How to develop a thematic framework - MY BEST ADVICE + coding in NVivo 12

A thematic framework should function like a self-contained map of both the study’s purpose and its findings—so a researcher shouldn’t need to re-read full transcripts to understand what matters. As coding develops in NVivo 12, the workflow shifts from building codes by scanning the data to “reading” the data through the coding framework: opening a specific code and reviewing the extracts assigned to it, rather than repeatedly opening and scrolling through entire interview documents.

That principle drives the practical advice in the lesson. The coding structure—how main themes/categories and their subthemes/child codes are organized—should communicate two things at a glance. First, it should make the study’s aims legible through the names and arrangement of the main themes. Second, it should make the findings visible through the wording and structure of subthemes/child codes, so someone reviewing the NVivo project file could infer what the research focused on and what was found without reading the raw interviews.

The second half turns to NVivo mechanics for “reading your codes,” starting with a basic navigation problem: knowing which parts of a transcript have already been coded, and which codes were applied there. NVivo can highlight coding in the source, but the lesson warns that this depends on whether the source was turned into a “case.” When sources are converted into cases, selecting “highlight all coding” can backfire because the entire transcript becomes coded under the case node—making everything look coded and obscuring the real, substantive codes.

To avoid that, the recommended approach is to highlight coding for selected items and manually choose only the actual codes created (excluding the case node). Once the coded segments are highlighted, a second NVivo feature—coding stripes—adds the missing layer: it shows, next to each coded extract, which specific codes were applied. In the example, different extracts carry different labels (e.g., “overwhelmed” and “creativity”), and the coding stripes can also reveal longer spans associated with a particular code.

Finally, the lesson shows how to read what each code contains inside the coding framework. In the Codes folder, a researcher can open a code (via double-click or right-click → “open node”) to view all extracts assigned to that code. NVivo provides multiple views for interpreting a code: a reference view for locating where it appears, a summary view that reports how many references and the coverage percentage (how much text the code spans), and a text view that becomes especially useful as the number of codes and files grows. With these tools, the coding framework becomes navigable—each code acts like a doorway into the relevant evidence—setting up the next step of refining the framework by decoding and merging codes.