Then, we reveal that a small minority of extremist stubborn users significantly drive opinion polarization into echo chambers even faster with prioritization, due to the interplay of several determinants of the dynamics. 3/4

We extend an existing opinion polarization model on adaptive networks to feature two new types of agents: stubborn and priority. We also introduce a novel measure of polarization combining bimodality of both agents and their average neighbour opinion. 2/4

What happens when a policy changes but the data is no longer available? We need to look at the mechanisms instead. In our newest pre-print (arxiv.org/abs/2312.09358), we study the effects of inadvertent social media content promotion on polarization & formation of echo chambers 1/4