A practical guide to implementing ChatGPT as a secondary coder in qualitative research

Abstract

Analyzing interview data provides in-depth insights into qualitative research topics, but is often a time-consuming and costly process. This research aims to enhance this process by leveraging transformative technologies influencing the accounting field, like ChatGPT. ChatGPT is capable of generating human-like text and performing text-based analyses by using a pre-trained model. Specifically, this research illustrates ChatGPT’s role as a secondary coder in deductive qualitative accounting research, analyzing interview transcripts using content analysis with a predefined coding scheme. Data was collected through semi-structured interviews with 36 business economics students. First, the primary researcher manually analyzed the data using a deductive approach. Next, a second human researcher and ChatGPT (ChatGPT-4o Plus) were appointed as secondary coders to independently verify and validate the coding. The paper demonstrates ChatGPT’s potential to assist in coding interview transcripts, emphasizing its role as a supplementary tool in qualitative research. The coding results from both secondary coders were compared with those of the primary human coder, revealing over 99% agreement for both secondary coders. In addition, a step-by-step illustration, best practices, prompts, and critical reflections are shared. This study serves as a foundational step in understanding and leveraging ChatGPT’s use in the analysis of interview transcripts.