The effect of warning design on subjective risk estimates

This presentation will highlight key findings from a set of decision experiments conducted using a large (n = 5564) sample of the U.S. public. In the experiment, each participant was presented with different graphical and verbal representations of tornado risk, and they made response choices at various distances to a series of fictitious storms. Using this spatial decision setting, several effects became apparent: people inferred uncertainty into deterministic information based on distance; lengthening deterministic warnings caused people closer to the storm to perceive less risk, while people farther from the storm perceived more; simple verbal guidance and explicit estimates of uncertainty both improved decisions in our simulated task; and finally, map design influenced the perception of risk – specifically, using cool colors made people perceive less risk. Implications of findings for warning practice and future research will be discussed.