The impact of increasing COVID-19 cases/deaths on the number of uncivil tweets directed at governments

Nishi, Kohei

Political expression through social media such as Twitter has already taken root as a form of political participation. However, it is less clear what kind of political messages people send out on social media and under what circumstances they do so. This study theorizes that when government policy performance worsens, people get angry or frustrated and send uncivil messages to the government. To test this theory, the current study classifies tweets directed at U.S. state governors as uncivil or not, using a neural network machine-learning model, and examines the impact of worsening state-level COVID-19 indicators on the number of uncivil tweets directed at the state governors. The results show that increasing state-level COVID-19 cases and deaths lead to higher numbers of uncivil tweets directed at state governors. This suggests that people evaluate the government's performance through actions other than voting.