Axial coding
Axial coding is the grounded-theory step in which a researcher links the codes from open coding to one another, gathering them into categories and working out what causes each category, the context it occurs in, how people respond and what follows. Selective coding then builds on it.
How it works
Open coding breaks the data into many small labeled pieces. Axial coding reassembles them around an axis: a category such as "switching banks" is examined for its conditions (what prompts it), the actions and interactions around it, and its consequences. Codes that describe the same phenomenon merge, and codes that explain one another are linked.
The researcher moves back and forth between the data and the emerging categories, comparing each new passage against what is already coded, a practice grounded theory calls constant comparison. Axial coding then feeds selective coding, where one core category is chosen to tie the whole analysis together.
A worked example
A credit union's research team has 140 open codes from 22 interviews with members who left. Axial coding groups them into 6 categories. One, "fee surprise", links a condition (a first overdraft), an action (calling the branch) and a consequence (opening an account elsewhere within 30 days), and it appears in 15 of the 22 interviews.
Related
Automate this in Pulse
An axial coder can take two Pulse themes that keep appearing in the same interviews, such as late delivery and canceled orders, and read those transcripts for the condition that links them.