How to conduct Thematic Analysis for SoTL Research

As researchers in the Scholarship of Teaching & Learning, there will invariably come a time when qualitative analysis will need to be conducted. For many researchers with a more quantitative background, or for those just starting in their academic career, “qualitative analysis” may be very new. This problem is exacerbated by the different forms of qualitative analysis offered. Anyone may pick up a book on interpretative phenomenological analysis, read the title and requirements of IPA, and then decide “maybe qualitative analysis isn’t for me”. However, there is an option available for researchers new to the field that is easy to understand, has a clear set of guidelines, and can be conducted on near any form of data. This is called “Thematic Analysis”.

Thematic analysis is a method of qualitative analysis that serves to identify, analyze, and report patterns (themes) within the data. It is useful as generally the first form of qualitative analysis that researchers learn. For anyone who knows programming, think of thematic analysis as the Python (or HTML) of qualitative research methods. For those with a less tech knowledge, thematic analysis is to qualitative analysis as cooking an egg is to becoming a chef. The next sections of this blog will discuss some pros and cons of this method, and then go straight into the 6 steps of conducting thematic analysis. It will end with some additional sources should you be interested in learning more.

Pros & Cons of Thematic Analysis


1. Flexibility. Thematic analysis can be done on almost anything. Whether the data set you have is interviews, documents, news items, or speeches, the process is still the same. This is incredibly useful for SoTL researchers as our data can and does come from anywhere. This flexibility also applies to philosophical underpin. In qualitative analysis, researchers commonly state what philosophy (constructivism, essentialism, etc.) they take prior to conducting research. Certain forms of qualitative analysis are situated within a philosophy that the researcher may not subscribe to. Thankfully, the structure of thematic analysis allows for different philosophical paradigm to be used.

2. Learning curve. Thematic analysis is also easy to learn. This is useful given the diverse education level and knowledge background of SoTL researchers. Academics who study SoTL come from different fields of academia. these different skill levels demand a form of analysis that lets them jump in without too much formal training. Conducting another form of analysis, like grounded theory or discourse analysis can be quite daunting and turn others off from qualitative analysis.


1. Reliability. The flexibility associated with thematic analysis is actually a double-edged sword. Because thematic analysis is less structured, individuals may be stuck with too many sources and no real way to synthesize it all appropriately. Trying to conduct thematic analysis with multiple data types (interviews, documents) means individuals may come to the issue of which item or theme to focus on. Furthermore, this flexibility means individuals conducting thematic analysis with the same data may come to wildly different results. This is a broader issue with qualitative analysis in general, but the flexibility of thematic analysis exacerbates these concerns.

2. Prestige. Thematic analysis, unlike other forms of qualitative analysis is not tied to a philosophical paradigm. This also comes at the cost of being less popular – what Braun & Clarke describes as “kudos” – than other forms of analysis such as grounded theory or discourse analysis. Thus, other researchers may not particularly know of “thematic analysis” even if they’ve inadvertently done it themselves. Furthermore, the prestige of a form of analysis can sometimes be used to lend credence to the work conducted. However, this is an outcome of ‘methodolatry’ – a commitment to method rather than the topic or research question (Holloway & Todres, 2003). The hope is that as different forms of qualitative analysis are better understood, individuals will not succumb to this heuristic.

6 Steps to conducting thematic analysis

Presented here are the 6 steps to conducting thematic analysis. Note however that these are merely guidelines. The key to conducting good qualitative analysis  is not following a step-by-step guide but maintaining an analytical mindset while doing so. Having a strong analytical mindset throughout the process will lead to the best results, whether or not you follow the process exactly.

  1. Familiarization with data

Familiarization requires immersion into the data. Repeated readings are recommended. The researcher must actively search for meaning, patterns, and other possible items of interest. Note taking, in the form of a notepad, sticky notes, or journal, is also necessary. Braun & Clarke (2006) recommend that, at this stage, the researcher should try and generate an initial list of ideas about what is in the data and what is interesting.

  1. Generation of initial codes

Codes identify data features (e.g. descriptors, patterns) of interest to the research question. Researchers must systematically work through the entire data set, giving full & equal attention to each data item and extracting codes. Advice include code as many themes as possible, code inclusively (include some context), and allow data items to be coded more than once when relevant.

  1. Searching for themes

This phase requires the researcher to summarize the different codes into overarching themes, connecting the themes with the purpose to create a thematic map of the data. A thematic map contains both themes and subthemes with an outlined depiction of each themes’ relations to each other.

  1. Reviewing themes

This phase involves a refinement of the themes over a two-part process. In part 1, researchers must review the data extracts, reading for the proposed themes. If the proposed themes are ill-fitting, an editing of the thematic map may be necessary. In part 2, researchers must consider the validity of individual themes in relation to the data set. This includes determining whether themes “work” within the data set and whether additional codes are found with the themes in mind.

  1. Defining and naming themes

This phase involves creating the final thematic map. This map requires an explanation of each theme and how each theme addresses the research question. Sub themes are also included and explanations are required for these as well. The end purpose is a clear definition of what the themes are and how they differentiate.

  1. Producing the report

The final phase involves producing the report of the analysis process. Recommendations include ensuring the report is concise, coherent, logical, non-repetitive, and interesting. Evidence in the form of data extracts are generally used to further elucidate points.


The information in these 6 steps provide the key points for conducting thematic analysis. For a more thorough explanation of the method, see Braun & Clarke (2006). More detailed information about qualitative analysis in general is available through the book Successful Qualitative Research: A Practical Guide for Beginners (Braun & Clarke, 2001).


Braun, V., & Clarke, V. (2001). Successful qualitative research: A practical guide for beginners.       Thousand Oaks, CA: Sage Publications

Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77-101.

Holloway, I., & Todres, L. (2003). The status of method: flexibility, consistency and coherence. Qualitative Research, 3(3), 345-357.

2 thoughts on “How to conduct Thematic Analysis for SoTL Research

  1. galiciablackman

    I first encountered Braun & Clarke a few months ago when I was trying to determine how to interpret some informal interviews. I think there is a natural organic appeal to this kind of analysis that makes the pros very attractive. But as you rightly pointed out in the cons, it is also easy to get carried away with the volume of content which can be generated. I like what phenomenology offers to be able to cope with data spinning out of control. It would be more along the lines of commitment to the topic and not only method, as your reference to Holloway &Torres indicates. With any research approach, knowing the limitations upfront is key to doing good work, so laying out the pros and cons like this is a smart start for anyone.

  2. cpostrow

    Great post! I think you already implied this, but I’d also add that you don’t necessarily need to (or should) use the same codes for all of your data. For example, in my recent narrative inquiry research I was careful to generate codes for each case/person I interviewed and analyze them separately before looking across the cases and comparing them to each other. In some forms of qualitative work you want to be careful not to shoehorn the data into a set of codes that may not apply to everything/everyone.

Leave a Reply