Quantifying Political Polarization through the Lens of Machine Translation and Vicarious Offense

Abstract This talk surveys three related research contributions that shed light on the current US political divide: 1. a novel machine-translation-based framework to quantify political polarization; 2. an analysis of disparate media portrayal of US policing in major cable news outlets; and 3. a novel perspective of vicarious offense that examines a timely and important question -- how well do Democratic-leaning users perceive what content would be deemed as offensive by their Republican-leaning counterparts or vice-versa?