When I applied thematic analysis to analyze the qualitative data, I labeled plenty of initial codes, such as "meat", "beef", "fish", and "chicken". However, these codes are at different levels, so I'm not sure how to organize these initial codes. Should I retain all these initial codes and summarize a theme named "meat"? If so, there is a code and theme that have the same name: "meat". And how should I calculate the frequency of these codes?
The attachment is two examples and which one is correct?
I suggest you look at Braun & Clarke's new (2022) book on Thematic Analysis. In general, they discourage this kind of "topical" labelling because it is unlikely to produce interpretive themes.
I'm going to go a step further. What are you trying to research? If you want frequencies as you pose, conduct a survey, i.e., quantitative data. Now if your meat example is connected to a larger idea, and I suspect it is, then build the context. [You can't have a superordinate category the same as a subordinate, which you have in your first example of meat and meat.] Maybe you're on the right track, but you do not give enough context.
Your title question: Can you abstract themes? Yes, all research generally involves some type of abstraction, or manifest and latent themes. Themes will not be as topical as you describe, but early on, you will over-categorize because you do not know what is important.
I have to admit that I did not notice that you were counting your codes when I wrote my first reply. In qualitative research, such counting is known as "content analysis," and that is very different from the approach that Braun and Clarke use for thematic analysis.
Xin, giving frequencies is commonly done in qualitative research. Johnny Saldaña even gave an example in an article where he was not sure about numbers in qualitative research.
Yet, many authors/researchers, see frequencies as missing the point. For example, what do frequencies mean in your context? One does not know; there is another major problem, often glossed over: Just because two people said something, did they both mean the same thing? [I have seen many times where I'm left scratching my head when frequencies are thrown out in qualitative research.]
My answer is you can do what you want to rigorously present your findings which tie into your research question. Many people do, but beware many methodologies/methods do not operate in such a manner, and just giving numbers even in quantitative research without context makes little sense.
Yes, you can use numbers if you create a codebook that you systematically apply to all of your data. This process of counting codes is what content analysis is all about.
Codebook gives you more viewpoint on how you will analyze your data. use it
- Badges
- Science topic
- Similar topics
- Thesis Research