Colour-Enhanced Star Plot Glyphs: Can Salient Shape Characteristics Be Overcome?

This paper reports two experiments that address the question of how the shape characteristics of star plot glyphs influence classification tasks and whether additional graphic features such as color can be used to counterbalance the effects of shape characteristics. In a previous study (Klippel, Hardisty, & Weaver, to appear, 2009) we found that salient shapes of star plot glyphs such as “has one spike” influence the classification of the data represented by star plot glyphs compared to star plot glyphs that represent the same data but do not exhibit salient shape characteristics. Hence, we found a similar phenomenon that has been reported for Chernoff faces: that assigning variables to facial characteristics differently can change the outcome of classifications tasks. The shape differences in star plot glyphs are induced by assigning variables to rays in different ways. While the first two experiments showed shape influences, we conducted two follow up studies to shed more light on the influence of shape. A third experiment (the first new experiment) simply addressed the question of how the classification of star plot glyphs would be affected if they are stripped of their meaning, that is, instead of classifying glyphs which represented data, participants were asked to group star plot glyphs as shapes. The forth experiment (the second new experiment) used the same icons as experiment one (from the previous study) and experiment three. This time the star plot glyphs were meaningful; they represented data but additionally, the rays in the star plots were color coded to focus attention on differences in salient shapes. The shape characteristic “has one spike” as a general characteristic cannot be applied as easily as all spikes have different colors. The results show that color enhanced star plot glyphs have positive effects on the processing speed and that they reduce the influence of salient shape characteristics. 1 Accepted by Cartographica, to be published 2009.