Abstract
Subjects sorted decks of cards depicting pie charts and divided bar graphs on two criteria: the proportion shown in the graph, and the graph's overall size, or scaling. Sorting times and errors were measured. For divided bars, performance was impaired when subjects were required to sort the proportion and the overall scaling varied. No such impairment occurred for pie charts. The results suggest that proportion and scaling are integral dimensions for divided bar graphs, but separable dimensions for pie charts. Subjects can judge angles or slopes with pie charts having different scaling, but must estimate a ratio prior to classification with different-scale divided bars. In sum, showing proportions with divided bar graphs can be problematic if the scaling of the graph varies, but pie charts are not similarly affected.
Get full access to this article
View all access options for this article.