An empirical study is presented which explores the use of motion, point size and grey scales for the display of statistical data; specifically, the perception of correlations between variables. The task is to measure the subject's ability to perceive differences between high correlation and low correlation for a two dimensional plot presented in different ways. One variable is always mapped to position with respect to the x axis while the other is mapped to one of six different display parameters depending on experimental condition; namely: position with respect to the y axis, frequency, phase, and amplitude of oscillation, greyscale value of the data points and size of the data points. Human observers appear to be quite insensitive to the relative frequency of moving points but quite sensitive to the correlation of phase angle. Some of the potential advantages and disadvantages involved in using motion for the display of multivariate data are discussed.
Journal or Conference Title
Jun 19 - Jun 23, 1989
London, Ontario, Canada
Limoges, Serge; Ware, Colin; and Knight, William, "Displaying Correlations using Position, Motion, Point Size or Point Colour" (1989). Graphics Interface. 770.