Some analysis methods introduced.
In the thematic analysis, the researchers isolate patterns, themes, from qualitative data. Inductive thematic analysis is data-driven approach as there the idea is to find the themes from the data without a help of theory (of course, the research question gives a focus what to look for the data) whereas theoretical thematic analysis looks for patterns relating to a theory and initial themes are be based on the previous research. The theoretical thematic analysis is also called as the template analysis. Template analysis is described below and here we focus only on inductive thematic analysis.
The thematic analysis has following phases.
- Familiarizing with the data. In this stage, all the data is read through or inspected in detail.
- Generating initial codes. In this stage, the interesting features of data are coded. Whenever a new feature is encountered a new code is created. After new codes have been added, the whose data is coded based on the most recent set of codes.
- Searching themes. In this stage, the codes are grouped together. The aim is to collect all data via codes that belong together in one group.
- Checking themes against the codes and the data
- Defining and naming the themes. The found themes are then given a descriptive name based on the data.
- Writing the report.
Typically thematics analysis work includes moving freely between different phases.
Quote Coded as "XXX" (XXX, man 27, p.5) xxx
In coding, it is important to keep information about where the quote came from so that it is easy to find later.
Searching for themes
Some practical way to group the themes
- Hand sort themes. Put the codes (and quotes relating to a code) on a piece of paper and sort the themes into groups. Emerging groups are themes.
- Draw mind maps about the codes. Add related themes as children of the same node. Those nodes later are named and are themes.
- Create a table from the data and group codes on the table, for example, use columns as themes.
Draw a thematic map and revise it until you have themes (see an example of a thematic map below)
Fig a thematic map here
- How well the grouped codes and data matches together?
- Is the theme homogenous?
- Do the codes fit together naturally?
- Do the themes describe data well?
- Do the themes match data well?
- Do the thematic map accurately represent the data?
- Can a researcher that was not involved in coding and searching these agree your logic behind coding and grouping codes into themes?
If some codes are different to other ones in the group, drop those codes out from the theme.
Defining and naming themes
An alternative approach
In this version, instead of directly coding the data, the data is first highlighted, then shortened by paraphrasing. Paraphrasing is used to form themes. After that, the paraphrased data snippets are changed to codes.
1. highlight things in the data that are interesting in terms of your research question
2. Create a table from the highlighted data. Group by questions and participants.
Participant A Participant B Question 1 - hightlight 1 - highlight 2 Question 2 - highlight 3 - highlight 4
3. Paraphrase the data so that the quotes inserted to the table are more concise, but have the same meaning
4. Combine the quotes in the table to themes, but keep the participant intact. In this stage delete duplicates. Note that the themes should not come directly from the questions.
Participant A Participant B Theme 1 - paraphresed 1 - paraphrased 5 - paraphresed 4 Theme 2 - paraphrased 3 - paraphrased 6 - paraphrased 2 - paraphrased 7
5. Code the answers
Participant A Participant B Theme 1 - code 1 - code 1 - code 4 Theme 2 - code 3 - code 5 - code 2 - code 2
Combine themes as much as possible. Repeat steps 4 and 5 until you have a small set of themes and codes that can be used to describe an answer to your research question.
6. reporting the result
Template analysis is a form of thematic analysis where (typically) textual data is analyzed based on predefined codes (categories). This predefined list of categories is called as a coding template. The initial codes are typically based on theory.
The analysis starts with reading the data and marking segments using codes in the template. The template is then modified based on reading data. The modified final template is then applied to all data.
Steps in template analysis
- Define initial template based on theory and the research question
- Transcribe the data
- Initial coding. In this stage in addition to coding data based on the initial codes, the template the template is modified when needed (introduce new codes and revise codes)
- Use the final template to the full data
- Interpret the coded data: group codes and create themes based on the groups; prioritize the themes (what are the important ones in terms of your research questions); look connections between themes. Again mindmap and physical sorting (codes on small pieces of paper and sorting them), for example, are useful.
- Write the analysis results.
More about thematic analysis
- Braun and Clarke, 2006
The are also software for coducting thematic analysis
Taxonomies, typologies, concept charts, flow charts
(Loland et al., 2006)
Data analysis using coding is explained in the observations section.
Qualitative content analysis