RAPID: Capturing Behavioral Response and Perceived Risk to Ebola Using Social Media

Kent State University, Kent OH

Investigators

Abstract

Affirming the fact that infectious disease respects no geographic boundary, residents of northeastern Ohio, specifically the Kent State University community, were faced with a potential Ebola exposure during September 2014 when nurse Amber Vinson visited relatives who worked on campus. Students, staff, faculty, and nearby residents responded to the potential risk of Ebola, like many people around the world, by changing their behavior based on available information and their perceived risk of infection. This Grant for Rapid Response Research (RAPID) project will investigate these mechanisms and improve our understanding of information spread through social media and how it translates to on-the-ground perceptions and behaviors. This is vital if social media data are to be used to enhance situational awareness in emergency response and disease surveillance systems. Within the context of these systems, social media may provide real-time information to decision makers, emergency responders, and health officials to speed the process of detection, prevention, control, and community action, thus benefiting society as a whole by reducing the health and economic impact of major risk events including epidemics, natural disasters, or human-caused disasters. In addition, this project will help address disaster related social vulnerability by identifying how information barriers may disadvantage certain social groups. Specifically, this project will compare and contrast the behavior response of those inside a closed system, such as a university or hospital, to those living outside that system. Expected results will help inform how to improve equity in critical information flows across social systems to eliminate such disparities. Multiple disciplines now regard social media as a lens into social communication, belief, and action. Yet, little research has been conducted to validate the potential and pitfalls of these data, in particular, how social and geographic variation between groups impacts the potential utility of social media as a data source. This transformative research will consider both these aspects within the framework of the Protective Action Decision Model for risk communication. By examining the linkages between the flow of information and the change in behavior in response to a major risk event, this project addresses an unresolved fundamental question: are data from the social media platform Twitter representative of on-the-ground perceived risk and behavioral response? This project addresses this question by validating Ebola-related tweets on and around campus with two timely surveys fielded to almost 60,000 people designed to collect risk perception and behavioral response data. Expected results will not only improve situational awareness for various scenarios including epidemics, and natural or human-caused disasters, but also advance theory surrounding each by validating social media as an appropriate data source. We will advance understanding of communication flow across socioeconomic, institutional, geographic, and physical-virtual boundaries for an at-risk community. Outcomes will be used to inform theoretical development across the disaster and health sciences, and general risk communication using social media by revealing the utility, nuance, and limitations of social media as a data source.