Identifying Bullies with a Computer Game

Current computer involvement in adolescent social networks (youth between the ages of 11 and 17) provides new opportunities to study group dynamics, interactions amongst peers, individual preferences. Nevertheless, most research this area focuses on efficiently retrieving information that is explicit large (e.g., properties graph structure), but not how use dynamics virtual network discover latent characteristics real-world network. In paper, we present analysis a game designed take advantage familiarity adolescents with online networks, describe data generated by can be used identify bullies 5th grade classrooms. We probabilistic model using in-game players (i.e., content chat messages) infer their role within classroom (either bully or non-bully). The evaluation our done previously collected from psychological surveys same population comparing performance off-the-shelf classifiers.