Toxicity is rampant on social media platforms. In this study, we use COVID-19 datasets discussing the use of face masks, vaccines, 5G, and Bill Gates conspiracy theories to examine the contagiousness of tweets with hateful or toxic content. While the majority of people believe that face masks or vaccines are effective in fighting against COVID-19, a smaller minority do not. Another major issue during the pandemic was the conspiracies surrounding 5G technology and Bill Gates. Studies indicate that the worries surrounding mobile technology have fueled certain conspiracy theories linking 5G and Bill Gates with the COVID-19 virus. Using these polarizing datasets, we measure the type and intensity of hate speech in each dataset. Then we propose a definition for a toxicity contagiousness score to study the propagation of toxicity in each dataset. Our study revealed that in the 5G dataset, low toxic content has a positive correlation with the contagiousness score. Highly toxic content in the Bill Gates and pro-mask datasets also showed a positive correlation. However, in the anti-community dataset, such as anti-mask and anti-vaccine, highly toxic content had a negative correlation with the contagiousness score. These findings shed light on how different types of content and contexts can influence the spread of toxicity in online communities.

PDF