Emotions and Incivility in Vaccine Mandate Discourse: Natural Language Processing Insights

Area of Science:

• Computational Social Science
• Psychology
• Public Health

Background:

• Vaccine hesitancy and political polarization hinder COVID-19 pandemic mitigation efforts in the US.
• Elected officials' differing responses to vaccine mandates amplify public distrust.
• Social media discourse on COVID-19 is polarized, with attitudes often driven by political ideology over scientific evidence.

Purpose of the Study:

• To investigate the psychological processes underlying uncivil discourse related to COVID-19 vaccine mandates.
• To examine the roles of anxiety, anger, and sadness in predicting online incivility regarding vaccine policies.
• To analyze incivility as a multifaceted construct in social media discussions.

Main Methods:

• Utilized natural language processing (NLP) with the Linguistic Inquiry and Word Count (LIWC) tool and the Google Perspective API.
• Analyzed 8,014 tweets containing COVID-19 vaccine mandate keywords from September 14 to October 1, 2021.
• Measured emotional valence (anger, sadness, anxiety) and dimensions of incivility (toxicity, insult, threat, etc.).

Main Results:

• Anger was significantly positively associated with all dimensions of incivility (toxicity, severe toxicity, insult, profanity, threat, identity attacks).
• Sadness was positively related to threat, while anxiety was negatively associated with identity attack and profanity.
• Emotions like anxiety, anger, and sadness were uniquely linked to specific dimensions of discourse incivility.

Conclusions:

• A multidimensional approach to incivility offers a promising framework for understanding and intervening in online vaccine discourse.
• Identifying specific emotions associated with incivility can inform the development of targeted interventions for public health professionals.
• Social media platforms can leverage NLP tools like the Google Perspective API for real-time feedback on uncivil health-related content.