Events

Public defence in Computer Science, M.Sc. Yan Xia

A data-driven look into political (de)polarization on social media shows the depth and resilience of divides

Public defence from the Aalto University School of Science, Department of Computer Science.
Two opposing human profiles formed from clusters of 0s and 1s, with streams of grawlix emerging from their mouths.
Abstract illustration of polarized discourse on social media. © Yan Xia

Title of the thesis: Understanding the (de)polarized social media

Thesis defender: Yan Xia
Opponent: Professor Taha Yasseri, Trinity College Dublin, Technological University Dublin, Ireland
Custos: Professor Mikko Kivelä, Aalto University School of Science

Political polarization poses a serious societal challenge around the globe, which has only grown more pronounced with the rise of digital communication. While extensive social science research has probed the phenomenon in theoretical or experimental settings, important questions remain: why does polarization naturally persist and deepen in online debates? Does polarization ever decrease without deliberate intervention, and if so, how?

This thesis draws on mixed-methods analyses of observational social media data to examine dynamics of (de)polarization in real-world contexts. The results reveal that polarization in online political discourse can be amplified through the viral spread of polarizing themes that enhance ingroup bonds and preclude outgroup engagement. Meanwhile, depolarization following unobtrusive cross-group interactions is limited, and even a substantial external threat depolarized only a subset of actors on a subset of arguments.

The findings depict a system of online discourse in which polarization is deeply entrenched, self-reinforcing, and resilient to depolarizing forces. Despite this pessimistic landscape, the thesis outlines pathways toward depolarization by advancing a computational framework that disentangles layers of information diffusion in polarized debates, and by identifying depolarization mechanisms in crisis contexts that may also operate under ordinary conditions. Further, the thesis makes a methodological contribution by evaluating the collection of methods employed, and illustrating how they can be triangulated and integrated to produce reliable findings in computational social science.

Keywords: polarization, social network

Contact information: yan.xia@aalto.fi 

Thesis available for public display 7 days prior to the defence at Aaltodoc

Doctoral theses of the School of Science

A large white 'A!' sculpture on the rooftop of the Undergraduate centre. A large tree and other buildings in the background.

Doctoral theses of the School of Science at Aaltodoc (external link)

Doctoral theses of the School of Science are available in the open access repository maintained by Aalto, Aaltodoc.

Zoom Quick Guide
  • Updated:
  • Published:
Share
URL copied!