Exploring Political Mistrust in Pandemic Risk Communication: Mixed-Method Study Using Social Media Data Analysis

Area of Science:

• Public Health
• Computational Social Science
• Health Communication

Background:

• Extends prior research on pandemic risk perception by the Finnish Institute for Health and Welfare.
• Focuses on the crucial role of trust in health authorities and its effect on public health outcomes during health crises.

Purpose of the Study:

• Investigate temporal and platform-specific variations in trust levels.
• Explore 12 subcategories of political mistrust, including accumulation, fluctuations, and topic relevance.
• Compare qualitative findings with computational analysis.

Main Methods:

• Analyzed 13,629 Twitter and Facebook posts (2020-2023) related to COVID-19.
• Utilized a fine-tuned FinBERT model (80% accuracy) for predicting political mistrust.
• Employed the BERTopic model for advanced topic modeling.

Main Results:

• Identified 43 mistrust topics across 9 major themes, including COVID-19 mortality, testing, and vaccine efficacy.
• Linked mistrust in authority to perceptions of disease severity, health measure adoption, and information seeking.
• Observed differing mistrust and misinformation patterns on Facebook versus Twitter due to platform-specific features.

Conclusions:

• Demonstrated the efficacy of natural language processing for managing large-scale data and misinformation.
• Reiterated the critical role of trust in health authorities for risk communication and public compliance.
• Stressed the need for transparent communication and a holistic approach to public health for effective crisis management.