Thematic mapping of interviews: connect themes, causes and exceptions
How to build an interview thematic map: move from codes to candidate themes, connect central ideas to causes, consequences and exceptions, then check the map against the data.
A theme explains a shared meaning
A list like “price, speed, support” is convenient, but it does not yet show what happened to participants or why. A thematic map links expectations, events, responses and consequences. It also brings cases that fall outside the main pattern into view.
Building the map takes interview review, coding and interpretation in light of the research question. A word cloud alone is insufficient. Keep notes and links to passages and distinguish participants’ words from your interpretation.
Move from passages to relationships and back
Draft candidate themes
Group codes by shared meaning. Write a sentence saying what each group explains about the research question. If it breaks into several unrelated ideas, consider splitting the theme.
Define each theme’s role
Identify central and supporting themes, possible causes, conditions, consequences and participant strategies. Label each arrow with the relationship, such as “reinforces,” “comes after” or “limits.”
Add evidence and exceptions
For every node, keep several passages from different interviews and at least one borderline case if you have one. Keep evidence that challenges the main account. It helps define the theme’s boundaries.
Check the map as a whole and by part
Read the map as an explanation, then return to the full interviews. Check for quotes taken out of context, sequences pieced together from different people’s accounts and themes broader than the data supports.
Mapping expectations of fast support
This is a fictional training example. The names, circumstances and outcomes are included solely to explain the method.
When they got back to me the same day, I was happy to wait. After two delays, I started posting the same question in a chat with colleagues.
A candidate theme could be “response predictability shapes workarounds.” Nodes might include the expected response time, repeated delays, falling trust in the channel and reposting the question. The links describe a possible sequence to check against other interviews.
An exception is a participant who waited longer without reposting because they could see the ticket status. That case refines the theme: knowing what is happening while you wait may matter alongside speed. This is a fictional example.
| Map element | Question | Evidence |
|---|---|---|
| Theme | What does it explain? | Interview passages |
| Link | How are the nodes related? | Labeled arrow |
| Condition | When does it occur? | Case context |
| Exception | What falls outside it? | Borderline passage |
Thematic map record
Keep a written record with the map. Without definitions, people can read the same diagram in different ways.
Research question: Map version and date: Central theme: Theme definition: Supporting themes: Labeled relationships: Conditions: Consequences: Supporting passages / timecodes: Exceptions and borderline cases: Alternative explanation: What to check in the next interviews:
Check the map against the data
- Each theme has a clear central idea.
- Relationships are labeled and can be explained.
- Quotes lead back to their full context.
- Exceptions are preserved and considered.
A closer look
Does a theme have to appear in most interviews?
Not necessarily. Importance depends on the question, depth and role of the theme. A rare case can reveal an important boundary.
Can I draw the map before finishing the coding?
Yes, as a working hypothesis. Label the version and expect to revise it after adding passages and rereading interviews.