When investigating more complex phenomena, data analysis can reveal concepts that can be developed to different levels of significance. For example, when developing categories you can distinguish categories, subcategories, their properties.
However, I have not find any recommendation in literature.
Comments and references to literature sources are welcome.
There are no rules for this sort of thing in qualitative research. Instead, I tell my students to think about how they will write the Results section of their article or thesis etc. In particular, if you have too many categories (themes) you will overwhelm the reader, and the same is true if a particular category has too many subcategories.
I suggest that you look at the qualitative publications in your field to see how many categories (themes) are typically reported, and model your own work on that standard.
Yes, it seems that there is no rules. However, some prominent researchers like Braun and Clarke (2006) presented a study with two levels related to themes and sub-themes.
Virginia Braun and Victoria Clarke (2006) Using thematic analysis in psychology. Qualitative Research in Psychology, 3 (2). pp. 77-101. DOI: 10.1191/1478088706qp063oa.
Further, Thomas (2006) presented a general inductive approach, and suggested that findings should include 3-8 categories. With 3-8 categories, one can expect at most 2 levels of categorization, in my opinion.
David R. Thomas (2006) A General Inductive Approach for Analyzing Qualitative Evaluation Data. American Journal of Evaluation, 27(2), pp. 237-246. doi:10.1177/1098214005283748.
Lets move to the field of software engineering (my expertise), in which qualitative methods have recently gained attention. Findings in reported studies vary in forms, size and granularity, which can be checked by reading articles in leading journals. Majority of published studies presented findings with larger number of categories (around 20) classified at one or two levels.
Thanks for suggestions, they confirm my feeling that it is important to justify research process, and develop findings that reflect investigated practice by using categorization that is most suitable.
Dear Zeljko,
I agree with David. However, there are some rules for the limiting situations. Any classification implies: a) a set of objects (tangible or intangible); b) a set of attributes associated; c) one criterion or several criteria based on which classification is done; d) a set of working hypotheses. A criterion expresses the property of belonging or membership to a certain category. For the same set of objects and attributes associated, by changing the criterion we change the result of classification. The first rule of any classification says that one object can belong to one and only one category. If objects are complex, than one may imagine several levels of classification. The second rule says that the lowest level of classification is reached when the entities contained in classification cannot be further decomposed. In my view, there is also necessity rule which says that one construct a classification with respect to a certain purpose. That means that the number of categories and levels stops when that purpose is achieved.
Wishing you inspiration and good luck!
- Badges
- Science topic