Display Format and the Perception of Numerical Data

Abstract

This paper presents two experiments which test the principle of compatibility of proximity. This principle states that when a task requires the integration of information, the display should present the data in close proximity. When, however, the information must be selectively attended or processed, the information should be displayed with low proximity. In the first experiment, subjects were shown two dimensional data for two hypothetical conditions and asked to extrapolate to a third condition. Performance on the task was more accurate when the data were displayed on a coordinate axis graph than when a bargraph display was used. The second experiment demonstrated that when cues in a multi-cue judgment task had to be selectively attended, performance was better with a bargraph display than with an integrated triangle display. No difference was observed when all the cues had to be attended. Discussion centers on display format and perception of relations between numerical data.