From Text to Context. Navigating Thematic Analysis and AI-Powered Tools in Qualitative Research

Qualitative research has long relied on traditional thematic analysis, a method that involves the manual identification and analysis of recurring themes within a dataset. While effective, this approach can be labour-intensive and subjective, often requiring researchers to sift through data multiple times. The advent of basic qualitative data analysis software and natural language processing tools offers an opportunity to streamline and enhance this process.
Another promising avenue is topic modelling, a technique that leans more towards quantitative statistical analysis. Unlike thematic analysis, topic modelling processes large text corpora and relies less on domain-specific expertise. When integrated with semantic context tools like Word2Vec, it provides valuable insights into the topic structure of the analyzed text. Moreover, it necessitates fluency with programming tools, such as Python libraries, broadening the skill set required for qualitative research.
The landscape of qualitative research is further revolutionized by the emergence of AI-powered tools like BERT and GPT. These large language models facilitate the processing of extensive textual data, opening doors to new techniques and approaches. They enable the development of (semi-)automated workflows that significantly reduce the researcher workload associated with data preparation, analysis, and synthesis, without compromising the qualitative essence of the research.
We explore the synergies and challenges at the intersection of traditional thematic analysis, topic modelling, and AI-powered tools. We discuss how these methods can be integrated to create a more efficient, objective, and comprehensive qualitative data analysis process.