Narrative-Integrated Thematic Analysis (NITA): AI-Supported Theme Generation Without Coding

The coding approach has dominated qualitative data analysis (QDA) since at least the introduction of grounded theory in the 1960s. The development of generative AI (GenAI) has led to an emerging debate about whether GenAI can help qualitative researchers shift away from the coding paradigm. Yet, the application of GenAI in QDA to date has largely focused on replacing human coding to improve efficiency and effectiveness, while the attention to facilitating the transition to alternative methods remains limited. In this article, we introduce a Narrative-Integrated Thematic Analysis procedure (NITA) that allows qualitative researchers to design, train and guide a GenAI in conducting thematic analysis. This NITA approach is regarded as a thematic analysis approach that positions researchers’ reflexivity, intellect and judgement at the centre of the analysis process. This approach combines a reflexive, iterative monitoring, evaluation and learning procedure (PERFECT) with a conversational method for interacting with and guiding GenAI. Using an publicly available interview dataset provided by Lumivero and custom ChatGPT, we applied this approach through six stages: planning an initial PERFECT procedure, preparation, generating candidate themes, constructing individual narratives, constructing meta-narrative, and writing up. Compared to our previous experiment on the same dataset using the coding approach, we have found that the themes generated this time had similar qualities. This reveal the transformative potential of GenAI to redefine thematic analysis.