Services / Qualitative coding
Qualitative coding and thematic analysis you can evidence
Themes that appear without a visible trail from the data are the first thing an examiner or reviewer challenges. We build the codebook, code systematically, and hand back the audit trail that connects every theme to the extracts supporting it.
- A documented codebook
- Coded dataset with extracts
- Theme structure with evidence
- Inter-rater reliability if required
Qualitative coding is the systematic labelling of text — interview transcripts or open-text responses — so that patterns can be identified and evidenced. It produces a codebook, a coded dataset, and themes that can be traced back to the specific extracts supporting them.
What is included
What you get with qualitative coding
Codebook development
Inductive, deductive or framework — chosen to match your methodology rather than applied by default.
Systematic coding
Applied consistently across the whole dataset, with definitions that stay stable as coding progresses.
Themes with evidence
Each theme supported by extracts, so you can show where it came from rather than asserting it.
Inter-rater reliability
Double-coding and Cohen's kappa where your method or funder requires a reliability figure.
An audit trail
The decisions taken during coding, recorded — which is what a methods chapter or a funder review actually asks for.
Explained, not just delivered
A walkthrough of the framework so you can defend it in your own words.
Who it is for
Two kinds of client, one standard of work
- Interview transcripts and nothing coded yet
- A methods chapter needing a defensible framework
- Open-text survey responses at volume
- A supervisor asking how themes were derived
- Consultation and engagement responses
- Staff or customer free-text feedback
- Case notes and service user interviews
- Evidence for a report where method will be scrutinised
The method
How coding actually works
Qualitative coding is the process of turning transcripts into evidence. Extracts are labelled with codes, codes are grouped into categories, and categories are developed into themes — with every theme still traceable back to the specific pieces of data that support it.
The traceability is what distinguishes analysis from impression. When an examiner or reviewer asks how you arrived at a theme, the answer is a path through this chain, not a claim that it emerged.
Framework, thematic or content
Choosing an approach
The approach should follow from your research question and your epistemological position, not from what is most familiar. The three we use most are genuinely different, and stating which one you used — and why — is part of a defensible methods section.
Reflexive thematic analysis
Braun and Clarke's six phases. Inductive, interpretative, and explicit that the researcher shapes the analysis. Best when you are exploring meaning rather than testing a framework.
Framework analysis
A matrix of cases against themes, developed partly a priori. Best for applied and policy research where you need comparison across cases and a visible audit trail.
Content analysis
Systematic, often counted, closer to the quantitative end. Best when frequency and coverage genuinely matter to the question.
“Themes emerged from the data” is the single most criticised sentence in qualitative methods sections. Themes are constructed by an analyst making decisions; saying which decisions, under which approach, is what makes the work assessable.
Practice
Building a codebook that survives the whole dataset
Most coding problems are codebook problems. Codes defined loosely at transcript three mean something different by transcript thirty, and the resulting themes are unstable.
Inter-rater reliability, where it is required
If your method or funder requires a reliability figure, a subset is double-coded independently and agreement calculated — usually Cohen's kappa for two coders. Kappa corrects for agreement that would occur by chance, which is why raw percentage agreement overstates reliability and is rarely accepted.
Not every qualitative approach requires it. Reflexive thematic analysis explicitly does not treat coding as a reliability exercise, and reporting kappa alongside it signals a confusion about the method. We will tell you which applies to your design.
The deliverable
What you receive
The last item is not optional in practice. Coding you cannot explain is coding you cannot defend, and the walkthrough is where you take ownership of the analysis.
Software and ownership
Working with your NVivo project
Coding is done in NVivo unless you need otherwise, and the project file is handed over at the end. You own the coded data, can interrogate it yourself, and can continue the analysis without us.
That matters more than it sounds for doctoral work. An examiner may ask to see how a theme was built; being able to open the project and show the coded extracts is a considerably stronger answer than describing the process from memory.
