Visual channels have a stable empirical accuracy ranking

What it is

A conditional empirical ordering of visual channels means that controlled quantitative comparisons produce different error distributions for different encodings. Cleveland and McGill found common-scale position more accurate than several nonaligned or nonspatial judgments, and later experiments replicated important relative differences. The ordering is not an unconditional human constant. It compares elementary perceptual tasks in particular graphics, under a specified task, stimulus construction, value range, readership, and display.

Why it happens

A graphic maps values to position, length, angle, area, luminance, or another visible property, which a reader decodes into a relation. Shared reference supports direct comparison; absent baselines, integration over area, and luminance discrimination introduce different error. An account based on “more mental steps” is only a possible mechanism, not evidence by itself. The same nominal channel can perform differently in bars, pies, rectangular areas, or small charts with grids, so relative accuracy and uncertainty are more defensible than a memorized universal chain.

Studying it

Specify the compared objects, response, data distribution, chart family, and display, then randomize stimuli and estimate error, time, and participant variation by condition. Classic ratio judgments can replicate position, length, angle, and area conditions. Heer and Bostock replicated some classic results in a less controlled crowdsourced environment, while also raising participant-quality and display-control issues. Report effect sizes and intervals, not rank alone; a novel mark does not inherit an old result merely because it resembles a named channel.

Where it stops holding

The ordering primarily addresses numerical comparison or ratio estimation, not trend, anomaly, correlation, memory, aesthetics, or decision quality. Direct labels, interactive lookup, and training change the task. Color vision, low vision, cognitive strategy, and visualization literacy can change performance. Recent replication finds group-level patterns compatible with classic results while some individuals credibly depart from the average ordering. “Stable” therefore means a reproducible population tendency, not identical performance for every reader.

Applying it

• State whether readers compare, order, estimate, or recognize shape, then seek evidence matching that task, graphic, and audience.
• For precise comparison, test common-scale position or clear length first. If angle, area, or luminance carries a critical magnitude, add labels, interactive lookup, or a more direct alternative view.
• Record the conditions supporting a channel choice instead of turning a textbook order into an exception-free standard. Test unfamiliar conditions with a focused task study.
• Evaluate error, time, and accessible alternatives at delivery size with intended readers. When reproducible local results conflict with the heuristic, update the design choice.

Related

• Same group: U1.01.2 The ranking comes from magnitude-judgment experiments, not design intuition · U1.01.3 Quantitative and categorical data need two different channel rankings · U1.01.4 The most important variable should get the highest-ranked available channel · U1.01.5 The ranking is an accuracy ceiling; alignment and mark size erode it
• Adjacent: U1.02.1 Position along a common scale is the most accurate channel · U1.03.2 Angle judgments are markedly less accurate than length under equal conditions
• Search terms: graphical perception · Cleveland McGill · visual channel accuracy