NLP and GenAI Analysis: Automating Thematic Analysis and Coding of Qualitative Educational Research Data

Abstract

Qualitative analysis of large datasets often involves manual coding and theme extraction, a process that is both time-consuming and subject to human interpretation. Traditional approaches to ensuring inter-rater reliability, where multiple analysts independently manually code data to achieve consensus, can be particularly laborious. Advances in Natural Language Processing (NLP) and Generative AI (GenAI) offer promising alternatives to streamline these processes by automating theme extraction and coding (Fitkov-Norris & Kocheva, 2023). Improving the efficiency and accuracy of qualitative analysis is crucial in Science Education Research, where effectively interpreting student open-ended survey responses and interviews is essential for promoting engagement with scientific concepts and research tasks (Sinatra et al., 2015).