Social Networks and Bullying Research

Bullying among students at a school has become a serious social problem. We (and our collaborators) consider school bullying using digital communication media (such as social networks, cell phones, short messaging systems, emails, blogs) and create a sociology-based network framework to help teachers and parents identify whether school bullying may exist among students.

The framework models interactions among students as a relationship network, constructs a relationship network from usage statistics of digital communication media, and identifies whether unique structural features exist in a relationship network that may indicate bullying among students. In doing so, the framework considers unique properties of digital communication media (e.g., recordability, anonymity, availability) that are different from those of non-digital communication media (such as letters and verbal communication) and that make it highly difficult to detect school bullying.

The framework is designed based on two key hypotheses: (1) school bullying imposes distinct structural features in a relationship network, and (2) a relationship network is constructed with some degree of accuracy from usage statistics of digital communication media without examining privacy-violating information. The framework only uses usage statistics that maintain privacy of communication, and it extracts such usage statistics from those publicly available, those collected by and traditionally made only available internally to a service provider, and those that the framework directly monitors.

We (and our collaborators) work with researchers in sociology and social psychology to empirically verify the hypotheses used in the framework, design the framework and empirically examine the framework. 

As the first step towards creating the framework, we and our collaborators developed an automated tool to analyze online posting of aggressive text messages to threaten a victim [1].  This tool identifies features of text messages which may be indicators of school bullying. Such features include the role of individual posting the message (e.g., a bully, victim, or defender), whether the text message contains insults directed at a victim, and whether the communication is anonymous. The developed tool was used to examine an online forum to show its effectiveness.

Selected Papers in Social Networks and School Bullying

[1] M. Moore, T. Nakano, A. Enomoto and T. Suda, “Anonymity and Roles Associated with Aggressive Posts in an Online Forum,” submitted to Elsevier Journal of Computers in Human Behavior, 2011.