Trust Definitions and Metrics for Social Media Analysis

Social media has changed the information landscape for a variety of events including natural disasters, demonstrations, and violent crises. During these events, people use a variety of social media, such as Twitter, to share information with the world. Given the massive amount of data generated, it is difficult to identify the valuable information in a sea of noise. In this study, we focus on a universal contributing factor to information value—trust— which is analyzed in three steps. First, a theory of trust in information is developed based on the application of existing literature to this unique and emerging issue. Second, a set of metrics is developed that focus on trusted relationships and behavioral indicators of trustworthiness within social media. Third, these trust metrics are tested on an anonymized data set and their results presented. ACKNOWLEDGEMENTS: This work was performed under US Navy contract number N00014-12-M-0259. The authors thank Dr. Rebecca Goolsby for her significant technical support and eager engagement on this project. This work was funded in its entirety by the Office for Naval Research (ONR). We also acknowledge the contribution of Professor Adam Henry of the University of Arizona who introduced the authors to the concept of trust in information.