Skin Colour Does Not Define Ethnicity: Quantifying Variation and Overlap Across Diverse Populations

Skin colour is a prominent human trait historically used to define ethnicity, yet its validity as a classification tool remains questionable.

We quantitatively analyse over 14 000 skin reflectance measurements from eight ethnically diverse groups in the International Skin Spectra Archive (ISSA), using a standard colour space designed to model human visual perception. We assess intragroup variation and intergroup overlap through two complementary approaches: individual-level perceptual differences and group-level shared gamut volumes.

Results show that within-group variability in chromaticity and lightness frequently exceeds between-group differences. At the individual level, 89.4% (95% CI: 81.5%-91.9%) of samples have perceptually indistinguishable counterparts across ethnicities. At the group level, the median shared gamut overlap is 60.5% (95% CI: 54.5%-63.6%), indicating substantial overlap in skin colour distributions. The two methods correlate strongly (r = 0.83, p < 0.001), confirming robust intergroup overlap.

Skin colour exhibits high within-group dispersion and extensive between-group overlap. These findings challenge the use of skin colour as a reliable indicator of ethnicity and underscore the need for objective, data-driven classification frameworks. They also highlight the complex, continuous nature of human skin variation, beyond simplistic ethnic categories.