Perception of Histograms in the Polar Coordinate System (Rose Plots) With Frequency Represented by Length or Area

Histograms are often visualized in a polar coordinate system (PCS) by representing each bin as a sector. This form of visualization introduces distortions that depend on whether frequency is represented by sector length (radius) or area. The present study evaluated the effects of such distortions on perceptual bias and accuracy. Participants were instructed to estimate the ratio between the two halves of a distribution using either aggregate length or area as a visual cue. Results showed that aggregate length-based reading with unlimited viewing time led to systematic overreporting and low accuracy, but that without cue aggregation did not. In contrast, area-based reading resulted in underreporting. The results were broadly consistent across different types of distributions. These findings highlight the perceptual advantages of equal-area forms for rose plot histograms, while even these forms may not be immune to perceptual biases.