Practicalities
Timescales and cost
Qualitative coding is priced by volume and complexity rather than by a fixed rate, because a 12-interview study and a 60-interview multi-site study are different undertakings. Send a sample transcript and the total count and we will quote a fixed price.
Sample size in qualitative work
How many interviews are enough?
This is the question we are asked most often, and the honest answer is that there is no number — but there are defensible ways to arrive at one, and “we stopped when we ran out of time” is not among them.
Saturation is the usual justification: the point at which additional data stops generating new codes or altering the theme structure. Its weakness is that it is often claimed rather than demonstrated. If you are going to cite saturation, you need to be able to say when it occurred and on what evidence — which means tracking new codes per interview as coding proceeds, not asserting it afterwards.
| Approach | Typical range | When it applies |
|---|---|---|
| Reflexive thematic analysis | 15–30 interviews | Exploratory work on a reasonably homogeneous group |
| Framework analysis | 20–50, spread across groups | Applied and policy research comparing defined stakeholder groups |
| Focus groups | 4–8 groups | Where interaction between participants is itself of interest |
| Case study | 1–5 cases, deeply analysed | Where depth within a bounded case matters more than breadth |
A heterogeneous sample needs more; a narrowly defined one needs fewer. What an examiner assesses is whether you can justify your number in terms of your question and your data, not whether it matches a table.
How it works
Three steps, entirely in writing
Send what you have
Attach the data, or just describe the project. No account, no mandatory call, and a five-field form rather than a fifteen-field one.
Approve a fixed quote
Usually within one working day, naming the statistician assigned, the deliverables and the deadline.
Receive the work
Output, syntax and a plain-English explanation — with follow-up questions answered at no extra cost.
Client feedback
What clients say about this work
“The coding process was systematic and well documented. Themes were clearly identified and supported by the interview data.”
“Fantastic qualitative analysis using a clear thematic approach. The coding framework made perfect sense.”
Proof
This work, on real projects
Clients are not named and no identifying detail is published, so what is described is the statistical problem — which is the part that shows whether a consultancy knows what it is doing.
Reflexive Thematic Analysis of Forty-Five Staff and Service User Interviews
Building a transparent coding framework across 700 pages of transcripts, where the aim was interpretation rather than counting.
Public services · Framework analysisFramework Analysis Comparing Perspectives Across Several Stakeholder Groups
Thirty interviews and eight focus groups from groups using different vocabulary, compared without flattening what made each distinct.
Third sector · Factor analysis and ordinal regressionExploratory Factor Analysis and Ordinal Regression on a National Survey
Reducing 75 overlapping questionnaire items to interpretable factors, then modelling an ordinal outcome correctly.
Your statistician
Who does this work
Member of the Royal Statistical Society
Psychometrics, structural equation modelling and research methodology, with doctoral supervision and journal review experience.
This page was reviewed for statistical accuracy by Elaine Halliburton on 14 August 2026.
Common questions
Answers before you ask
Which approach do you use?
Braun and Clarke reflexive thematic analysis, framework analysis, or content analysis — matched to your research question and stated in the deliverable.
Do you use NVivo?
Yes, and we can hand back the NVivo project so you own the coded data. We also train researchers in NVivo if you would rather do the coding yourself.
Will the themes be mine or yours?
The analysis is collaborative — we build the framework and code it, then walk you through it so you can challenge and refine it. Themes you cannot explain are of no use to you.
Can you handle transcription too?
We can arrange it, though it is quoted separately. Coding assumes clean transcripts.
How is it priced?
By volume and complexity of the data. Send a sample transcript and the total count and we will quote a fixed price.
Related
Often needed alongside this
NVivo training
NVivo training built around your own project rather than a demo dataset.
Fixed quoteStatistical consultancy
Design, analysis and reporting for charities, universities, the NHS and business.
From £950Data cleaning & preparation
Messy data made analysis-ready, with a documented log of every change.
Fixed quoteSend the data. Get a fixed quote.
Attach your dataset or just describe the project. A named statistician replies with a price and a deadline, usually within one working day.