Why foreign facial norms do not transfer to Thai faces

"Below average" means nothing until you can say whose average — and the answer changes the result entirely.

Last updated 2026-09-21

What a reference value is, and is not

A reference value in the anthropometric literature is a mean and standard deviation measured from one group of people, with that group, their ages and the measurement method stated. It describes that group. It is not a target every human should approach.

The misreading happens when the figure is lifted out of its context. One group's average nose width, relabelled "the standard", instantly makes everyone outside that group look abnormal — when what changed was the label, not the face.

Why the population matters

Physical anthropology has measured facial proportions across populations for a long time, and the averages for many values are not the same between groups. That is a descriptive fact and says nothing about any group being better than another.

The practical consequence is direct. Score a Thai face against another population's averages and part of the difference you see is the difference between groups rather than anything about you — and afterwards the two cannot be separated.

This is why a tool issuing a "perfection score" without naming its comparison group is unusable, however flattering the number.

The cohort in use, and its limits

DooDee’s reference figures come from published work measuring 240 Thai adults — 120 women and 120 men — aged 18–35.

The limits deserve equal billing:

A narrow age band — 18–35. Outside it the comparison loosens, because facial proportions change with age.
Cohort size — 240 is enough for workable averages, but it is not large enough to describe the full variation of a population.
The method has to match — reference figures are measured from defined anatomical points. Where the landmarks detected in an image are not those same points, the comparison shifts systematically.

Why only 12 values can be compared

Of 51 measured values, 12 have both a published mean and a standard deviation for a comparable group. The rest are not unmeasurable — there is simply nothing to compare them against.

Having no reference does not make a value useless. It changes which question it answers. A value without a cohort still says how far left differs from right, or how this month differs from last — which, for anyone tracking a result, is often more useful than a percentile.

A percentile computed from a handful of people is not a percentile, so one is not shown as fact until at least 30 scores sit behind it.

The question to put to every tool

"Compared against whom, how many of them, and at what ages?"

If it cannot answer, the number is an opinion formatted to look like a measurement. If it can, you immediately know how far the result applies to you.

See your values against a Thai cohort