Thematic Analysis: A Complete Step-by-Step Guide for Qualitative Research (2026)
Thematic analysis is the most widely used method for analysing qualitative data in social sciences, education, psychology, health research, and beyond. It is flexible enough to work across theoretical frameworks, accessible enough for first-time researchers, and rigorous enough to meet the standards of peer-reviewed publication. Yet for many students conducting their first qualitative study, the six phases of thematic analysis can feel opaque — you have the data, you have Braun and Clarke’s paper bookmarked, but you are not sure what “coding” actually looks like in practice or how to move from a list of codes to a coherent set of themes.
This guide walks through every phase of thematic analysis in concrete terms, with worked examples, common mistakes to avoid, and guidance on how to write up your themes in a thesis or dissertation.
What Is Thematic Analysis?
Thematic analysis (TA) is a method for identifying, analysing, and reporting patterns (themes) within qualitative data. First formalised by Virginia Braun and Victoria Clarke in their seminal 2006 paper in Qualitative Research in Psychology, TA has since become the de facto default for qualitative analysis across many disciplines — particularly because it is not tied to a single theoretical framework the way grounded theory or phenomenology are.
TA works with any qualitative data type: interview transcripts, focus group recordings, open survey responses, diary entries, social media posts, policy documents, or archival texts. The output is a set of themes — not categories, not content codes, but rich, interpretive accounts of what the data reveals about your research question.
A critical distinction: thematic analysis is not simply identifying what participants mentioned most often. A theme in TA captures something important about the data in relation to your research question, whether or not it appears frequently. As Scribbr’s thematic analysis guide emphasises: prevalence is not the same as significance.
Inductive vs Deductive Approaches
Inductive (bottom-up): Themes emerge from the data without a pre-existing framework. You approach the data with open curiosity and allow the patterns to guide your analysis. Best used when exploring a new phenomenon or when you want your analysis to be grounded in participants’ own meanings.
Deductive (top-down): You begin with an existing theory, framework, or research question and code the data in relation to it. Best used when testing or extending a theoretical framework, or when answering a specific pre-defined research question.
Most student research uses an inductive or mixed approach. Be explicit in your methodology chapter about which approach you are taking and why — this is a core methodological decision that affects how you conduct every phase.
Phase 1: Familiarisation with the Data
Before any analysis begins, you must know your data deeply. This means reading every transcript, watching every video recording, reading every survey response — multiple times. Take notes as you go: initial impressions, recurring words, surprising statements, and potential areas of interest. This is not yet coding; it is orientation.
For interview or focus group data, transcribe verbatim — including pauses, laughter, and interruptions. Verbatim transcription is important because meaning is embedded in how people speak, not just what they say. If you are using audio-to-text software, always review and correct the output before analysis.
Phase 2: Generating Initial Codes
Coding is the process of identifying meaningful segments of data and labelling them. A code is a short phrase that captures what is interesting about a data extract in relation to your research question.
Worked example:
Data extract (interview with a first-year university student): “I found it really hard at first because everyone seemed to already know each other, and I felt like I was the only one who didn’t know how things worked. It took me about six weeks before I stopped feeling like a fraud.”
Possible codes: social isolation in transition / perceived exclusion from established peer networks / imposter syndrome / temporal adjustment (6-week threshold) / uncertainty about institutional norms
Annotation: Notice that multiple codes can apply to a single extract. Initial coding is generative — capture everything that might be relevant. You will reduce and organise in later phases.
Code every data extract systematically. In manual analysis, use coloured highlighters or margin notes. In software, use the tagging function. At the end of Phase 2, you typically have dozens to hundreds of codes across your dataset.
Phase 3: Searching for Themes
Now you move from the level of individual codes to broader patterns. Group your codes into potential themes — clusters of codes that capture something coherent and significant about the data.
A useful technique is to print all your codes on index cards or Post-it notes and physically sort them into groups. Alternatively, use a spreadsheet: list all codes in one column, and move them into theme groupings in adjacent columns. This visual organisation helps you see what belongs together and what does not.
At this stage, themes are provisional — you are not committing to a final structure. You might have 8–15 provisional themes with some very small, some very large, and some overlapping.
Phase 4: Reviewing Themes
This phase has two levels. First, re-read all the coded extracts within each provisional theme and check whether they cohere — do they really tell the same story? If a theme contains contradictory material, it may need to be split. If two themes contain very similar material, they may need to be merged.
Second, read the entire dataset again through the lens of your provisional themes and check that your themes are a fair and full account of the data as a whole. Are there important patterns in the data that none of your themes captures? Are there extracts that do not fit anywhere — and if so, do they represent a theme you have missed?
End Phase 4 with a thematic map: a visual representation of your themes, sub-themes, and how they relate to each other and to your research question.
