DECISIONS, VARIABILITY, AND VISUALIZATION: A NOVEL INSTRUMENT FOR DECISION MAKING UNDER VARIABILITY

To make statistically sound choices, decision makers must respond to the consequences of variability—to target variability. Previous work has shown that targeting is beneficial, but there are questions about its prevalence and contributing factors. With a novel instrument and a demographically representative, randomly assigned n=306 sample, we find that most U.S. adults target variability by default, even with minimal information (95% CI [78%, 92%]). Surprisingly, the same information presented via a standard bar graph inhibits targeting, while showing raw data enhances targeting. These results support a resource view of statistically sound decision making (over a deficit view), and contradict widespread ideas in data visualization (that raw data will overwhelm).