Poster #59 - Impact of Mental Health Indicators in Online Citizen Petitions on Government Responses: A Machine Learning Approach

• In Event: Friday Poster Reception

Abstract

With the rise of digital governance, online platforms have become a crucial medium for governments to interact with citizens. This study investigates how mental health indicators within citizen petitions influence government responses on the national online government board. Using text data from the China Leadership Message Board, we employed Large Language Models (LLMs) and machine learning techniques to identify mental health risk factors in citizen submissions. We then analyzed the relationship between these factors and government response strategies, specifically focusing on the effectiveness and specificity of the responses with both Ordinary Least Squares regression and machine learning correlation analysis method. Our findings reveal that mental health indicators in petitions are negatively correlated with the effectiveness of government responses, indicating that officers provide less effective measures or clear feedback when addressing petitions highlighting mental health concerns. Conversely, a positive correlation was observed between mental health indicators and the specificity of responses, suggesting that while responses are less effective, they may often be tailored rather than template responses.

These results suggest that the government’s ability to identify mental health issues in petitions is limited due to their complexity and subtlety compared to more straightforward sentiments or emotions. Additionally, even when mental health concerns are recognized, the government may find it challenging to provide effective responses due to limitations in authority or the sensitivity of the issues. This study underscores the importance of equipping government agencies with the tools to identify mental health indicators in citizen petitions and interactions, enabling them to offer more useful and tailored responses. Based on these findings, we propose recommendations to improve the interaction between citizens and the government, ultimately contributing to the modernization of the policy process.

Author

• Wenjin Chen, The Hong Kong University of Science and Technology
  Presenting Author