RAPID: Panic & Pandemic: Information, Diffusion Dynamics, and the Management of Public Health Crises

Ohio State University, The, Columbus OH

Investigators

Abstract

During a public health emergency such as the Ebola outbreak, effective management of the information environment is critical not only for preventing public panic but also for getting citizens to receive and retain information about the disease that can help to alter their behavior, thus lowering their risk of infection and ultimately helping to halt the spread of disease. It is widely established that information coming to individuals via their social networks has a strong eect on their beliefs, attitudes, and behavior. To examine the complex interplay between ocial (e.g. from CDC) and social information, this experiment will manipulate both the amount of social information subjects receive and also the social topology network in which they are embedded. The micro-level processes identied by this experiment will then be used to simulate the interplay of official and social information on the population level using agent based models. The study will (1) provide a precise time-series of information and attitude spread; (2) provide a precise distribution of the number of social reinforcements necessary for an individual to update their beliefs; (3) provide a clear measure of the eect that the volume of communication from public health agencies has on individual attitudes; (4) provide a clear measure of the eect that communications from health agencies have on public trust; (5) provide a clear measure of the eects of network structure and density, having the possibility prescribing dierent informational interventions in dierent network topologies. Finally, the agent-based simulations conducted using the experimental results can provide public health managers with an effective way to (6) forecast the spread of information, (7) anticipate the population-level effects of theirinformational interventions, and (8) understand and manage the spread of panic amidst a possible pandemic. This experiment follows a 3 by 4 design in which the participants are randomly assigned into groups of high, low, or random social clustering as well as high, low, or no exposure to Ebola-related information from a public health source. Each type type of treatment group will be run for two trials, meaning that the diffusion-of-information process will be performed, observed, and repeated a total of 24 times. The experiment will be conducted online and run for three weeks. Subjects will log in daily to a dedicated website where they will receive Ebola information from a public health agency and be able to see their social contact's attitudes related to the Ebola virus (e.g. their degree of concern, their trust in the government and public health responses, their answers to several factual questions about the disease and short Twitter-like posts about their views). On a daily login, subjects will view any informational messages they receive (as determined by treatment category) and read the opinions proered by their social connections. This design allows one to track the diffusion-of-information process in great detail over the social network as well as produce an understanding of how and when information from a public health source can change individuals' attitudes and beliefs. The results from the controlled experiment can be used to simulate the spread of information among the general public.