WEA - Training Data Sets of Risk Communication and Perception

The goal of this project is to characterize risk perception and response to a designated hazard. The classification of these individual perspectives across the public sphere contributes to the development of a model that evaluates and predicts risk response to an official alert or warning. Cataloging and understanding this response reveals what the public considers important and its priorities in the face of danger, which may differ greatly from the priorities and perspectives of the experts who send the alerts. Such understanding will also allow inefficiencies to be eliminated from risk communications. This approach places emphasis on the decisions the public makes upon receiving an alert or warning and investigates the cognitive factors that influence the decision-making process.

With sponsorship from the U.S. Department of Homeland Security, Science and Technology Directorate, the Pacific Northwest National Laboratory is addressing these challenges by developing a risk ontology that allows for better understanding of risk response and thus an improvement in risk communication between the expert and the layperson. By building this ontology from messages tweeted from real users in real time, it captures actual, untailored responses to an identified hazard. In addition, by mining tweets specific to one hazard, the characteristics of one community can be assessed to not only include geographic location, but also encompass the attitudes, belief system, culture, and demographics that contribute to the community’s perception of risk and thus its response to expert warnings.