Author: Astrid Coxon

Common thematic analysis pitfalls – and how to avoid them

If you’re a qualitative researcher, you will be familiar with Braun and Clarke’s reflexive thematic analysis (RTA). Originally outlined in their 2006 methodological paper “Using thematic analysis in psychology”, it is probably the most common ‘first foray’ into qualitative research for psychologists. Since 2006, Braun and Clarke have refined and reworked their 6-step method, producing papers, books, YouTube videos (from both Clarke and Braun) and an entire website dedicated to supporting the conduct of quality RTA.

Despite this clear and helpful guidance, we regularly come across a handful of common errors in the conduct and write-up of RTA studies here at PGPR.

1. Themes do not emerge!

If you’ve ever used the term “the themes that emerged…” or similar in your RTA study, you’re in good company – it’s an error so commonly commented on that it’s spawned its own meme.

Pin badge (now discontinued, sadly) designed by Dr Heidi R Green, “Science On A Postcard”

But why does this seemingly innocent phrase cause so much uproar in the RTA community? It comes down to its positivist implications: saying that the themes emerged, were identified, or even were found in the data, suggests that, if we look hard enough, we can find objective themes which truly exist, regardless of the researcher’s analytic process.

Themes, in an RTA sense, are not real – they are constructed through the analytic process. Braun and Clarke, true to form, say it best: “In reflexive TA, themes are generated, created or constructed (for example), they are not identified, found or discovered, and they definitely don’t just ‘emerge’ from data like a fully-grown Venus arising from the sea and arriving at the shore in Botticelli’s famous painting” (Braun & Clarke, 2022). Is this semantic pedantry? Some might argue that it is, but this edict has become the shibboleth by which seasoned RTAers recognise whether or not their students have familiarised themselves with Braun and Clarke beyond 2006!

2. So where do themes come from..?

They are developed! You won’t know what your themes will be until you’ve engaged in the analytic process, so none of this “we expect the themes to be…” nonsense in your introduction, please. Depending on your research aims and specific approach (inductive, deductive – RTA is delightfully flexible), you may have specific areas of the data you want to focus on, but try to also allow for unexpected insights. And definitely steer clear of topic summaries as theme names – a theme named “how patients use online patient portals” doesn’t tell an interpretative story (and probably doesn’t go much beyond the initially stated research aim). But theme names such as “taking ownership of my healthcare” communicates a core message – a message that cannot be conceptualized prior to analysis.

To put it simply: themes come from a combination of the data, the analytic process, the sociocultural context in which the data was collected and analysed, and the specific context that you, as the analyst, bring to the analytic process. If, for example, Johanna and I were to analyse the same set of transcripts, we would very likely construct a different set of themes. If we were to analyse the data together, a third distinct set of themes would probably be developed. And that is because…

3. You are a tool of the analytic process

…not a contaminant to be contained. Again, this comes down to a question of philosophical frameworks – RTA is not constructed on a bedrock of positivist paradigms.

Photo by Lachlan Donald on Unsplash

Unlike in much quantitative research, the aim of RTA is not to control for or eliminate researcher bias – the reflexive aspect of RTA acknowledges that the researcher plays an active role in developing themes. As the analyst, your position, context, and yes, bias, bring something unique to the analytic process – so own it, don’t deny it. Engage reflexively throughout the analytic process to explain how and why you developed the themes as you did. This can be done as a group process, with multiple coders or analysts, but not in an effort to “control for bias”. Inter-rater reliability has no place here.

4. Leave the concept of saturation to the grounded theorists

In a similar vein to “inter-rater reliability”, we all too often read references to saturation in RTA studies. I wonder if this is in an effort to appease positivist editors or supervisors, or a desperate bid to cling to something concrete to guide sample sizes. You might be showing your positivist hand if you insist that “interviews will be conducted until data saturation is reached”. The concept of saturation is that there comes a point after which “no new information” is produced by further interviews (or focus groups, or whatever mode of data collection is being undertaken). This makes sense in a grounded theory study, where the analyst is interested in finding the solution to a specific problem and will use theoretical sampling to simultaneously collect data as they analyse, thus allowing them to narrow down further and further towards an understanding of their new theory.

However, in RTA, Braun and Clarke have been pretty clear on the topic: “…when it comes to reflexive TA, data saturation is not a particularly useful, or indeed theoretically coherent, concept” (Braun & Clarke, 2019). The more data you collect, the more possibilities for interpretation arise – the themes you develop after 10 interviews may look very different from the themes you might have developed if you conducted 20 interviews. You’ll just have to make a pragmatic, interpretative judgement (and possibly be guided by Braun & Clarke’s Thematic Analysis FAQ). Try to be comfortable with the uncertainty.

