Abstract:
We identify critical issues in comparative research on relational information displays (RIDs). The key argument is that when conducting an analysis of the cognitive process of people viewing different displays, their perceptual processes must be held constant so that they do not affect the results. We propose that in order to help researchers more easily compare display types (e.g., graphs) for how effectively they convey information, two factors must be considered. First, each element (e.g., each bar in a bar graph) in graphs that are being compared has to be equally discriminable. Second, the number of elements in the graphs being compared has to be the same; the maximum number of elements is limited by from the graph that uses a presentation format (e.g., density) that has the fewest number of discriminable levels. We present a psychophysics experiment that identified differential discrimination thresholds for density levels.

Keywords: relational information display, human factors, psychophysics

You can access this technical report via: PDF