Differences of Opinions: How Visualizations of Uncertainty in Ratings Data Affect Choice

When people make decisions based on data, they often disregard variability in the data and place disproportionate emphasis on central tendency. However, visualizations differ in how they depict variability, and in some tasks and domains, research has also observed that people can be sensitive to variability, such as with customer ratings. In the current study, participants were randomly assigned to different visualizations of ratings data, and we examined how variability in the ratings data affected participants' decisions between two options. Results showed that participants were sensitive to variability in ratings data when presented through graphical visualizations but not when presented numerically. The histogram condition produced the strongest effect, with participants showing markedly stronger preferences when variability was low. These effects were not moderated by participants' statistical training, decision domain, nor domain expertise. We explore the implications of these findings for people's perceptions of uncertainty during decision making tasks.