ATTRACTION SCIENCE · PROPORTIONS

Facial averageness: why composite proportions rate high (and where they stop)

In craniofacial science, facial averageness does not mean looking plain: it means having vertical thirds, horizontal fifths, midface ratio and eye spacing that sit near the geometric centre of human proportions with no jarring outlier. Here is what composite-face studies proved, where averageness stops and how you can check your own harmony corridors with the free Facial Ratios calculator →

Two meanings of the word average

No word in facial aesthetics causes more confusion than 'average'. In casual conversation, calling a face average means its overall rating sits around the 50th percentile - neither striking nor unusual.

In evolutionary psychology and anthropometry, averageness is a geometric distance metric: how close a face's landmark configuration sits to the mathematical mean configuration of its age, sex and ancestry group. A face whose eye spacing, midface length, philtrum and jaw width all land near the population median has high mathematical averageness - and human raters consistently judge that balanced geometry as well above 50th-percentile attractiveness.

What composite-face experiments proved

Across four decades of controlled experiments, researchers have separated three effects that used to be tangled together when photos were blended.

Key experimental findings on facial averageness
StudyMethodMain finding
Langlois & Roggman (1990)Digital pixel composites of 2, 4, 8, 16 and 32 facesAttractiveness rose steadily with sample count: 16- and 32-face composites beat nearly every constituent face
Perrett, May & Yoshikawa (1994)All-60 composite vs top-15 composite + 50% caricatureAttractive faces share systematic non-average traits beyond the mean, so averageness is high-scoring but not optimal
Rhodes, Sumich & Byatt (1999)Landmark warping toward mean geometry with original skin texture keptPure shape averageness raises ratings even when skin smoothness and symmetry are held constant
Rhodes (2006) meta-analysisQuantitative synthesis across Western and non-Western cohortsModerate-to-large positive effect (r ~ 0.38 to 0.52) for geometric averageness in both male and female faces

Why human vision prefers central proportions

Two complementary explanations survive in the literature: both can be true at the same time.

The cognitive-fluency (prototype) account shows that the human visual cortex processes category-typical shapes faster and with lower neural effort than outlier shapes - an effect also seen for birds, watches, cars and dot patterns (Halberstadt and Rhodes 2000). Easy perceptual processing registers subjectively as visual pleasantness.

The biological-signalling (koinophilia) account notes that large craniofacial deviations on a single axis are statistically more common after developmental stress, whereas landing near the population centre across a dozen independent growth zones requires steady, unperturbed development.

Where averageness stops: dimorphism and halos

If mathematical averageness explained 100 percent of attractiveness, two faces at the population mean would outscore every model on earth. In practice, Perrett's 1994 study and later 3D morphometric work show that top-decile faces combine a high-averageness harmony scaffold with selective departures in sex-typical or youthful directions.

In male faces, stronger gonial definition, brow-ridge structure and cheekbone prominence sit slightly past the population mean without breaking vertical thirds. In female faces, slightly fuller lips, larger relative eye aperture and softer lower-third taper sit slightly past the female mean on the neoteny-dimorphism axis.

Put simply: mathematical averageness eliminates 'failos' (jarring proportion outliers), while dimorphism, angularity and grooming supply the 'halos' that make a face memorable.

Why camera lens distortion mimics low averageness

An arm's-length phone camera at 30 to 40 cm enlarges the nose by roughly 25 to 30 percent relative to the ears and narrows bigonial jaw width, artificially pushing central-face ratios out of their normal corridor.

Stepping back to 1.5 to 2 metres (or using a rear camera timer at chest-to-eye level) restores natural orthographic proportions so you measure your actual facial harmony rather than wide-angle lens bulge.

Frequently asked

Does facial averageness mean looking plain or ordinary?

No. In face-perception research, averageness means mathematical proximity to a population's central proportions - eye spacing, vertical thirds, nose-to-mouth ratio and jaw taper sitting near the geometric mean rather than at extreme tails. In everyday speech people use 'average' to mean middle-of-the-pack attractiveness (5 out of 10), whereas a mathematically averaged 32-face composite routinely scores in the top 10 to 20 percent of human rating panels.

Who discovered that composite faces look attractive?

Francis Galton first noticed the effect in 1878 while superimposing photographic plates onto one sheet. Over a century later, Judith Langlois and Lori Roggman (1990) tested digital pixel-averaged composites of 4, 8, 16 and 32 young adult faces and showed that raters judged the 16- and 32-face composites significantly more attractive than almost all of the individual faces used to build them.

Are the most attractive faces simply the most average?

Not quite. David Perrett, Keith May and Sakiko Yoshikawa (1994) built a composite of 60 faces and a second composite from only the 15 highest-rated faces in that group. Raters preferred the top-15 composite over the all-60 composite and preferred an exaggerated morph pushing those distinctive traits slightly further still. Mathematical averageness builds a strong floor (around the 80th percentile), while sexual dimorphism, neoteny and feature contrast separate the very top.

Is the composite effect just blurred, smooth skin?

Texture smoothing contributes part of the effect, because averaging 32 photos cancels out individual blemishes and uneven lighting. However, follow-up experiments by Gillian Rhodes and colleagues (1999, 2006) warped a single real person's skin texture toward and away from the population-mean landmark geometry without blurring texture at all: pure geometric averageness still raised attractiveness ratings substantially across cultures.

How does facial averageness show up in a LooksMeter scan?

It corresponds directly to the Harmony pillar: when facial thirds, horizontal fifths, midface ratio and eye spacing all sit inside their central reference corridors simultaneously, nothing pulls the eye out of balance. Distinctiveness then comes from angularity, dimorphism and presentation.

Email me the cheat-sheet
FREE SPEC SHEET · INSTANT INBOX DELIVERY

Get the cheat-sheet + a first-time member perk

Get our facial harmony & styling reference sheet emailed to you.

or email spec sheet

Explore more

Keep reading

All guides →

This page is general information about facial aesthetics terminology, not medical or cosmetic advice. Scores from our tools are calibrated reads of one photo that describe presentation, never your worth.