To apply thematic analysis in a qualitative psychology thesis: familiarise yourself with the full dataset, generate initial codes systematically, search for candidate themes across codes, review those themes against the coded extracts and the whole dataset, define and name each theme precisely, then write the analysis with data extracts as evidence. This six-phase sequence, set out by Braun and Clarke (2006), is one of the most widely used approaches to thematic analysis in psychology.
Step 1: Transcribe and Familiarise Yourself With the Full Dataset
Transcribe every interview or focus group recording in full (verbatim, including filler words if your design calls for that level of detail), then read the complete set of transcripts at least twice before coding anything. Example sentence for your methodology chapter: “All interviews were transcribed verbatim and read in full twice before initial coding began, following Braun and Clarke’s (2006) familiarisation phase.” Resist the urge to start coding on your first read — familiarisation is meant to give you a sense of the dataset as a whole before you start breaking it into pieces.
Step 2: Generate Initial Codes Systematically
Work through the dataset line by line (or extract by extract), attaching a short code label to any segment of text relevant to your research question. Codes at this stage are close to the data’s own language, not yet abstracted into themes. Illustrative example sentence: “Initial codes were generated inductively, working systematically through each transcript, with 47 initial codes produced across the 12 interviews.” Whether you code inductively (codes emerging from the data itself) or deductively (codes drawn from an existing framework or your research questions) should be stated explicitly and applied consistently — mixing the two without saying so is a common source of examiner questions at viva.
Step 3: Search for Candidate Themes Across Your Codes
Group related codes into candidate themes — patterns of shared meaning across multiple codes and, ideally, multiple participants. A theme is not simply a topic that came up; it is an organised pattern that says something specific about your research question. Example sentence: “Codes relating to loss of routine, disrupted sleep, and reduced social contact were grouped into the candidate theme ‘disruption to daily structure.’” At this stage, sketch your candidate themes as a simple table or diagram linking each theme to the codes that feed into it, so your reasoning is traceable later.
Step 4: Review Themes Against Extracts and the Whole Dataset
Review each candidate theme twice: first against all the coded extracts assigned to it (do they genuinely fit together, or does the theme need splitting?), then against the entire dataset (does the theme hold up across the full set of transcripts, or only in one or two interviews?). Example sentence: “Each candidate theme was reviewed against its coded extracts and then against the complete dataset; the initial theme ‘coping strategies’ was split into two distinct themes, ‘active coping’ and ‘avoidant coping,’ after this review revealed they represented meaningfully different patterns.” This two-level review is the step most first-draft theses skip, and it is usually the first thing a supervisor asks about.
Step 5: Define and Name Each Theme Precisely
For every theme that survives review, write a clear definition capturing exactly what the theme is about, what it is not about, and how it relates to your research question, then choose a concise, informative name — not a generic label like “Theme 1” or a vague phrase that could describe several different things. Example sentence: “The theme ‘disruption to daily structure’ was defined as participants’ accounts of how the loss of an external routine altered sleep, eating, and social-contact patterns, distinct from the separate theme of emotional response to that disruption.”
Step 6: Produce the Written Analysis With Data as Evidence
Write your findings chapter theme by theme, using direct participant quotations as evidence for each analytic claim you make, and connect your analytic narrative back to your research question throughout rather than only in the discussion chapter. Illustrative example sentence: “As one participant described, ‘I stopped keeping any kind of schedule, it all just blurred together’ (Participant 4), illustrating how the loss of routine was experienced as a loss of temporal structure rather than simply reduced activity.” Every theme should appear with more than one supporting extract, ideally from different participants, not a single quotation standing in for the whole pattern.
Worked Illustrative Extract-to-Theme Table
The table below shows an illustrative, condensed path from one raw data extract through to a named theme, to make Steps 2 through 5 concrete in one view. Build a version of this table for your own dataset as a working document — it does not usually appear in the thesis itself, but it is exactly the kind of evidence a supervisor or examiner may ask to see if they question how you arrived at a theme.
| Stage | Illustrative content |
|---|---|
| Raw extract | “I stopped keeping any kind of schedule, it all just blurred together.” |
| Initial code | Loss of daily routine |
| Related codes grouped | Disrupted sleep timing; reduced structured activity; blurred weekday/weekend distinction |
| Candidate theme | Disruption to daily structure |
| Final theme definition | Participants’ accounts of how the loss of an external routine altered sleep, activity, and time-perception patterns |
Writing a Reflexivity Statement
Many psychology qualitative theses, particularly those following a reflexive thematic analysis approach, expect a short reflexivity statement in the methodology chapter: a paragraph naming your own position in relation to the topic and participants (for example, your own familiarity with the population studied, or assumptions you brought into the analysis) and how you attempted to remain aware of that position while coding. Example sentence: “As a final-year psychology student with no prior clinical experience with this population, I approached the data without a pre-existing clinical framework, and kept a reflexive journal throughout coding to note where my own assumptions might be shaping theme development.” This is not a confession of bias to apologise for — it is a standard, expected component of the method that shows the reader you engaged with the data as an active interpreter rather than a neutral counter of themes.
