The Impact of Color-Coded Uncertainty on Understanding and Decision-Making

Abstract

Currently, severe weather risk is communicated using “Watches” and “Warnings,” although their effectiveness is debated. Indeed, research suggests that explicit numeric probabilities improve people’s understanding of risk as well as the quality of their decisions (Joslyn & LeClerc, 2013). In many applied contexts however, color-coded risk is promoted as more accessible despite minimal evidence supporting this claim. This experimental study compared the effect of three forecast formats – watch & warning, color-coding, and percent chance – on understanding of event likelihood, trust in the forecast, and decision quality. Participants experienced forty virtual storms with the potential to produce tornadoes. For each storm, participants made a series of decisions about taking shelter. Participants made the most frequent cautious decisions in the watch & warning condition. However, decision quality, understanding and trust were highest in conditions with numeric probabilities.