Perceived area systematically underestimates actual area

What it is

Conditional compression in area judgment describes conditions in which physical area increases faster than the population-average subjective magnitude. Psychophysics often models this compression with a power exponent below one, while graphical-perception experiments often find area-ratio judgments less accurate than common-scale position or length. The evidence is related but not interchangeable: one estimates a stimulus-to-sensation function, the other answer error in a chart, and neither establishes a fixed underestimate for every person and graphic.

Why it happens

Area requires integration over two dimensions. Readers may recruit diameter, side length, boundary, or reference objects, none of which varies linearly with area. Shape, aspect ratio, fill, neighboring marks, and comparison range alter the available cues; response formats add ratio knowledge and numerical expression. Compression is therefore an empirical pattern produced by interacting perceptual and task processes, not proof that vision always takes a square root or that one exponent transfers across stimuli.

Studying it

Separate direct magnitude estimation, pairwise ratio estimation, ordering, and difference detection. Randomize area ratio, absolute size, shape, aspect ratio, and context; retain individual responses and fit psychophysical functions or error models with intervals. Stevens-style magnitude estimation supplies a power-law framework, Cleveland and McGill compare graphical answer error, and Spence's study of lines, disks, cylinders, and boxes shows that task and stimulus can yield exponents near one even for apparently two- or three-dimensional elements. A replication should report method, scale range, and sample rather than extract one exponent.

Where it stops holding

Compression need not appear on each judgment. Its size or direction can weaken, disappear, or change with reference, range, shape, and method. Ordering by size is generally easier than estimating an exact ratio, although dense, overlapping, or tiny marks can still break order judgments. Labels, legends, and interactive lookup alter the task, so bare-area experiments do not predict the whole interface. Bubble occlusion, map placement, and treemap hierarchy belong to their chart-specific mechanisms.

Applying it

• Decide whether readers need coarse order, ratio estimation, or exact values. Do not leave the latter two to area alone; add a position or length view, values, or queryable detail.
• Keep the shape and aspect ratio of area marks comparable, and avoid decoration, perspective, or irrelevant size changes that add cues unrelated to data.
• Test the target area range at delivery size on the actual task. Inspect response distributions and tail failures by audience and device rather than only the mean.
• Do not apply one psychophysical exponent as automatic “compensation.” It changes literal mapping and can overcorrect another task or reader; validate and disclose it if used.

Related

• Same group: U1.04.2 Mapping values to diameter inflates them quadratically · U1.04.3 3D volume encoding carries the largest error
• Adjacent: U1.01.2 The ranking comes from magnitude-judgment experiments, not design intuition · U2.06.1 Bubbles carry a third variable on the least precise common channel
• Search terms: area perception · magnitude estimation · psychophysical scaling