Evidence review

Facial averageness and attractiveness: what composite-face studies show

Why mathematically blended composites outscore most individual faces, how shape warping separated geometry from smooth skin and why top-decile faces still depart selectively from the mean.

Published
Last reviewed
Version
1.0
Reading time
9 min
Author
LooksMeter Research

Key findings

  1. 01Digitally blended composites of 16 and 32 adult faces were rated more attractive than nearly every individual face that went into the blend.1
  2. 02The preference is not merely a blur or skin-smoothing artefact: warping a single real face's proportions toward the population mean while preserving its original skin texture still raised attractiveness ratings.3
  3. 03Geometric averageness remains predictive even when bilateral symmetry is controlled statistically, showing that central proportions and left-right symmetry are distinct cues.3, 6
  4. 04Mathematical averageness is high-scoring rather than optimal: a composite built from the 15 highest-rated faces outscored a composite of all 60 faces while exaggerating those differences raised ratings further.2
  5. 05Raters also prefer mathematically averaged prototypes of birds, fish and wristwatches, pointing to general perceptual fluency alongside face-specific selection.4

Section 1

The claims

In everyday speech, calling a face average means placing it around the middle of attractiveness ratings. In craniofacial and perceptual science, averageness means something geometric: how close a face's feature spacings and proportions sit to the mathematical mean of its population. Confusing those two meanings leads either to the myth that plainness is ideal or to the opposite myth that overall balance does not matter.

Table 1Core claims tested in the facial-averageness literature.
ClaimWhat the studies reportEvidence
Mathematically averaged facial proportions score well above 50th-percentile attractivenessSupported across digital composites, 2D shape warps and cross-cultural meta-analyses in both male and female faces.1, 3, 6Consistent
The composite effect is only smooth, averaged-out skin textureRefuted. Holding skin texture and bilateral symmetry constant while warping landmark configuration toward the mean still increases rated attractiveness.3Not supported
The population-mean face is the single most attractive configuration possibleRefuted. Composites of high-rated faces outperform whole-sample composites, showing systematic directional preferences beyond the mean.2, 5, 7Not supported
Preference for average shapes reflects general visual processing fluency as well as biologySupported. Averaged exemplars of non-face objects also draw higher attractiveness ratings from human observers.4Consistent

Consistent: several independent studies point the same way. Mixed: studies point in different directions. Limited: one or two studies, not yet repeated. Not supported: tested, with no effect found or one too small to matter.

Section 2

Composite-face experiments

Langlois and Roggman (1990) digitised young adult male and female portraits and created pixel-aligned composites of 2, 4, 8, 16 and 32 faces.1 As more faces entered the composite, rated attractiveness climbed monotonically: the 16-face and 32-face blends outscored almost every individual portrait in the set.

Subsequent reviews across Western and non-Western rating panels confirmed a moderate-to-large positive association between proximity to population-central proportions and attractiveness ratings in both sexes.6, 7 Because combining independent faces cancels extreme values on every axis at once - very wide or very narrow eye spacing, unusually long or short midfaces, steep or compressed lower thirds - a composite automatically sits inside the central band on every ratio simultaneously.

Section 3

Separating geometric shape from skin smoothing

Two early criticisms were raised against pixel composites: first, blending 32 photographs blurs pores, blemishes and shadows into porcelain-smooth skin while second, blending left and right deviations makes the composite more symmetric than an ordinary face.

Rhodes, Sumich and Byatt (1999) tested both objections directly by warping individual faces toward or away from the sample's mean shape while retaining each person's real unblended skin texture and factoring out bilateral symmetry.3 Moving a face's geometry closer to the mean raised attractiveness and lowered perceived distinctiveness even with zero texture averaging and with symmetry held constant. Grammer and Thornhill (1994) likewise analysed landmark coordinates alongside ratings and found independent contributions of averageness, symmetry and sex-typical traits.5

Section 4

Why mathematical averageness is a floor, not a ceiling

Perrett, May and Yoshikawa (1994) demonstrated the exact boundary of averageness in both British and Japanese samples.2 They built one composite from 60 faces and a second composite from only the 15 highest-rated faces in that cohort, then generated a third image exaggerating the shape vector from the 60-face mean through the top-15 composite by a further 50 percent.