5. There isn’t just one approach to Thematic Analysis

…although you’d be forgiven for thinking so. Braun and Clarke’s 2006 methodological paper “Using thematic analysis in psychology” has several hundred thousand citations now, and their approach is used across a broad range of disciplines. However, theirs is not the only way, even if it is my favourite – clearly, I’m not alone in this! There are other modes of thematic analysis, some of which have more neo-positivist roots. Thematic analysis’ multiplicity is part of the reason why, post-2006, Braun and Clarke chose to differentiate by naming their method reflexive thematic analysis.

If you are undertaking an RTA study, be explicit about this by naming it as such – and try to cite sources beyond that 2006 paper, to demonstrate your reflexive chops. And don’t forget: be reflexive!

If you want help ensuring that your RTA is working as well as it can, just get in touch with us via the box below. We’ll be happy to help.

Five more common problems for postgrad students and how PGPR can help

In a previous post, we covered some of the common problems we support postgrad qualitative students with, to help you to tighten up your writing. In this post, we’re covering a few more.

1.      Overuse of conjunctive adverbs

Conjunctive adverbs are words or short phrases used to transition from one clause to another. You’ve used them all your life, and you might not have known they had a term – words like “however”, “moreover”, “in conclusion” and “additionally” are all conjunctive adverbs, and can be useful for creating narrative direction or emphasis. However, over-use can make your writing clunky, cluttered and overly formal. Moreover, starting every sentence with a conjunctive adverb can be a jarring reading experience, detracting from the content of your writing.

Take those two previous sentences as an example; arguably (there’s another one), the “however” is a useful signifier of a change in direction (from the “usefulness of conjunctive adverbs” to “writer beware”). The following sentence didn’t need to start with “moreover” – it stands up fine on its own. I could have started the sentence before this one with another conjunctive adverb, such as “Arguably, the following sentence…”, but did it need it? Arguably, it didn’t. Next time you review your writing, try highlighting every time you use a conjunctive adverb – can some of them bear pruning?

Conjucntive adverbs can be useful linking devices but use them sparingly!

Photo by Zulmaury Saavedra on Unsplash

2.      Badly positioned ordinal and sequencing adverbs

In academic writing, we often use ordinal or sequencing adverbs to structure points in our argument – firstly, secondly, lastly, finally. They can help your reader to identify discrete points within a broader claim, or guide them through the order of your ideas. But use them wisely!

  • If you’re going to use ordinal adverbs, use them clearly and consistently – for example, if you tell your reader that there are three main criticisms of a theory, don’t start with “firstly” and then fail to indicate “secondly” and “thirdly”, as they may not be able to clearly identify where one criticism ends and another begins.
  • If you start a clause with “finally” or “lastly”, it really should be the final or last clause of that argument – don’t use “finally” in the middle of a paragraph and then go on to make several other points related to the argument.

3.      Making analytical claims that aren’t properly supported by data

A common refrain from markers and reviewers everywhere is “citation needed”, usually referring to claims made that aren’t supported by published peer-reviewed literature. If you’re a qualitative researcher, you also need to ensure that analytical claims are clearly supported by your data. When you make statements about the phenomenon you’re researching, is it evidenced in the data you’re presenting, or does it just “feel right”? Make sure that your reader can follow your reasoning by presenting relevant quotations from your participants and explaining your analytic process.

4.      Hedging

In writing, “hedging” is using language that indicates uncertainty. You might be familiar with the term “hedging your bets”, to protect yourself from potential losses by placing bets on multiple possible outcomes.

Sometimes, hedging can be appropriate, such as when you are exploring multiple possible interpretations of findings (e.g., “this may indicate…”). However, don’t hedge everything – overuse of hedging can make your writing less direct or certain, and this isn’t appropriate when you are stating things that are certain or unambiguous.

For example, instead of writing “In this chapter, I hope to demonstrate…”, simply state “In this chapter, I will demonstrate…” – not only will this make you sound less apologetic, but it also improves clarity and conciseness, reducing the burden on your reader (who otherwise has to cut through the waffles of mights and maybes!)

