Can thematic analysis be deductive?
Yes, thematic analysis can definitely be deductive. Thematic analysis is a flexible method that can follow either a deductive (theoretical, top-down) or inductive (bottom-up) approach.
Key Takeaways
- Deductive TA: Yes, thematic analysis can be deductive: the researcher applies a pre-existing theory or framework to guide coding, rather than letting themes emerge from the data alone.
- Top-down vs bottom-up: Deductive coding works top-down from theory; inductive coding works bottom-up from the data itself.
- When to use: A deductive approach suits research that tests or confirms an existing theory, such as coding therapy transcripts against a CBT framework.
- Advantages: It is efficient, keeps the analysis focused on the research question, and supports direct comparison with earlier studies.
- Risk: A fixed framework can miss unexpected themes that fall outside it, so many researchers combine deductive and inductive coding in a hybrid approach.
What is deductive thematic analysis?
In a deductive approach, the researcher starts with a pre-existing theory or framework and uses it to guide coding and theme development (Braun & Clarke, 2006). The theory sets the agenda first. Themes are not simply drawn from the data; they are shaped by the researcher’s prior knowledge.
Braun and Clarke (2006) illustrate the contrast with one line of data. Given “I answer emails at 10 p.m. because the laptop is right there,” an inductive coder might code it as “always-on availability.”
A deductive coder would code it differently. Working from boundary theory, they might label the same line “boundary permeability,” a concept imported from the theory.
But most analyses are a blend. Even a mostly deductive study should stay alert to patterns the framework did not predict (Braun & Clarke, 2006).
For example, imagine analyzing interviews about workplace stress using a deductive approach. You might start with established categories from stress theory, such as “workload,” “interpersonal conflict,” and “work-life balance,” as your initial coding framework.
This fully deductive form is sometimes called “theoretical thematic analysis.” A pre-existing theory sets the coding frame, not the data itself.
When is it appropriate to use a deductive approach to thematic analysis?
A deductive approach to thematic analysis is most appropriate when the researcher has a strong theoretical framework that they want to use to guide the analysis.
It is also appropriate when the research question is focused on testing or confirming existing theories.
A classic example is cognitive-behavioural therapy research.
A researcher analysing therapy transcripts through a CBT lens might code specifically for “automatic thoughts,” “cognitive distortions,” and “behavioural experiments” (Braun & Clarke, 2006). Working from a fixed, theory-driven code list like this is sometimes called codebook thematic analysis.
How is deductive thematic analysis different from inductive thematic analysis?
In inductive thematic analysis, the themes are generated from the data itself, without any preconceived notions.
In contrast, in deductive thematic analysis, the themes are determined before the data is analyzed.
Deductive thematic analysis is a top-down approach, while inductive thematic analysis is a bottom-up approach.
The table below summarizes the main contrasts.
| Inductive | Deductive | |
|---|---|---|
| Starting point | The data itself | A pre-existing theory or framework |
| Direction | Bottom-up | Top-down |
| Best suited to | Exploring what the data contains | Testing or confirming an existing theory |
| Main risk | Missing an organizing pattern | Missing themes outside the framework |
Can I use a combination of inductive and deductive approaches in thematic analysis?
Yes. In practice, many researchers use a combination of inductive and deductive approaches to thematic analysis.
This is sometimes called a “hybrid approach.”
For example, a researcher might start with a set of pre-existing themes, but then remain open to new themes that emerge from the data.
The choice of whether to use an inductive, deductive, or hybrid approach depends on the research question, the existing theoretical framework, and the researcher’s epistemological stance.
What are some examples of deductive thematic analysis?
A key published example of deductive thematic analysis is Clarke and Kitzinger’s (2004) study of how lesbian and gay parents were represented on television talk shows.
- Aim: To examine how lesbian and gay parents were represented and represented themselves on daytime television talk shows.
- Method: Rather than letting themes emerge, the analysts coded talk-show transcripts against the pre-existing concept of heterosexism. This is the assumption that heterosexual parenting is the unmarked norm against which gay and lesbian parenting must justify itself.
