Graphs have been an essential tool for the analysis and communication of statistical data for about 200 years. Despite widespread use and their importance in science, business, and many other walks of life, relatively little is known about how people perceive and process statistical graphs. This article reviews several empirical studies designed to explore the suitability of various graphs for a variety of purposes, and discusses the relevant theoretical psychological literature. The role of traditional psychophysics is considered, especially in connection with the long-running dispute concerning the relative merits of pie and bar charts. The review also discusses experiments on the perception of scatterplots and the use of multivariate displays, and points out the need for more empirical work.