Phase 5: Defining and Naming Themes
For each theme, write a clear definition: what is the central organising concept? What does this theme capture that the others do not? What is its scope — what counts as belonging to this theme and what does not?
Theme names should be informative, not generic. “Challenges” is a poor theme name — it could apply to almost any qualitative study. “Navigating institutional opacity in the first year” is specific, interpretive, and communicates the substance of what the theme captures.
Phase 6: Writing Up
The write-up of thematic analysis is not a list of themes followed by quotes. It is an analytical narrative that weaves evidence and interpretation together. The standard structure for each theme is: (1) introduce the theme and its significance, (2) present data extracts as evidence, (3) analyse what the extracts reveal — this is the analytical work, not a summary of what the participant said, (4) link back to the literature.
Strong write-up pattern:
“A prominent theme across interviews was the experience of institutional opacity — the sense that the rules governing academic life were both consequential and unknowable. Participants consistently described encountering unspoken conventions they had no way of anticipating. As one student explained: ‘Nobody tells you how to actually talk to your supervisor — like, whether you email or just knock, or how often you’re supposed to send them things. I got it wrong a few times and felt terrible’ (P4, Female, Year 1). This finding resonates with Reay et al.’s (2009) concept of ‘academic habitus’ — the embodied sense of belonging in higher education that first-generation students disproportionately lack on entry.”
Annotation: Notice the structure — theme described, extract quoted with participant identifier, analytical interpretation (institutional opacity as a concept), and connection to relevant literature. This is what thematic analysis write-up looks like.
Software for Thematic Analysis
| Tool | Cost | Best For |
|---|---|---|
| NVivo | Paid (free student licence at many universities) | Large datasets; mixed methods; complex coding |
| ATLAS.ti | Paid (student discount available) | Visual coding; multimedia data |
| MAXQDA | Paid (student licence available) | Mixed methods; strong visualisation |
| Delve | Free (basic) / Low-cost pro | Beginners; simple thematic analysis |
| Excel / Google Sheets | Free | Small datasets; students without software access |
See our guide on qualitative research methods for a broader overview of qualitative approaches. For the methodology chapter of your thesis, our research methodology chapter guide covers how to write up your TA approach with appropriate epistemological justification.
Common Mistakes in Thematic Analysis
- Treating themes as categories or content codes — TA themes are interpretive claims about what the data means, not frequency counts of topics mentioned.
- Jumping to themes without systematic coding — Phase 2 cannot be skipped. Themes built without thorough coding are impressionistic, not analytical.
- Not anchoring themes to data — every theme needs supporting data extracts. “The theme of identity emerged” is an unsupported assertion without quotes.
- Generic theme names — “Challenges,” “Benefits,” “Experiences” are not informative theme names. Be specific about what the challenge/benefit/experience IS.
- Failure to reflect positionality — in qualitative research, your background and assumptions influence what you see in the data. A brief reflexivity statement in your methodology demonstrates methodological maturity.
Frequently Asked Questions
How many themes should thematic analysis produce?
Most thematic analyses produce 3–8 main themes. Fewer than three often suggests over-generalisation; more than eight can become unwieldy and may signal that codes, not themes, are being reported. Sub-themes can add granularity within main themes without expanding the overall theme count. The right number is whatever accurately and fully accounts for what is meaningful in your data in relation to your research question.
What is the difference between thematic analysis and content analysis?
Content analysis is primarily concerned with what is present in the data — it often involves counting how frequently certain words or topics appear. Thematic analysis is concerned with what the data means — it identifies interpretive patterns that capture something significant about the phenomenon. Content analysis can be quantitative; thematic analysis is inherently qualitative and interpretive.
How do you ensure reliability in thematic analysis?
Qualitative research does not use reliability in the psychometric sense, but demonstrates rigour through: thorough documentation of analytical decisions (an audit trail), member checking (sharing findings with participants for validation), reflexivity (acknowledging your positionality), peer debriefing (discussing themes with a colleague or supervisor), and thick description (providing enough detail for readers to assess the transferability of findings).
Can thematic analysis be used with survey data?
Yes. Thematic analysis is commonly applied to open-ended survey responses. The same six-phase process applies. Note that survey responses tend to be shorter and less rich than interview data, which may limit the depth of analysis possible — but for research questions where breadth of perspectives matters more than depth, survey-based TA is entirely appropriate.
Do I need software for thematic analysis?
No. Thematic analysis can be conducted entirely manually using printed transcripts, highlighters, and index cards — or using Word, Excel, or Google Docs. Software like NVivo, ATLAS.ti, or Delve makes the process easier and more systematic with larger datasets, but the analytical thinking is done by the researcher, not the software. Never describe your methodology as “I used NVivo to analyse the data” — the tool does not do the analysis.
Write Up Your Qualitative Findings with Tesify
Tesify helps you structure and refine qualitative write-ups — from methodology justification to theme narratives — with academic precision and clarity.
Leave a Reply