Trolls' Online Aggression Effects

This paper builds upon the existing literature on Internet trolls to define them as social media users who system-atically express aggression online, and develops a computational model to study the effect of trolls on the aggression level on social media. We perform a series of computational experiments to identify the bounds for the effect magnitude. We also collect data on the online discussions on the Russian social media platform VK from 2020 – 2021 to calibrate the developed model. Our calibration results reveal that the most likely magnitude of trolls' effects on the level of aggression requires a double-size increase in the proportion of trolls for a single-size increase in the proportion of aggressive comments.