  • Raters preferred the top-15 composite over the all-60 composite in both populations.2
  • Raters preferred the exaggerated shape morph over the top-15 composite itself, proving that top-rated faces share directional features (such as cheekbone contour, jaw taper or relative eye aperture) that sit systematically past the population average.2, 7

In practical measurement terms, central proportions prevent a single jarring outlier from dominating a viewer's attention, while sex-typical contours, feature contrast and grooming create positive standout appeal above that harmonious baseline.6, 7

Section 5

Why human vision favours central proportions

Halberstadt and Rhodes (2000) showed that humans also judge averaged drawings of dogs, birds and wristwatches as more attractive than individual exemplars.4 Category-central prototypes are encoded more fluently by visual cortex so perceptual fluency carries an immediate positive affective signal. Evolutionary psychologists add that maintaining central proportions across many cranial growth zones simultaneously also correlates with stable developmental buffering, making prototype fluency and biological signalling compatible accounts rather than rivals.6, 7

Section 6

Reading a harmony result on your own photo

Close-range phone cameras introduce perspective bulge that artificially inflates nose-to-face width by up to 30 percent at 30 cm, making a well-proportioned face measure as an outlier.8 Taking a front photo at eye level from at least 1.5 metres lets you compare your actual vertical sections, horizontal fifths and eye spacing against the central percentile corridors reported in LM-R-001.

Section 7

Limits of this review

Most morphing experiments test static frontal photographs under studio lighting rather than faces in motion or conversation. Population means also shift across ancestry groups, so comparing a face against a single narrow sample conflates ancestry differences with departures from averageness.6

References

  1. 1Langlois JH, Roggman LA. Attractive faces are only average. Psychological Science. 1990;1(2):115-121. doi:10.1111/j.1467-9280.1990.tb00079.x
  2. 2Perrett DI, May KA, Yoshikawa S. Facial shape and judgements of female attractiveness. Nature. 1994;368(6468):239-242. doi:10.1038/368239a0 · PubMed
  3. 3Rhodes G, Sumich A, Byatt G. Are average facial configurations attractive only because of their symmetry?. Psychological Science. 1999;10(1):52-58. doi:10.1111/1467-9280.00106
  4. 4Halberstadt J, Rhodes G. The attractiveness of nonface averages: implications for an evolutionary explanation of the attractiveness of average faces. Psychological Science. 2000;11(4):285-289. doi:10.1111/1467-9280.00257 · PubMed
  5. 5Grammer K, Thornhill R. Human (Homo sapiens) facial attractiveness and sexual selection: the role of symmetry and averageness. Journal of Comparative Psychology. 1994;108(3):233-242. doi:10.1037/0735-7036.108.3.233 · PubMed
  6. 6Rhodes G. The evolutionary psychology of facial beauty. Annual Review of Psychology. 2006;57(1):199-226. doi:10.1146/annurev.psych.57.102904.190208 · PubMed
  7. 7Little AC, Jones BC, DeBruine LM. Facial attractiveness: evolutionary based research. Philosophical Transactions of the Royal Society B: Biological Sciences. 2011;366(1571):1638-1659. doi:10.1098/rstb.2010.0404 · PubMed
  8. 8Ward B, Ward M, Fried O, Paskhover B. Nasal distortion in short-distance photographs: the selfie effect. JAMA Facial Plastic Surgery. 2018;20(4):333-335. doi:10.1001/jamafacial.2018.0009 · PubMed

Every reference was checked against Crossref before publication. Those indexed in PubMed were checked there as well. Findings are summarised in our own words. Read the original before relying on a number.

How to cite

LooksMeter Research. Facial averageness and attractiveness: what composite-face studies show. Version 1.0. 19 November 2026. https://www.looksmeter.com/research/facial-averageness-evidence

About this document

LooksMeter is a consumer appearance analysis product. This document is written by the LooksMeter team for general information. It has not been peer reviewed or reviewed by a medical professional. The measurements it describes are not clinically validated and are not a diagnosis or a medical device. See a qualified clinician about any health or surgical decision.

Corrections are welcome at [email protected]. Material changes are recorded by raising the version number above.

Continue