Research Guide: Coding and Thematic Analysis

Once you have collected qualitative data: through interviews, ethnographic observation, or cultural artifacts: you need a systematic method for transforming it into findings. This guide follows the Braun & Clarke framework for thematic analysis, one of the most widely cited and practically useful approaches in qualitative research.

The six steps are:

• ·Familiarize yourself with the data
• ·Create initial codes
• ·Decide what to code, and add new codes
• ·Collate codes with supporting data
• ·Group codes into themes
• ·Evaluate and revise your themes

Step 1: Familiarize Yourself with the Data

Read through your entire dataset before coding anything. Take notes on initial impressions, recurring ideas, and things that surprise you. This read-through is not passive: you are actively building a mental map of the material. If you have transcripts, read them in full. Do not start coding until you have a feel for the whole.

Step 2: Create Initial Codes

From your initial read-through, identify codes: short labels that capture what a particular piece of data is about. Do certain subjects in your data share a complaint? Do they describe a process in similar ways? These become candidate codes. At this stage, cast a wide net. You can always merge or discard codes later; the goal is to avoid missing important patterns by coding too narrowly.

Step 3: Begin Coding, and Add New Codes

Start labeling your data with codes. As you work through the material, you will encounter passages that do not fit your existing codes: add new ones. This is expected and healthy. Coding is iterative, not mechanical. Keep a code register that tracks what each code means, so you apply it consistently across the dataset.

Step 4: Collate Codes with Supporting Data

For each code, collect all the data passages that have been labeled with it. Evaluate whether you have enough data to support the code as meaningful. Some codes will have rich evidence; others will have one or two instances and may need to be merged into a broader code or dropped. The goal is codes that are both grounded in evidence and analytically meaningful.

Step 5: Group Codes into Themes

Group related codes into themes. A theme should be nuanced and complex: not just a topic, but a claim about the data. For example, "overwork" is a topic; "a culture of overwork sustained by implicit social norms" is a theme. Themes should capture something meaningful about the pattern in the data, not just describe what is there.

Step 6: Evaluate and Revise Your Themes

Thematic analysis is iterative. Read back through your data with your themes in hand, and ask: do these themes tell the whole story? Are there data points that do not fit? Are any themes too broad or too narrow? Revise until your themes collectively account for the most important patterns in the data.

Writing Your Paper

Your write-up should not just list themes: it should build a narrative. Use your data to construct a causal or interpretive account of what is happening and why. Quote liberally from your data to show the reader that your themes are grounded, not invented. The best qualitative papers make the reader feel that the interpretation is not only plausible but inevitable given the evidence.

To preserve the anonymity of interview subjects, you may want to create amalgamated quotes: combining parts of several different quotes to form a new composite quote, and changing non-important identifying details. This preserves the meaning while protecting the individual. Label these as paraphrased or composite when you use them.

Qualitative Coding Software

If you conduct qualitative research regularly, dedicated coding software will significantly improve your workflow. Common tools include NVivo, Atlas.ti, and the open-source MAXQDA. These allow you to attach codes to passages, visualize code relationships, and search across your dataset by code: capabilities that become essential once your dataset exceeds a few dozen pages.