Yongwen Zheng
Pronouns: she/her
Research Mentor(s): Justine Davis
Co-Presenter:
Research Mentor School/College/Department: DAAS / LSA
Presentation Date: April 20
Presentation Type: Poster
Session: Session 2 – 11am – 11:50am
Room: League Ballroom
Authors: Justine Davis, Yongwen Zheng
Presenter: 65
Abstract
Social media is increasingly more prevalent in the world, but we have not yet systematically studied its effect on election violence. The goal of this research is to expand understanding of the micro-foundations of election violence by examining how partisans in Côte d’Ivoire use social media, specifically Facebook, for election topics related to violence. To do so, we’re using several methodologies such as hand-coding of posts, computation text analysis, and structural topic modeling. We will collect photos from Facebook posts and use pre-trained convolutional neural network models and machine learning techniques to examine those photos. We demonstrate that the discussions of violence-related topics are more prevalent among opposition-supporting groups. Opposition supporting groups are more likely to discuss violence, describe specific episodes of violence, and use hostile, xenophobic, and incendiary language than incumbent supporters. Alarmingly, we find that violence-related topics are associated with more shares on Facebook. This research has shown that leveraging social media can influence violence in elections. The research of polarization in social media can be applied to other countries. The result can help governments and social media companies better understand and regulate social media platforms and avoid cases of misinformation.
Arts and Humanities, Interdisciplinary, Social Sciences
