Confusing Categories and Themes
J. Morse
DOI: https://doi.org/10.1177/1049732308314930
IF: 4.233
2008-06-01
Qualitative Health Research
Abstract:Despite the fact that categories and themes are different—they are used for different purposes in the research process, are developed from data using different strategies, and capture different forms of knowledge—I see that the terms are sometimes used almost interchangeably in completed research. This matters for the conceptual and theoretical structure of the completed research, and if there is a lack of cohesion between the methods and the results, the study becomes less comprehensible. Basically, a category is a collection of similar data sorted into the same place, and this arrangement enables the researchers to identify and describe the characteristics of the category. This, in turn, enables the category itself to be defined, and then compared and contrasted with other categories, or if broad in scope, to be divided into smaller categories, and its parts identified and described. A theme, on the other hand, is a meaningful “essence” that runs through the data. Just as a theme in opera occurs over and over again, sometimes in the foreground, sometimes in the background, and sometimes co-occurring with other tunes, so does the theme in our research. It is the basic topic that the narrative is about, overall. This comparison of categories and themes becomes clearer if we carry our opera metaphor one step further. Once I heard on the radio a content analysis of an opera. The writer had sorted all the trills and all the “Ah-has” into separate categories, and the result was ludicrous. But this example makes the difference between a category and a theme immediately obvious. Now, categories are important for determining what is in the data (the “what”). So they are used in ethnography and in the initial analytic phase of grounded theory. The ultimate use of categories is in the development of a taxonomy, in which the researcher identifies relationships between categories and smaller units, or subcategories. In some models (and again grounded theory is a good example), categories are sorted in trajectories over time, with some categories (or forms of the category) preceding or following others. Pushing the analysis further, the researcher can then determine what triggers the change, and move into theory development. But themes are used in the later phase of grounded theory to tie it all together (as the core variable or Basic Social Process), and they are the basic strategy of analysis in phenomenology, where the purpose is to elicit meaning or the essence of the experience for the participant. Analytic strategies for categorizing and “themeing” differ. Categories are developed using content analysis, in which similar chunks of text are ordered or placed proximally. They are separated from the interview or document itself; they are positioned so that example after example of the same thing may be examined, and the major commonalties may be identified, coded, explicated, and described. If a category becomes large (i.e., it contains a lot of examples), it may be separated into smaller units or subcategories. To identify a theme, the researcher reads the interview or document paragraph by paragraph, asking, “What is this about?”, and thinking interpretively. Analytic strategies may ease the process. Using the computer program’s highlighting features, emphasize key words and phrases. Then, using the footnote feature of the word processing program, make footnotes about the major (and sometimes the minor) emphasis of each section of text. Then the “convert footnotes to endnotes” feature places all of the footnotes together, and enables the footnotes themselves to be examined as a whole. This eases the process of making the themes identifiable. If we compare these two strategies for creating a category and identifying a theme, a category may appear at one part of a process (or appear in different forms in different stages), while a theme should go right through the data. Can a theme also be a category? Let’s consider love as an example. Examining it as a category, we would have categories of instances of love, targets of love, and perhaps even instances of love scorned. We may have the data to identify the characteristics of love of money (greed), puppy love, being in love, love of work, or love of play. We may even be able to differentiate between affection and passion. Creating such definitions decontextualizes, strips the link to a particular individual, but