M.A. Keijzer, MSc
Will the personalization of online social networks intensify opinion polarization?
Personalization has dramatically changed the way information is distributed online. In order to accommodate the needs of their users, search engines, webshops, and online social networks provide information tailored to their individual users. Browsing a web personalized to your needs and tastes eases the process of navigating relevant content within the vast amount of information available online. Scholars and pundits have warned that browsing a web of content made up by your past likes creates the possibility of getting stuck in a 'filter bubble' or 'echo chamber'. This, in turn, could create polarization of opinions on a macro scale. The opinion dynamics literature, however, leaves a great deal of uncertainty on the consequences of web personalization. The two most prominent approaches in the field actually make opposite predictions about the effect of personalization on the macroscopic distribution of opinions. In this project we aim to explore when and why personalization leads to polarization, and subsequently develop minimally invasive methods to prevent polarization. We develop agent-based models of online interaction and test the most crucial assumptions for its micro foundations using laboratory - and online experiments.
|Last modified:||15 August 2017 2.31 p.m.|