Abstract
Cyberbullying is not a crime, but has significant potential to harm victims’ mental health in the online world as enabled by information and communication technology (ICT). This research in progress aims to derive a predictive mechanism that can protect potential victims from abuse and harm by cyberbullying. The study is based on the collection and processing of 140,000 tweets, and uses a logistic regression model to predict a tendency for cyberbullying based on the manifestation of emotionally charged language on Twitter. Our findings show high potency and statistical significance in the identification of charged language that has the potential to victimize others. The study contributes to a preventative confirmation of cyberbullying, in an effort to provide early warnings for parental mediation and/or mitigation agencies, including school counselors, online bystanders, and law enforcement agencies.
Recommended Citation
Ho, Shuyuan Mary; Kao, Dayu; Chiu-Huang, Ming-Jung; Li, Wenyi; Lai, Chung-Jui; and Ankamah, Bismark, "Charged language on Twitter: A predictive model of cyberbullying to prevent victimization" (2019). WISP 2019 Proceedings. 21.
https://aisel.aisnet.org/wisp2019/21