Qualitative analysis often relies on the use of codes to make sense of raw data through a transparent and organized process that helps researchers build empirical insights. Among the many methodologies available to qualitative researchers, grounded theory can be adapted to generate new theories about the phenomenon being studied. In qualitative research that adopts a grounded theory method, axial coding is the second part of the coding process that comes after open coding and before selective coding. Axial coding is an important set in the process of grounded theory, so it is important to outline what axial coding is and how it is essential to your research project if you are following grounded theory.
Remember that analyzing qualitative data relies on the use of codes that summarize the meaning gathered in data collection. The process of coding must be conducted in a transparent and rigorous manner.
Axial codes are often associated with grounded theory development, which researchers employ when they want to build theory arising from the data they have collected. Open, axial, and selective coding are the three key steps in grounded theory coding. Open coding breaks down data into discrete parts of meaning while selective coding employs codes to contribute to a theoretical framework. Within a coding paradigm under grounded theory procedures, axial coding aims to tie these steps together by organizing initial codes into categories which researchers can develop into a theory through their empirical analysis.
By the time you get to the axial coding stage, you should have a set of open codes and, for the moment, they may have seemingly little to do with each other. The other issue after initial coding is that you might have very many discrete codes that make the identification of key themes arising from your data difficult. Now is the time to organize them into broader categories.
Axial coding is simply the process of identifying larger themes relevant to your research question. These form the "axes" around which the codes can be organized. Suppose you are analyzing interview data about perspectives regarding personal behavior and well-being. Along the way, you have created codes such as "morning exercise," "sleep," and "vegetarian diet." There are also codes such as "video games" and "smartphone usage," but for now, you might posit a connection among the first three codes and organize them under the category "healthy lifestyle."
This axial code represents a potentially key theme that may be significant to your data analysis and theory generation. More importantly, organizing existing codes from the open coding phase into axial codes helps to establish connections between those discrete elements. Enabling researchers to identify relationships elevates the initial codes from otherwise separate and unrelated entities to a more cohesive unit for a comprehensive understanding of the data.
While grouping codes together into categories, you can employ a process called constant comparison, especially as you incorporate new data into your project. Ask yourself if the new data (or a re-reading of the existing data) affirms or challenges the categories you identify by comparing discrete segments of data with one another. For example, does your definition generated by the category "healthy lifestyle" apply to the new data, or do you need to revisit which codes belong in the category? Maybe when people in your interview study talk about a vegetarian diet, they do so in ways that are more about environmental consciousness and less about personal well-being.
In such a case, you may want to reconsider whether that code belongs in this category. On the other hand, new data may identify codes for a healthy lifestyle, such as "spending time outdoors" or "mindfulness," that may not have existed in the initial data. Constant comparison in this case ensures that you are perpetually developing your categories consistent with the data you analyze.
As with any stage of coding, axial coding as a process of organizing codes into categories relies on subjective interpretation. Since grounded theory is a bottom-up and, therefore, open-ended approach to data analysis, different people looking at the same data may interpret the data in distinct ways.
To address this challenge, researchers should always make their iterative process for coding clear to their research audience. What codes did they create from open coding and why? In what ways were the open and axial codes that were created relevant to the research question being addressed? How did categories develop in the process of reading the data? Answering these and other questions related to research transparency is not aimed at making the research more "objective," but to clarify to your audience how you arrived at your axial codes and provide clear guidance to other researchers about the grounded theory process.