Why optical illusions work: size, length and brightness depend on context

Start by holding the central object still

The Ebbinghaus illusion compares central circles surrounded by differently sized neighbours. The important comparison is the target circle, not the diameter of the entire arrangement. If both target and context change at once, it becomes difficult to tell which change explains the difference a viewer reports.

In the Ebbinghaus matching activity, the left central radius stays at twenty drawing units while the right is adjustable. Remove the surrounds after choosing a match and inspect the recorded physical difference. A primary study of geometrical illusions provides research context for comparing perceptual effects. The board is a self-adjustment illustration, not a replication of that study or a calibrated measure of your visual system.

Distinguish a shaft from the fins around it

The Müller–Lyer illusion concerns apparent shaft length under different fin arrangements. Overall decorated extent and the horizontal segment between endpoints are different quantities. A diagram can encourage a reader to judge the former while a caption claims to compare the latter, creating a mismatch between the intended encoding and the perceived object.

Use the Müller–Lyer experiment to adjust the lower shaft before hiding its fins. The upper segment remains 240 units long. If your chosen lower segment differs, the reveal records that difference rather than secretly changing the line. An exact physical match is also a legitimate response. Visible numerical controls, unlimited viewing and uncontrolled screen geometry prevent this small activity from estimating a universal illusion magnitude or diagnosing perceptual ability.

Worked example: test a pair of chart markers

Imagine a dashboard using two circular markers to represent equal values. One sits among tiny marks and the other among large marks. The following is a proposed design review, not an empirical claim that all users will make the same error.

1. Render both target markers with the same physical radius and verify that equality in the drawing code. Preserve a copy with the surrounding marks visible so the comparison reflects the intended layout.
2. Ask a reader to compare apparent size, then remove or standardise the surrounding marks without changing the targets. Record the reader’s judgment separately from the known dimensions rather than treating either as a substitute for the other.
3. Repeat the review in the actual panel size, spacing and theme. If the target is a precise comparison, add direct labels or a simpler common-scale encoding rather than expecting context to become irrelevant after an explanation.

Equal channel values do not guarantee equal appearance

The simultaneous brightness contrast board places patches on dark and light surrounds. The reference uses RGB channels of 128; the comparison can be adjusted. Equalising the surrounds preserves those channel values, letting you distinguish a physical drawing specification from an apparent match. The displayed numbers are not calibrated luminance readings.

This distinction matters when a design uses colour to suggest status, emphasis or equivalence. A swatch reviewed on white can behave differently inside a dark panel. Display settings and ambient lighting add further variation. Test the actual arrangement and use redundant labels when interpreting colour correctly matters. A perceptual demonstration does not replace established accessibility checks, nor should a viewer’s subjective match be presented as a monitor calibration.

A practical context review for visual comparisons

Treat a context review as an investigation with an explicit target. Otherwise a redesign can accidentally change scale, spacing, contrast and meaning at once, leaving no clear explanation of what improved. Preserve the source values, identify the visual judgment a reader must make and test that judgment in the complete interface rather than in a detached asset preview.

• Specify the physical quantity being encoded and the apparent comparison the reader needs to make.
• Keep the target fixed when testing how surrounding size, decoration or brightness changes its interpretation.
• Inspect both light and dark contexts at the actual display dimensions and expected reading distance.
• Use direct values, shape or text when a precise interpretation should not depend on a colour impression alone.
• Record observations as observations; do not label a small informal review a validated perceptual experiment.

Context also changes what counts as the object

Sometimes the issue is not an object’s apparent size or brightness but which region becomes the object at all. Figure–ground organisation can assign a shared boundary to a central vessel or to adjacent profiles. OpenStax’s Gestalt chapter introduces this organisation of visual scenes. It provides a different question from matching two measured lengths.

The related figure–ground activity uses an authored reversible shape and attention cues. Its buttons change contrast as well as instructions, so a click is not proof that an involuntary perceptual reversal occurred. That limitation is useful: an honest design demonstration separates what the interface changed from what the reader experienced. Across all these examples, the transferable habit is to keep drawing values, task instructions and perceptual judgments individually inspectable, then consider how they interact in the real layout.