Test Methodology & Limitations

Last updated: 2026-09-30

Purpose

ColorVision provides original, browser-based activities for learning about colour vision and for noticing possible colour-discrimination differences. They are educational screening activities, not reproductions of proprietary clinical test materials and not a medical diagnosis, certificate, or occupational assessment.

How the activities work

The digit-plate activity uses original dot patterns generated in the browser. The hue-order activities ask a participant to arrange digital colour samples between fixed references. Results are calculated from responses made in that browser session and are shown as educational feedback.

Conventional Ishihara-style pseudoisochromatic plates are designed primarily to screen red-green colour-vision differences. Conventional Farnsworth arrangement tests use standardised physical materials and controlled viewing conditions. ColorVision does not claim that a browser display reproduces those conditions.

Why online results are limited

• Displays vary in calibration, brightness, gamut, colour management, and viewing angle.
• Ambient light, glare, blue-light filters, accessibility filters, and browser settings can alter what is seen.
• Visual acuity, fatigue, instructions, and response conditions can affect an activity result.

For these reasons, do not use a ColorVision result for medical, licensing, employment, school-placement, or legal decisions. If you notice a new change in colour vision or receive a result that concerns you, consult an optometrist or ophthalmologist.

Sources and review

The educational descriptions are maintained by the ColorVision Editorial Team and link to the underlying sources on each guide. The content has not been individually reviewed by a named licensed clinician unless a page explicitly says otherwise. We update this page when the activity design or its stated limitations materially change. See tool capabilities and limitations for the inputs and outputs of every browser-based tool.

• National Eye Institute: Color Blindness
• National Academies / NCBI Bookshelf: Procedures for Testing Color Vision
• NCBI Bookshelf: The Perception of Color

Exact implementation used by ColorVision

The descriptions below document this site’s code, not the scoring rules of the clinical tests that inspired the general activity formats.

24-plate digit activity

• Every plate is drawn at run time on a 380 × 380 canvas. A text mask separates figure dots from background dots.
• The renderer attempts to pack up to 1,300 non-overlapping circles with radii from 3 to 10 pixels. Dot colours are randomly selected from the activity’s fixed foreground and background palettes, with a small proportion swapped as visual noise.
• The displayed match percentage counts exact responses across all 24 plates. The plain-language session summary excludes the two control plates and groups the remaining response-difference rate into under 12%, 12–39%, or 40% and above.
• Those bands are product wording thresholds, not sensitivity, specificity, diagnosis, or clinical severity thresholds. The plates are regenerated and can look different on another attempt.

40-cap hue-order practice

• The activity generates 40 HSL colours at nine-degree hue intervals with 72% saturation and 52% lightness. The first and last colours are fixed; the other 38 are shuffled.
• For each movable cap, the placement-error value is the squared index distance to its left neighbour plus the squared index distance to its right neighbour. A correctly ordered interior cap has a baseline value of 2.
• The arrangement score is the sum of those values minus the 76-point perfect-order baseline. The labels use product thresholds of 0, 1–16, 17–40, 41–100, and above 100.
• The polar plot maps each cap to its generated hue position and scales radius by the square root of its placement-error value. Its orientation is descriptive geometry only and does not classify a cone condition.

15-cap hue-order activity

• The activity generates 15 HSL colours at 24-degree hue intervals with 80% saturation and 50% lightness. The first colour remains fixed.
• The score counts pairwise inversions among the other 14 colours: each pair adds one when it appears in the opposite order from the generated reference sequence.
• Labels use product thresholds of 0–2, 3–6, and above 6 inversions. They describe this arrangement and are not clinical D-15 crossing rules.

Symbol and children’s shape activities

• The symbol activity contains eight original SVG items: four red-green-oriented pairs and four blue-yellow-oriented pairs. It reports exact symbol matches by group.
• The children’s activity samples 40 points along each of six original shape outlines. A trace is counted when at least 62% of those samples are within 25 SVG units of a pointer path containing eight or more points.
• These rules measure answers and tracing within the interface. They have no validated clinical cutoff.

Change log

• 30 September 2026: published exact generation and scoring details; removed diagnostic, subtype, severity, pass/fail, and occupational interpretations from the original browser activities.