The Promise of AI-Assisted Qualitative Analysis in Educational Research: An Example of Thematic Analysis

• In Event: AERA Roundtable Session Thursday 4:15 pm - Gold 4
  In Roundtable Session: Qualitative Inquiry in the Age of AI: Reflexivity, Collaboration, Cultural Responsiveness, and a Counter Balance

Abstract

AI-assisted analysis has been integrated into commercial qualitative analysis software for years, but researchers have raised concerns over its validity. Moreover, its (non)applicability in non-English texts remains unclear. In this study, we applied the Social Cognitive Career Theory to understand important factors of career development for college students in Taiwan. Interviews were transcribed verbatim in Mandarin Chinese, then uploaded to ChatGPT for thematic analyses. The findings revealed that, with clear procedures, ChatGPT effectively identified patterns across four constructs: interest, self-efficacy, goals, and actions, and demonstrated the ability to analyze Chinese texts, including detecting implicit expressions. While ChatGPT demonstrated promising analytical capability, the study also highlighted the importance of researcher oversight to ensure coherence and accuracy.

Authors

• Thi Phuong Vy Nguyen, National Taiwan Normal University

• Ying-Shao Hsu, National Taiwan Normal University

• Hsun-Yu Chan, National Taiwan Normal University

• Yi-Fang Lee, National Taiwan Normal University

• Hoang Bao Ngoc Nguyen, National Taiwan Normal University

• Ching Hsu, National Taiwan Normal University

• Hsiao-Ping Yu, National Taiwan Normal University

• Kai-Lin Yang, National Taiwan Normal University