the process of interpretation poses significant ethical challenges because it involves a process of transformation” (2017, p. 282). There is no simple or pure description; we always interpret from a position - or, perhaps more accurately, an aggregate of positions. This means interpretation is inevitably a political act. A “concern for representing participants” has been described as “perhaps the most significant ethical dilemma we face” as qualitative researchers (Swauger, 2011, p. 500). Ethical codes highlight our professional obligation to protect people from harm through the misuse or misrepresentation of our research (e.g. British Psychological Society, 2018). As ethical qualitative researchers, we need to think about the representational ethics and politics of our analyses. Reflexive TA research is a process of meaning-making and meaning-telling - not just summing up the things people said. Representational ethics is a question of how we tell a story that does not do harm. But there are different layers to thinking about representation, ethics, politics and harm.

What do two interconnected strands of ethics and politics of representation relate to?

• Participants in our research: how do we tell a research story that remains true to participants’ stories, without simply repeating what they say? We don’t have to tell a story that our participants would agree with, however. Telling that kind of story can feel ethically troubling, especially if you are doing more critical versions of TA (Weatherall, Gavey, & Potts, 2002), but remember these two elements:

  • First, our purpose in doing reflexive TA is to tell a story of patterned meaning, based on a range of data and participants. It is not to provide a case-study of a single person’s experience. This means there will be elements of the analyses we present that are both familiar and unfamiliar in different places, to different participants, that sometimes resonate with their own views and experiences, and sometimes don’t.

  • Second, our task is one of interpretation, of making sense of what our dataset (not just an individual participant) tells us, based on the skills and wider knowledge we bring to the process (Chamberlain, 2011; Willig, 2017). We are not simply a refining sieve that the data pass through on their way from the participants to the report. Depending on the form of reflexive TA we undertake, our analysis will be closer to, or further away from, data meanings as expressed by participants.

What ethical responsibility do researchers have?

Regardless, we have an important ethical responsibility to ensure, first, that participants understand the purpose of the research and the broad form of the likely analysis we will undertake (likely, because sometimes this changes a lot). For example, you might tell participants that you intend to report the reoccurring patterns in people’s experiences across all of the data and offering some observations about what these experiences might mean in the wider context. (Note we do not mean you need to tell the participants exactly what the analysis will be before you’ve done it; that would be impossible!) Second, that we don’t undertake interpretation in a way which patronizes or somehow belittles participants.

What does the concern for negative impacts on individual participants relate to?

The concern for negative impacts on individual participants relates to the wider communities that participants are members of, and how the stories we tell about participants might be harmful or beneficial to these wider communities.

What does the wider society consider?

The wider society. This considers the implications for members of the community or communities that participants are part of. Also, how our representational practices might work for - or against - a more socially-just society. At the broadest level, this is about taking care that our research does not reinforce existing negative stereotypes of communities or groups. This is about recognizing that in our societal contexts, different interpretations have different consequences, depending on who the participants are, and indeed who you are, as the researcher. Researcher identity is part of the issue of representation and ethics and connects to various positions of social marginality and privilege.

What are three useful concepts related to interpretation and analysis and connect more broadly to research design and research practice before, and after, analysis?

These concepts are insider/outsider positions, the Other, and intersectionality. Each is relevant and important to consider in relation to interpretation, ethics and politics, but frames or approaches these in slightly different ways.

What is the insider/outsider position?

The concept of insider and outsider is the most straightforward. We are an insider researcher if we are a member of the group we are studying, and an outsider researcher if we are not a member of the group we are studying. But identity is messy and we are often a complex mix of both insider and outsider (e.g. Hayfield & Huxley, 2015; Hellawell, 2006; Obasi, 2014; Paechter, 1998, 2013; L. T. Smith, 2013), and sometimes insider or outsider identities shift and change. It’s often assumed that being an insider researcher is simply better (inviting trust or bringing an insider advantage; Clark, Ellis Peel, & Riggs, 2010), and that it confers a more ethical position. But any position raises challenges and complexities (e.g. Wilkinson & Kitzinger, 2013). From the point of view of interpretation, it is important to recognize the position(s) we speak from in relation to insiderness and outsiderness, and engage in reflexivity in relation to understanding, interrogating and acknowledging how this has shaped our research processes (see Hellawell, 2006). But it’s not so simple.

