Polarization in Social Media: A Virtual Worlds-Based Approach

Jacob, Dennis and Banisch, Sven (2023) Polarization in Social Media: A Virtual Worlds-Based Approach. Journal of Artificial Societies and Social Simulation, 26 (3). ISSN 1460-7425

[thumbnail of 11.pdf] Text
11.pdf - Published Version

Download (4MB)

Abstract

As social media becomes increasingly integrated within the fabric of our digital lives, it is clear that these platforms have a great impact on our mental well-being and interpersonal relationships. However, recent events and studies suggest that these changes are not always for the better as social media might contribute to social polarization. In this work, we leverage agent-based modelling (ABM) techniques to simulate the associated opinion dynamics of polarization in social media platforms. To accomplish this, we first develop a methodology for distinguishing between different types of polarization. This enables a more nuanced investigation into the interplay between behavior online and behavior offline. We next expand on the public-private split model by introducing a novel “virtual worlds” framework for representing an online social media platform. Agents from the neighbor constrained “real world” can “log-in” to these virtual worlds with a certain probability and participate in a complete network; this reflects the unique socioeconomic and geographic anonymity provided through social media. Additionally, global homophilic influence is incorporated and its relationship with local virtual world structure is considered. We finally perform a sensitivity analysis over a set of model parameters, and find that the incorporation of virtual worlds can result in the simultaneous presence of different types of polarization in the real and virtual worlds. These findings align with studies on social media from the literature, and suggest that the online platform provided by social media poses unique challenges with regards to investigating the presence of polarization.

Item Type: Article
Subjects: Universal Eprints > Computer Science
Depositing User: Managing Editor
Date Deposited: 12 Oct 2023 05:16
Last Modified: 12 Oct 2023 05:16
URI: http://journal.article2publish.com/id/eprint/2335

Actions (login required)

View Item
View Item