This thesis studies the behavior of the public under the threat of an extreme event with the use of laboratory and on-line experiments. Chapter 2 and 3 are about decision making in the possibility of a natural climate disaster. Chapter 4 is about decisions in an infectious disease pandemic. Chapter 2 relates to the impact of different warning systems on public risk preferences during sudden natural disasters. It shows that while increasing the information content of warnings is usually beneficial and increases 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 in information. Chapter 3 is about changing the mode and content of information distribution influences the reaction time of the public. We have conducted research on this topic and found that adjusting the emergency warning system can reduce unnecessary waiting by the public, accelerate their response time, and thus reduce unnecessary casualties and property losses. Chapter 4 is about the influence of an early warning system and information intervention on public behavior in major public health events (such as COVID-19). This result shows that the public act selfishly during the pandemic but can be influenced by suggestions. This finding explains why many countries have been mired in the effects of the pandemic, with massive job losses and labor shortages.IFREE - International Foundation for Research in Experimenta