On The Correlates of Group-based Emotions of Social Movements in Social Media

This project explores the application of large language models for detecting emotions and predicting the collective actions within the context of three distinct social movements: Fridays For Future, Hong Kong 2019 Social Movement and the Chilean 2019 Social Movement. Using large amounts of text data from Twitter, I examine the relationship between emotions expressed in tweets and the occurrence of violent and non-violent collective actions. The study focuses on four key emotions: anger, fear, joy, and sadness, which I argue are foundational in human perception and motivation. This project attempts to bridge the gap between the large quantities of social media data related to social movements currently available and previous socio-psychological theories of participation in collective actions.
My analysis suggests that anger is the main expressed emotion in the data of all three social movements. However, there is significant variation on what is the most significant second emotion across social movements. Furthermore, several emotions were found to be predictive of either violent or non-violent collective actions, although these relationships are not consistent across social movements. Finally, analysis shows a strong feedback loop, computed using auto-correlation analysis, between reports of collective actions and the occurrence of future collective actions.
These findings contribute to the understanding of how emotions manifest and influence the social movement’s collective actions, and how the reporting of collective actions itself could influence the occurrence of future collective actions.