Don’t let your reader get lost in the hedge maze

Photo by Field Cottage on Unsplash

5.      Using too many summaries and repetition

Clear metalinguistics – explaining “this is what I’m going to cover” and “in summary, this is what I’ve covered” – can be useful to guide your reader through your writing, but don’t overdo it. If you have just explained something in good detail, don’t start the subsequent subsection with a summary of what you’ve just covered. Not only is this repetitive, but the extra clutter it creates potentially makes it harder for your reader to follow the narrative thread of your writing.

 

Need help pruning your hedging or ordering your adverbs? We can help you to make your writing shine – just contact us via the box below!

Using anthropomorphism in academic writing

It’s not uncommon to read “this study concludes…” or “these findings suggest…” in academic writing. I would hazard a guess that nearly every researcher has used similar phrases at some point in their career. You may well have used them yourself and thought nothing of it. However, these ubiquitous and seemingly innocuous phrases cause quite a stir in certain circles.

They are just two examples of anthropomorphism: incidences where the author attributes human characteristics to inanimate or nonhuman objects or concepts. Under APA 6 guidelines, academics would be chided for implying that their data had the power of speech – how can data say anything? Anthropomorphism was to be avoided at all costs.


It can be tempting to give human qualities to anything and everything!

Photo by Brett Jordan on Unsplash

And logically, rationally, this makes sense. Data, findings, “the study” – all these things lack agency. Research doesn’t think, explore, or conclude; researchers do! Active verbs belong with people, not abstract concepts or inanimate objects.

If you are only now learning of these particular guidelines, you may be groaning, rolling your eyes, and wondering if you have to go back through your entire dissertation to awkwardly juggle clauses and remove agency from “the study”. If you’re writing in the first-person, this is fairly straightforward. What was “this chapter describes…” becomes “in this chapter, I describe…”.

However, whilst first-person writing is an increasingly popular choice for qualitative researchers, not all supervisors, examiners and journal editors will accept this format…a topic for a whole other blog. De-anthropomorphising sentences in the third person is possible, but it can lead to some verbose and awkward phrasing. For example:

This research explores the impact of dog ownership on people’s happiness.

Could become the slightly wordier:

In this study, the researchers explored the impact of dog ownership on people’s happiness.

Or the clumsy, back-ended passive-voice version:

          In this research, the impact of dog ownership on people’s happiness was explored.


Every dog has its data
Photo by Jamie Street on Unsplash

Perhaps you’re wondering what all the fuss is about. Surely this is pedantic in the extreme; we know that it’s the researchers, not the research, who are doing the exploring, describing and suggesting! Anthropomorphism is a useful and engaging stylistic shorthand which can facilitate clarity and conciseness, rather than create confusion. It’s unlikely that we will read “the study suggests…” and go away thinking “gosh, that study has acquired the power of speech! And isn’t it remarkably articulate?”

And to an extent, it seems the APA style guide finally agrees (or should that be “the authors of the APA style guide agree” …?)… to an extent. Under APA 7 guidelines, the rules around anthropomorphism have been relaxed, meaning your study can now be allowed to speak. Only a little, though, and only under strict circumstances. There are limits to the active verbs which can be ascribed to non-human entities. For example, it’s now fine to state:

The data suggests…

The findings indicate…

The table presents…

Some phrases which might be acceptable (but use with caution):

This chapter describes…

The questionnaire assessed…

But definitely steer clear of:

The study concluded… (this suggests that there is one objective conclusion to be drawn from the data).

The literature claims… (this suggests that from a whole body of literature on a given topic, one unified view can be drawn).

Whilst there is arguably room for clarification around these new guidelines, I wouldn’t hold out hope for a comprehensive list of suitable verb-subject pairings. A good rule of thumb may be to question why you’re employing anthropomorphism – does it clarify your writing (perhaps by avoiding long, awkward sentence structuring), or does it muddy your intended meaning? If you spot anthropomorphism in your writing, is there another way this could be phrased? While there is no longer a blanket ban on anthropomorphism in APA, it’s generally good practice to reflect on your writing, and what you may be implying by employing this device.

You can read more about the changes to the rules here.


Sometimes it’s a case of moving the pieces around.
Photo by Olav Ahrens Røtne on Unsplash

If this all feels overwhelming, or confusing, or you’re not sure where to start, then there’s always help available. Get in touch with us about your writing concerns! Whether it’s specifically about anthropomorphism or more broadly about APA style, we’re happy to provide expert guidance and friendly support. Just fill in the box below to contact us, and we’ll get back to you ASAP.