Abstract
People struggle to reason about uncertainty, yet they must make decisions based on uncertain information in daily life as well as in high-stakes situations. One method of dealing with uncertainty is through data visualizations, which can help to summarize complex data. Though the importance of visualization design has been established, there remain questions about how certain aspects of design-for example, hue and lightness-affect interpretations, especially when uncertainty is present. We investigated the effects of hue and lightness variation on participants' decision-making when viewing hazard maps that depicted wildfire risk. In three experiments, participants were shown a series of hypothetical wildfire risk maps that varied in lightness or in hue, and were asked to decide whether to evacuate based on these maps. We found that lightness variation within and between maps affected decision-making such that people chose to evacuate at higher rates when maps varied in lightness and when they were darker overall (Experiment 1). The hue of maps varied, and people chose to evacuate more often when lightness-varying maps were red than when they were blue (Experiment 2). When we removed borders around the risk bands within the maps, evacuation rates increased for darker maps and decreased for lighter maps (Experiment 3). Taken together, our results reinforce the notion that visualization design can affect not only interpretations of risks, but also the decisions that result from those interpretations.