Presentation Information
[2A04]A Comparative Study of Quantitative Text Analysis and Generative AI Analysis of Public Poll’s Free Text Comments on Nuclear Energy
*Yuki HASHIMOTO1, Masako IKEGAMI1 (1. Institute of Science Tokyo)
Keywords:
Public poll,Free text comments,Quantitative text analysis,Qualitative text analysis,Generative AI
Since 2006, Public Poll on Nuclear Energy have been conducted, incorporating both multiple-choice questionnaires and free text comment sections. However, the latter has not been sufficiently analyzed or examined in past reports. To address this gap, we conducted quantitative text analysis using KH Coder and both quantitative and qualitative analyses using Chat-GPT. By comparing and evaluating these approaches, we demonstrated the utility of analyzing free text comment sections and validated the methodology and reliability of AI-assisted text analysis.
The results confirmed that the content of the free text comments largely aligns with the trends in the multiple-choice survey results. Additionally, we revealed deeper insights and changes in public opinion toward nuclear energy that had not been captured by conventional methods. Furthermore, the use of AI facilitated automated execution of certain aspects of both quantitative and qualitative text analysis, demonstrating the potential for more diverse approaches to analyzing free text comments.
The results confirmed that the content of the free text comments largely aligns with the trends in the multiple-choice survey results. Additionally, we revealed deeper insights and changes in public opinion toward nuclear energy that had not been captured by conventional methods. Furthermore, the use of AI facilitated automated execution of certain aspects of both quantitative and qualitative text analysis, demonstrating the potential for more diverse approaches to analyzing free text comments.