Farid Zaripov
Roughly 1 in 12 men and 1 in 200 women have some form of color vision deficiency — around 300 million people worldwide. Very few of them see the world in grayscale. The far more common experience is that certain pairs of colors become difficult or impossible to tell apart, which turns an ordinary red-and-green status indicator into a coin flip.
How Color Vision Works, Briefly
Normal color vision relies on three types of cone cell in the retina, each most sensitive to a different part of the spectrum: long wavelengths (red), medium (green) and short (blue). The brain compares their responses to derive color. Color vision deficiency occurs when one cone type is missing, or when its sensitivity is shifted so that it overlaps too closely with another.
Most forms are inherited and encoded on the X chromosome, which is why they are far more common in men — a man has one X chromosome and no second copy to compensate. Deficiency can also be acquired later in life through cataracts, diabetes, multiple sclerosis, optic nerve damage, or as a side effect of certain medications.
The Four Types
| Type | What is affected | Prevalence (men) | Confused pairs |
|---|---|---|---|
| Deuteranomaly | Green cone shifted | ~5 % | Red/green, brown/green |
| Protanomaly / protanopia | Red cone shifted or absent | ~1–1.5 % | Red/green; reds look darker |
| Deuteranopia | Green cone absent | ~1 % | Red/green, purple/blue |
| Tritanomaly / tritanopia | Blue cone affected | <0.01 % | Blue/green, yellow/pink |
| Achromatopsia | No functioning color vision | ~1 in 30 000 | All color; often with light sensitivity |
Two practical points follow. First, red-green deficiency dominates — over 90 % of all cases — so if you optimize for one thing, optimize for that. Second, protanopia has a consequence that surprises designers: reds appear not just shifted but darker, so red text on a dark background can drop below usable contrast entirely.
What Online Tests Actually Tell You
The classic clinical instrument is the Ishihara plate — a circle of colored dots containing a number visible to normal color vision and invisible, or different, to deficient vision. The Farnsworth-Munsell 100 hue test instead asks you to arrange color chips in order, and measures how far the arrangement deviates.
Both work on paper under controlled lighting. On a screen, several things interfere:
- Display calibration. An uncalibrated monitor shifts the very colors the test depends on.
- Ambient light. Testing under warm indoor lighting or in direct sunlight changes perceived hue.
- Blue-light filters and night modes. These deliberately distort color and will produce a false result.
- Screen type. OLED, IPS and TN panels render saturated reds and greens differently.
So an online color vision test is a useful screening tool — it will reliably tell you whether something is worth investigating — but it is not a diagnosis. If a screening result suggests deficiency, an optometrist can confirm the type and severity with calibrated plates and an anomaloscope. That distinction matters practically, because some occupations (pilots, certain electrical and marine roles) require formal certification that no online test can provide.
To get the most honest result from a screen test: use a properly lit room without direct glare, disable night mode and blue-light filters, set the display to a neutral color profile, view at normal brightness, and do not lean in — squinting at individual dots defeats the design of the plates.
Designing So It Does Not Matter
The single most useful principle in accessible design is that color must never be the only carrier of information. That is not a stylistic preference; it is WCAG success criterion 1.4.1. Every color-coded distinction needs a second channel:
- Shape or icon. A tick and a cross are unambiguous where green and red are not.
- Text labels. "In stock" beats a green dot.
- Position. Consistent placement lets people learn the meaning without reading the color.
- Pattern or texture. In charts, hatching separates series that hue alone does not.
- Brightness difference. Two colors that differ in lightness stay distinguishable even when hue collapses.
Contrast Is Not the Same Problem
Contrast ratio measures relative luminance, and luminance is largely preserved in color vision deficiency. This means a palette can pass every contrast check and still be unusable: red and green of identical lightness have poor color separation but a contrast ratio near 1:1 against each other, which contrast tools do not flag because they compare text against background rather than one hue against another. Checking contrast covers legibility; simulating the deficiency covers confusability. Both are needed.
Combinations to Avoid
- Red and green at similar lightness — the classic failure, and the most common one in status indicators.
- Green and brown, green and orange — routinely confused in deuteranomaly.
- Blue and purple — purple is perceived as blue when the red channel is weak.
- Light green and yellow.
- Blue and gray in tritanopia.
Combinations That Work
Blue and orange is the reliable workhorse pair: it survives all common deficiency types and remains distinguishable in grayscale. Blue and yellow, and any pairing with a large lightness difference, also hold up. When a design genuinely needs red and green semantics — profit and loss, pass and fail — keep the hues but make one clearly darker than the other, and add an icon.
Verifying Your Work
Three checks catch nearly everything:
- Simulate. View the interface as it appears with each deficiency type. A simulator applies the standard transformation matrices for protanopia, deuteranopia and tritanopia so you can see which distinctions collapse.
- Convert to grayscale. A crude but effective proxy — if two elements become identical in grayscale, they rely on hue alone.
- Read it aloud. If describing the interface requires saying "the green one" or "the red button", the color is doing work no other channel is doing.
Everyday Consequences
Beyond interface design, color vision deficiency affects ordinary tasks in ways that people with normal vision rarely anticipate: judging whether meat is cooked, spotting ripe fruit, reading heat-map charts and transit maps, distinguishing status LEDs on devices, matching clothes, and interpreting the red-green markings on wiring. Traffic lights are usually manageable because position is standardized — which is itself the design principle above, applied a century before anyone wrote it down.
For people who need to identify a color rather than merely distinguish it, a color namer reads the actual value from an image or the screen and states what it is — useful when the question is "is this shirt green or brown" rather than "do these two things differ".
Summary
Color vision deficiency is common, overwhelmingly red-green, and rarely total. Online tests screen usefully but do not diagnose — a clinical test does that. And the design fix is not a special palette for a minority: it is making sure every distinction carries a second signal beyond hue, which produces interfaces that are also clearer in bright sunlight, on cheap displays, and in grayscale print.
