Aggression and racism on social media increase polarization, thereby threatening the interconnectedness and inclusiveness of a network society. In this project members of the Institute for Societal Resilience work together with members of the Computational Lexicology & Terminology Lab of the VU to detect aggression and racist discourse on social media by means of automated content analysis. Proxies for aggression in social media content may include, for example, the use of terms of abuse (i.e., swear words). Similarly, detecting racism on social media relies on developing a dictionary containing different word categories specifically related to racist discourse. The project goes beyond the mere development of automated methods for detection by investigating environmental factors (e.g., societal events) that may predict aggression and racism on social media.

Team: Jaap Ouwerkerk, Ivar Vermeulen, Antske Fokkens, Peter Kerkhof