Overview
Eight earlier qualitative studies of Korean women leaders collected 200 narratives across sectors and found challenges in work-life balance, leadership development, and career development that were largely attributed to a gendered workplace. This poster, with Jieun You, Sumi Lee, and Yonjoo Cho, reanalyzes those narratives with topic modeling and compares the computed topics with the themes the qualitative studies had identified.
The narratives were converted into paragraphs and cleaned (stop-word removal, lemmatization). Keywords for each topic were examined with four indexes: highest probability, FREX, lift, and score. Models with different numbers of topics were compared, and the topics were then labeled from their keywords.
Key points
- The team reached consensus on 30 topic labels, for example women’s promotion and work environment.
- The topics overlapped with themes from the earlier studies, such as women’s leadership and work-life balance.
- The computational analysis also surfaced aspects that the qualitative studies may have passed over, such as living abroad and multicultural experience.
- Topics were limited in capturing subtle and meaningful aspects of the leaders’ experiences, so the findings are presented as preliminary.
- The poster discusses computational analysis as a counterweight to the subjective nature of qualitative research.
Event
2024 Academy of Human Resource Development International Research Conference in the Americas (poster session).
Citation
You, J., Chae, C., Lee, S., & Cho, Y. (2024). A computational analysis of Korean women leaders’ narratives using topic modeling. Poster presented at AHRD 2024.