Data encoded by area and volume can be systematically underestimated
Aliases: volume perception · graphical encoding · bubble chart
What it is
When quantity is encoded by circle area, bubble size, or three-dimensional volume, observers often recover ratios less accurately than when comparing length, especially underestimating differences at large area or volume. This area perception bias is related to non-linear psychophysics: geometrically correct scaling can still look like a smaller difference.
Why it happens
Length can be compared along one dimension; area and volume require integration of several dimensions, while outline, occlusion, and perspective interfere. Subjective mapping often compresses broad area ranges, so doubling area need not feel doubled. Volumetric graphics further invite estimation from a visible two-dimensional projection.
Studying it
Ask people to compare ratios or estimate multiples across length, area, and volume encodings; measure error, time, and ordering consistency. Control colour, position, overlap, and labels; vary viewpoint for 3D forms. Evaluate with real comparison tasks rather than visual appeal alone.
Where it stops holding
Area is not unusable: with few categories, large differences, numerical labels, or coarse tasks it can work well. The risk is precise ratio judgment or closely spaced values. Experience, animation, and interactive queries can alter performance, so the bias is not an unfixable fate.
Applying it
- For precise comparison, prefer position or length on a common baseline over bubble area, pie volume, or 3D bars.
- When area is required, add values, ordering, reference lines, or interactive highlighting so size does not carry all reading responsibility.
- Test close and wide-range values, checking whether people systematically compress large-value differences.