Just noticeable differences in 2D and 3D bar charts: A psychophysical analysis of chart readability

The comparative readability of 2-dimensional (2D) and 3-dimensional (3D) bar charts was assessed using the method of constant stimuli, and the corresponding Weber constants and just-noticeable-differences were computed. It was predicted that the just-noticeable-difference for 3D charts would be larger than for 2D charts. 36 bar charts (18 2D and 18 3D) were prepared for individual presentation on an overhead projector. Each char contained two bars, one of standard size (25 units) and a second that varied in size from 24 to 26 units in increments of 0.25 units (1% of the standard). 57 undergraduates in psychology were shown the 36 charts in a random sequence for 3 sec. each, separated by 3-sec. intervals. Participants recorded compte size judgements of bars for each chart. Multivariate analyses identify an advantage in accuracy (in the order of magnitude of approximately 10%) for 2D rather than 3D bar charts, after controlling for sex, age, and use of corrective lenses. The computed Weber constants for these judgements were similar to those computed for visual intensity over 100 years ago, but - as predicted - the just-noticeable-difference for 3D charts was larger (implying that larger differences are needed on 3D charts to distinguish closely related bars). Furthermore, the use of traditional psychophysical approaches (such as the method of constant stimuli) in assessing graphical aids would appear to be justified.