Troll and divide: the language of online polarization.

The affective animosity between the political left and right has grown steadily in many countries over the past few years, posing a threat to democratic practices and public health. There is a rising concern over the role that "bad actors" or trolls may play in the polarization of online networks. In this research, we examined the processes by which trolls may sow intergroup conflict through polarized rhetoric. We developed a dictionary to assess online polarization by measuring language associated with communications that display partisan bias in their diffusion. We validated the polarized language dictionary in 4 different contexts and across multiple time periods. The polarization dictionary made out-of-set predictions, generalized to both new political contexts (#BlackLivesMatter) and a different social media platform (Reddit), and predicted partisan differences in public opinion polls about COVID-19. Then we analyzed tweets from a known Russian troll source ( = 383,510) and found that their use of polarized language has increased over time. We also compared troll tweets from 3 countries ( = 79,833) and found that they all utilize more polarized language than regular Americans ( = 1,507,300) and trolls have increased their use of polarized rhetoric over time. We also find that polarized language is associated with greater engagement, but this association only holds for politically engaged users (both trolls and regular users). This research clarifies how trolls leverage polarized language and provides an open-source, simple tool for exploration of polarized communications on social media.