February 1, 2026 | 3 minute read
Invitation to Grounded Theory, chapter 5: The Logic of Grounded Theory Coding Practices and Initial Coding
by
Text Exploration
In this chapter, the author introduces the approach she uses for coding transcription data.
A code establishes a relationship between the data and the research participants. At its base, it means naming segments of data (usually line by line) with a label that “simultaneously categorizes, summarizes, and accounts for each piece of data.” This is an interpretative process. Codes are loosely related to the research data. They emphasize processes, actions, and feelings, and are written with a creative voice.
Coding is a tool for “interrogating, sorting, and synthesizing hundreds of pages of interviews, fieldnotes, documents, and other texts.” This is a process of examining the data in detail, and making the interpretation explicit. The coding process defines what is happening in the data, and coding “generates the bones of your analysis” which are then assembled into a structure of a theory (through theoretical centrality and integration).
Coding occurs through an initial phase of naming, and a selective phase to identify the most significant or frequent codes. Codes are created, rather than applied, which is the primary distinction between a grounded theory approach and a thematic analysis approach.
Continual coding “prompts you to keep interacting with your data” through an interactive space, and the author describes that a researcher may feel “awestruck about the privilege of learning about it.”
Initial coding should focus on language related to action. The codes are provisional and grounded in the data. They should be created quickly and spontaneously, and should be written using gerunds to make the data descriptive and active. This is a heuristic device.
Constant comparative methods begin to develop distinctions between data; this process focuses on similarities and differences.
In Vivo codes describe the special terms that are used by participants, and “help us to preserve participants’ meaning of their views and actions in the coding itself.” They represent social worlds and organizational settings.
Three observations; first, this is the only real practical description of technique that we’ve encountered yet, so that’s useful. Next, as I wrote before, her writing is terrible. It just circles the same ideas over and over. Finally, this is strongly related to the process we used at Narrative of interpretation and synthesis, and it feels very familiar and “safe” to me. I’m going to take this process at face value when I begin to make sense of student data from my next study, and I’ll explore different tools to support this, rather than my old-school Excel format.