Self-Effects on Social Media and Political Polarization

Important results from the project

The first objective (to identify the social media users and issues that indicate polarization on social media) was fulfilled through an analysis of Facebook during the 2016 Brexit referendum and an analysis of reactions to US politicians’ images across Facebook and Instagram. The second objective (to identify how platform design affects polarization) was fulfilled through an analysis of how the user networks and algorithms of platforms affect citizens’ political expression. The third objective (to communicate the results of the project) has already been mentioned.

Expected long term effects

First, we found that users who express polarized opinions are small in number but leave the majority of political comments on social media. They also actively express their political opinions in spaces designed for their political opponents. Second, we found that users across Facebook and Instagram react differently to the emotions in politicians’ faces, indicating a generation polarization in what values citizens prefer in their political leaders. Third, we found that platform design shapes the relationship between political ideology and political expression across platforms.

Approach and implementation

The project was designed and implemented through six studies that tested social media and polarization in different ways. These are: the computational analysis of Facebook comments, computer vision on Facebook and Instagram political ads, a study of reactions to images on Facebook and Instagram, two behavioral economics experiments testing the effect of emotions on political expression and affective polarization, and a large-scale survey of political expression across four countries and seven platforms.