Semantic vs Latent Themes: A Choice You Must State
Thematic analysis can identify themes at the semantic level (what participants explicitly said) or the latent level (underlying ideas, assumptions, or meanings beneath the explicit content). State which level your analysis operates at, and apply it consistently — a study that reads some themes semantically and others latently, without explaining why, invites an examiner to ask how the two were distinguished.
A Note on Coding Reliability and Reflexive Thematic Analysis
Older guidance on thematic analysis sometimes calls for a second coder and an inter-rater reliability statistic (such as Cohen’s kappa) as a marker of rigour. Braun and Clarke’s later work on reflexive thematic analysis (developed further after their original 2006 paper) explicitly argues against this practice for a reflexive approach, on the grounds that a single, reflexively engaged researcher’s coding is a feature of the method, not a weakness to be corrected by a second coder. Whichever position you take, state it explicitly in your methodology chapter and justify it against the specific version of thematic analysis you are citing — do not claim adherence to reflexive thematic analysis while also reporting an inter-rater reliability score, since the two are commonly read as inconsistent.
Common Mistakes When Applying Thematic Analysis
- Treating “themes” as a list of interview questions or topics discussed, rather than analytic patterns you constructed from the codes.
- Skipping the theme-review phase (Step 4) and moving straight from candidate themes to a final write-up.
- Reporting inter-rater reliability statistics while citing a reflexive thematic analysis approach that explicitly rejects them.
- Presenting a single supporting quotation per theme instead of multiple extracts from different participants.
- Not stating whether the analysis is inductive or deductive, or switching between the two without explanation.
How Much of Your Dataset Needs a Direct Quotation in the Write-Up?
There is no fixed proportion, but each theme should be supported by enough extracts that a reader unfamiliar with your full dataset can see the pattern for themselves, not just take your analytic claim on trust — as a practical guide, show more than one extract per major theme, drawn from more than one participant where the data allow, though your own supervisor’s expectations for your specific programme should take priority over any general guide.
Frequently Asked Questions
Do I need special software to do thematic analysis?
No — thematic analysis can be done manually with highlighters and a spreadsheet, though qualitative analysis software such as NVivo, ATLAS.ti or MAXQDA can make organising a large number of codes and extracts more manageable for bigger datasets.
How many interviews do I need for a thematic analysis thesis?
There is no universal number; the usual guidance is to continue data collection until no meaningfully new themes are emerging from additional interviews (thematic saturation), with the specific number depending on your population, question complexity, and your faculty’s own expectations.
Can I combine thematic analysis with quantitative data in the same thesis?
Yes, as a mixed-methods design, provided you clearly justify how the qualitative and quantitative components relate to each other (for example, one explaining or contextualising the other) rather than presenting them as two disconnected studies.
What is the difference between thematic analysis and content analysis?
Thematic analysis focuses on identifying patterns of meaning across a dataset with flexible, often qualitative, judgement about what counts as a theme; content analysis (particularly in its more quantitative forms) often involves counting the frequency of specific codes or categories, sometimes with a more structured coding frame decided in advance.
Should I use inductive or deductive coding for my thesis?
Choose based on your research question: inductive coding suits exploratory questions where you do not want to impose an existing framework, while deductive coding suits questions built around testing or applying an existing theoretical framework to new data.
Do all six phases need to happen in strict sequence?
Braun and Clarke describe the phases as recursive rather than strictly linear — you may move back to an earlier phase (for example, back to coding) if a later phase reveals the earlier work needs revision, and this is normal, not a sign of a flawed process, provided you describe what actually happened in your methodology chapter.
How do I decide how many themes to report?
Let the strength and distinctiveness of the pattern decide, not a target number — a thesis with a small number of strong, well-supported themes is stronger than one with many thin, overlapping ones.
Can two people code the same dataset without it being “reflexive” thematic analysis anymore?
Yes — using a second coder and coding-reliability checks is associated with earlier or more structured versions of thematic analysis, not a requirement of the method in general; state clearly which version of thematic analysis you are following and be consistent with its own stated principles.
What if a participant’s account does not fit any of my themes?
Note it explicitly as a deviant or contrasting case in your write-up rather than omitting it — discussing cases that do not fit the dominant pattern is a sign of careful, honest analysis, not a weakness in your findings.
Once your themes are defined and your extracts selected, Tesify can help you turn this analysis into a fully drafted findings chapter — you remain responsible for verifying every quotation and analytic claim it produces. Draft your findings chapter with Tesify.
For the validated instruments a mixed-methods psychology thesis might pair with this qualitative analysis, see which validated Malay-language scales you can use in a psychology FYP. For a complete worked psychology thesis example this analysis chapter would sit inside, see what a complete psychology FYP looks like. For choosing qualitative analysis software to manage your coding, see NVivo vs ATLAS.ti vs MAXQDA compared. For the literature review chapter your thematic findings will be discussed against, see how to write a literature review for your FYP or thesis.