What does a framework of insider/outsider risk?

A framework of insider/outsider risks obscuring the social reality that not all identities are created equal - and that although there is considerable variation in the privilege afforded to any individual member of a certain social group, social groupings overall vary in their level of privilege or marginalization within society. How much insider and outsider positions matter can vary along these lines of privilege and marginalization.

What is ‘The Other’ or ‘Otherness’?

Another useful concept to make sense of the ways power is inflected through societally-significant differences is ‘Otherness’ or ‘The Other’. A concept of the Other has been theorized and described within various scholarly traditions, from western philosophy to postcolonial studies to feminism. Typically, the term the Other is used to describe a person or persons who belong to a social group that is marginalized or otherwise outside the dominant norms (e.g. who, in westernized nations, is female, Indigenous/a Person of Color, working class, etc.). It captures privilege and disenfranchisement and does so through theorizing the individual within current and historical societal organization and structures of power.

What is meant by ‘representing the Other’?

Theorizing Otherness is not just about insider or outsider status, but about power. Issues relating to representing the Other capture, “whether, and how, we researchers should represent members of groups to which we do not ourselves belong - in particular, members of groups oppressed in ways we are not” (Kitzinger & Wilkinson, 1996, p. 1; see also Rice, 2009). For reflexive TA researchers focused on participant lived experiences and/or sense-making, the politics and ethics of representing the Other is an important consideration. Simply avoiding research that might involve representing the Other is not the simple solution. As the majority of academic researchers come from positions of some social privilege, just researching as insiders would perpetuate inequity, where knowledges and experiences of those socially marginalized or invisiblized remain hidden. At the same time, the notion of typically relatively privileged, White, middle-class researchers benignly ‘giving voice’ to participants from social marginalized groups - and neglecting to interrogate Whiteness and the lives of the privileged - has rightly been problematized (e.g. Clark & Braun, 2019a; Fine, 1992).

What is ‘epistemological violence’?

The concept of epistemological violence has been developed to capture the way interpretation of data from or related to, and subsequent representations of, the Other, can do harm (L. T. Smith, 2013; Teo, 2010, 2011). The closely related concept of epistemic violence evokes harm related to knowledge and discourse, connected to systems of power and oppression (see Spivak, 1988), and epistemic exclusions through who gets to be a legitimized knower (e.g. Ahmed, 2000; Ymous et al., 2020). Any research involving representing the Other needs to be undertaken with great care, without assumptions of entitlement to ask any questions and make any interpretations with any populations, and with reflexivity.

What is ‘intersectionality’?

Intersectionality theorizes discrimination and social marginalization outside of simple additive models and instead captures both experiences of social disadvantage and privilege, and theorizing these as almost two sides of the same coin. It particularly captures the ways in which different forms of privilege and disadvantage intersect. For example, different systems of marginality and privilege - around race and gender say - are not separable, instead they inextricably shape the meaning and experience of each other.

In terms of interpretation within reflexive TA, what might intersectionality offer or mandate?

At the most basic, intersectionality is a way of engaging with participant data that is nuanced and contextualized and incorporates privilege and marginalization into our sense-making. It recognizes the locatedness and partiality of what we claim. Coming back to questions of ethics, politics and practice, it requires reflexivity to recognize our interpretative work is always going to be partial and imperfect and reflects our situatedness.

Why is it important to consider the language you use?

Part of the wider context for research is language, and the way it can often unintentionally marginalize. Therefore, it is important to consider the language you use, and how it might convey stigma, or might marginalize. A good starting guideline is to work with the language people themselves use.