Date created: | Last Updated:

Category: Project

Description: A common task with scatterplots is communicating trends in bivariate data. However, the ability of people to visually estimate these trends is under-explored: in cases where visual estimations are systematically different from statistical models, designers may need to intervene to de-bias these estimations, or otherwise inform viewers of the difference between statistical and visual models of trend. We examine the visual estimation of trends, and the potential of mark shape or data distribution to bias these estimates, through a set of laboratory studies and show the results in this project.