LAUSR

Abstract We study decisions under different weather warning systems that vary in format and/or information conveyed using a laboratory experiment. Participants have to decide between a safe but costly option (spending to protect from a storm) and a risky option (of not spending for protection). We ran three treatments based upon the severe weather warning system for the UK that the Met Office has been using since 2011 - a risk matrix to communicate the impact and likelihood of an event. In Treatment 1, participants received a colored table with a check in the box of the matrix that showed the likelihood and impact level of the warning. In Treatment 2, participants had the colored table and the color of the warning communicated but without a check in the exact box. In Treatment 3, participants only had the color of the warning communicated without seeing the associated table. Overall our work shows that while increasing the information with content of warnings is usually beneficial and increases the trust in the warning system, it must be done with caution since better decisions (judged by higher profits) are not always made with an increase of information.