> Qualitative Analysis

Initial Coding

First-cycle coding — the shared mechanic, and where the two paths' coding stances diverge. · 9 min

First-cycle coding is your first systematic pass: break the data into chunks and attach a short label to each (Saldaña, 2021). Two workhorse semantic styles — codes that stay close to what was said:

Descriptive coding — you summarize the chunk in a word or phrase (help-seeking delay).
In vivo coding — the code is the participant’s own words ("think I was dumb"), preserving their voice (Charmaz, 2014).

The mechanic — segment, then label — is common to both traditions. But the coding stance already diverges here, and neither is the neutral default:

The codebook / coding-reliability path favors semantic codes like the ones above: close to the surface, so they can be defined and applied consistently across coders. They feed the codebook in the next lessons.
Reflexive TA leans on latent coding — reading for the meaning underneath the words. The Reflexive Coding lesson recodes these same excerpts that way (e.g. "think I was dumb" → managing how I am seen).

The worked example below uses the semantic stance — read it as one reading, not the default.

Code the example

Same excerpts as before, now coded the semantic way. Click any highlighted span to see the code, its type, the memo justifying it, and what theme it rolls up to on the codebook path. (The roll-up is shown for orientation; in real work, themes come later — don’t pre-judge. Path B’s latent recoding of the same spans is in Reflexive Coding.)

Coded segment — click to inspect

It was a linked-list thing, a null pointer exception. . Honestly for missing something basic. Before that I . When I finally asked, and I felt silly for waiting.

For debugging I don't really have a system. until something worked — . The compiler kept throwing this , and I , so I just and pasted in whatever came up.

The thing that gets me is the error messages. — it doesn't say where or why. . So I , just guessing really. After a while , and .

Click a highlighted segment to see how it was coded and why.

Inductive, deductive, abductive?

Where do codes come from? Three stances, and most analysts move between them:

Reasoning	How codes arise	Source
Inductive	codes emerge from the data	(Thomas, 2006)
Deductive	a pre-set framework is applied first	(Hsieh & Shannon, 2005)
Abductive	looping between data and theory — where most analysts land, guided by “sensitizing concepts”	(Timmermans & Tavory, 2012), (Blumer, 1954)