Grouping by color alone fails for users with color vision deficiency

What it is

Among the ways similarity gets used as a grouping cue, color is the one most often deployed on its own — everything belonging to one category gets painted the same hue, and matching hue is meant to signal "these belong together." That practice carries a hidden assumption: that everyone can tell the chosen hues apart. That assumption fails for users with color vision deficiency — if the hues chosen for grouping happen to fall on a pair that a given type of deficiency confuses, those users don't see several distinct colors sorted into groups; they see what looks like one indistinguishable patch of color. The grouping cue does not merely weaken — it fails outright.

Why it happens

Whether a color pair can be told apart depends on which two colors they specifically are, not on the blanket question of whether someone "has a color vision deficiency." Most types of deficiency compress a specific pair of hues (certain combinations of red and green, or blue and yellow) so closely together perceptually that they nearly or fully overlap, while lightness differences usually remain distinguishable. That means the failure of color-only grouping is not random: if the grouping colors a designer picked happen to sit on the same confusion line, they can be completely indistinguishable to users with the matching deficiency even though the hues look highly distinct to someone with typical color vision — while if the grouping colors also carry a clear lightness difference, users may still be able to separate them by lightness even when the hues themselves fall in the confusable range.

Studying it

A common way to assess whether such grouping is safe is to run a screenshot of the interface through a color-vision-deficiency simulation tool covering the common deficiency types, then check whether the colors used for grouping remain distinguishable after the transformation. But simulation tools are built on generalized models of deficiency and cannot fully represent any specific individual's actual perception. A more rigorous check recruits real users with the matching type of deficiency, has them perform the grouping judgment directly, and compares their results against what the simulation predicted — guarding against a false "safe" conclusion drawn purely from over-reliance on the simulator.

Where it stops holding

If the items being grouped carry some other distinguishable property beyond color — different shapes, positions that are already separated, text labels — users still have another cue to fall back on even if the color grouping fails, and the cost is not severe. But if color is the sole basis for grouping — several data series in a chart distinguished only by color, with identical shape, position, and labels — a user with color vision deficiency loses the grouping information outright, with no other cue available to compensate.

Applying it

• For any grouping done through color alone, layer on at least one non-color distinguishable difference — shape, pattern, text label, position — so users with color vision deficiency can complete the grouping judgment without relying on color.
• When choosing grouping colors, prioritize spreading them apart in lightness rather than only in hue — lightness is a dimension users with color vision deficiency can still reliably perceive.
• To verify: place the design's screenshot side by side with a color-vision-deficiency-simulated version and check whether the colors used for grouping remain instantly distinguishable after the transformation; where possible, repeat the same grouping judgment task with real users who have the matching deficiency as a cross-check against the simulation.

Related

• Same group: A2.15.2 Spacing-only visual grouping does not transfer to a screen reader that reads content in linear sequence · A2.15.3 Low-contrast container cues can be entirely invisible to users with low vision · A2.15.4 A critical grouping relationship needs a non-visual, perceivable encoding as a fallback · A2.15.5 Purely visual grouping can fail after zooming or resizing text, once spacing ratios shift
• Nearby: A1.06 Color vision deficiency · J2.03 Information must not depend on colour alone · U4.04 Color-blind-safe palettes
• Search terms: color vision deficiency · color-blind safe · redundant coding