Discovering Hidden Patterns: Applying Topic Modeling in Qualitative Research

Abstract

In qualitative studies, researchers must devote a significant amount of time and effort to extracting meaningful themes from large sets of texts and examining the links between themes, which are frequently done manually. The availability of natural language models has enabled the application of a wide range of techniques to automatically detecting hierarchy, linkages, and latent themes in texts. This paper aims to investigate the coherence of the topics acquired from the analysis with the predefined themes, as well as the hierarchy between topics, the similarity, and the proximity-distance between topics by means of the topic model based on BERTopic using unstructured qualitative data. This paper aims to investigate the coherence of the topics acquired from the analysis with the predefined themes, as well as the hierarchy between topics, the similarity, and the proximity-distance between topics by means of the topic model based on BERTopic using unstructured qualitative data. The qualitative data for this study was gathered from 106 students engaged in a university-run pedagogical formation certificate program. In BERTopic procedure, the paraphrase-multilingual-MiniLM-L12-v2 model was used as the sentence transformer model, UMAP was used as the dimension reduction method, and HDBSCAN algorithm as the clustering method. It was found that BERTopic successfully identified six topics corresponding to the six predicted themes in unstructured texts. Moreover, 74% of the texts containing some certain themes could be classified accurately. The algorithm effectively discerned which themes were analogous and which had significant distinctions from others. It was concluded that BERTopic is a procedure which is capable of identifying themes that researchers may not notice, depending on the data density in qualitative data analysis, and has the potential to enable qualitative research to reach more detailed findings.