- Results: Participants used discursive strategies of normalization. They emphasized how lesbian- and gay-headed families conformed to the norms of white, middle-class heterosexuality, responding to homophobic and heterosexist critiques of their parenting.
- Conclusion: Because the framework told analysts what to look for, the coding showed whether parents’ talk resisted the heterosexist framing or, at points, reproduced it.
What are the advantages of using a deductive approach to thematic analysis?
There are several advantages to using a deductive approach to thematic analysis.
- Efficiency: It can be faster to analyze data when the researcher already has a strong theoretical framework in mind.
- Focus: It helps keep the analysis focused and relevant to the research question.
- Comparability: It supports direct comparison with previous research, building cumulative knowledge in a field.
- Trade-off: It may miss unexpected themes that do not fit the predetermined framework.
What are the disadvantages of using a deductive approach to thematic analysis?
There are also some disadvantages to using a deductive approach to thematic analysis.
First, it can limit the flexibility of the analysis. The researcher may be less likely to identify new or unexpected themes that emerge from the data.
Second, if the researcher’s pre-existing themes are not well-founded or are not relevant to the data, the analysis may be biased or inaccurate.
Finally, the themes determined in advance might not capture everything in the data.
What are some tips for conducting deductive thematic analysis?
- It is important to be transparent about the choices you made during the research process.
- Explain why you opted for specific methods and discuss implications for future research.
- You should also be consistent in applying these choices throughout the analysis.
1. Clearly Define the Research Question:
A well-defined research question is crucial for any research, but it is particularly important in deductive thematic analysis.
This is because the research question guides the choice of theoretical framework and the coding scheme. The coding scheme is the list of categories and rules used to sort the data.
2. Identify the Existing Theoretical Framework:
The next step is to identify the pre-existing theoretical framework that you will use to guide your analysis.
This framework should be relevant to your research question and provide a clear set of concepts or themes that you can use to code the data. Coding means attaching a short label to each relevant segment.
Sources like literature reviews or existing research in your field can be valuable resources for identifying relevant theoretical frameworks.
For example, a study on panic buying used the Theory of Reasoned Action and Protection Motivation Theory to guide their analysis.
3. Develop a Coding Scheme:
Once you have identified your theoretical framework, you need to develop a coding scheme that is based on the concepts or themes within that framework.
The coding scheme should include a list of codes that represent the different themes you are interested in.
This will help you to systematically apply the framework to your data.
A panic buying study developed codes like ‘Provoke perception‘, ‘Anxiety‘, and ‘Eminence‘ based on concepts from Theory of Reasoned Action and Protection Motivation Theory.
4. Familiarize Yourself with the Data:
Just like in other forms of thematic analysis, you need to become familiar with your data before you start coding.
This means reading through the data multiple times and taking notes on anything that stands out to you.
You may also want to create summaries of the data or develop mind maps to help you visualize the data.
5. Code the Data:
Once you are familiar with the data and have a coding scheme developed, you can begin coding the data.
This involves reading through the data and assigning codes to the different segments of the data that reflect the themes in your coding scheme.
Some segments of data might be relevant to multiple codes.
Codes at this stage may be semantic, staying close to what was said, or latent, interpreting an underlying idea (Braun & Clarke, 2006). Even within a deductive framework, a new code can still be added inductively if something unexpected keeps appearing in the data.
6. Analyze the Coded Data:
After you have coded all of the data, you need to analyze the coded data to identify patterns and relationships.
This involves looking for connections between the different themes and how they interact with each other.
You can use a variety of methods to analyze the coded data, such as creating tables, charts, or diagrams.
7. Interpret the Findings:
The final step of thematic analysis is to interpret the findings of your analysis in relation to your research question and theoretical framework.
This involves discussing the implications of your findings and how they contribute to the existing body of knowledge.
It may involve identifying areas where the data supports the pre-existing theory, areas where it challenges the theory, and areas where the theory might need to be refined or expanded.
