Alignment,Scaling,and Size Effects in Discrimination of Graphical Elements

Abstract

Recent work in graphical perception has attempted to identify the mental operations used by an observer when extracting information from a graphical display (e.g., Hollands and Spence, in press; Simkin and Hastie, 1987). The current research varied the alignment, scaling, and size of proportions shown in pie charts and divided bar graphs. Subjects were required to discriminate between two proportions (i.e., which proportion is larger?), each shown relative to its own whole. Response times and errors were measured. Results from Experiment 1 show that for both pies and divided bars, the time penalty for discriminating unaligned proportions was dependent on the size difference between the two proportions, with a greater penalty with a smaller percent difference. Results from Experiment 2 show that different scaling slowed subjects considerably, especially when the size difference was small, and especially with divided bars. The results are interpreted in terms of hypothesized alignment, scaling, and discrimination operations. The practical implications for graphical design are also discussed.