Fold a face photograph in half and mirror one side onto the other, resulting in an appearance that feels somewhat eerie instead of enhanced. Nearly everyone has attempted this at some time, typically using their own likeness. The response is seldom one of pleasure.
For approximately thirty years, symmetry has been the prevailing explanation for what makes a face appealing, recited in magazine articles and introductory psychology classes with unwavering assurance. A comprehensive cross-cultural study has now subjected this notion to a more rigorous evaluation than is typically encountered, and symmetry did not emerge favorably.
## What the researchers assessed
Karel Kleisner from Charles University in Prague spearheaded an international team that collected 1,550 facial photographs from ten distinct populations: Brazil, Cameroon, Czechia, Colombia, India, Namibia, Romania, Turkey, the UK, and Vietnam. Each image was unaltered, captured directly facing the camera with a neutral head position and direct gaze, and manually marked with 72 anatomical landmark points.
From these landmarks, three specific elements were extracted. Symmetry. Gender typicality, which indicates how masculine or feminine the fundamental shape appears. And distinctiveness, which denotes how far a face deviates from the mathematical mean of its respective population.
Attractiveness assessments were provided by opposite-gender volunteers from the subject’s own country, which is more significant than it may seem. A Namibian face was evaluated against Namibian standards rather than a universal standard largely derived from European features.
## Symmetry did not endure the analysis
In [a paper published in Evolution and Human Behavior](https://doi.org/10.1016/j.evolhumbehav.2023.10.001), the team revealed that distinctiveness consistently lowered attractiveness ratings across all ten populations, without exception or cultural variation. Femininity enhanced the ratings for women’s faces. Masculinity had little effect on men’s faces.
Once the other factors were included in the model, symmetry was categorized as noise.
In articles by [PsyPost](https://www.psypost.org/revisiting-the-science-of-attraction-averageness-is-key-to-facial-beauty-global-study-finds/) covering the findings, this outcome contradicted a notion many consider established. Moreover, the paper carefully frames the issue: what the raters favored was nearness to a local average, not symmetry.
## Why previous studies concluded differently
Earlier experiments often created their own faces. A researcher takes one image, digitally alters it to create a more symmetrical version and a less symmetrical version, then surveys individuals for their preferences. Under such conditions, symmetry is favored, which is how [David Perrett and colleagues at St Andrews](https://doi.org/10.1016/S1090-5138(99)00014-8) arrived at their widely referenced 1999 conclusion stating that symmetrical faces are preferred.
The problem is, a transformed pair of images varies on only one dimension. Real faces exhibit differences across numerous dimensions simultaneously, and an observer has much more information competing for attention when looking at one. Manipulated stimuli reveal the visual cues a viewer can detect, which differs from what individuals rely on in real-world scenarios.
## A second team conducted another study
A team led by Benedict Jones at the University of Strathclyde in Glasgow released a follow-up in [Scientific Reports](https://www.nature.com/articles/s41598-025-86974-0) in February 2025. Fifty male and fifty female photographs, all unaltered, were evaluated by 400 volunteers, with shapes assessed from the images rather than artificially crafted.
Averageness predicted attractiveness ratings. So did femininity, but only in women’s faces. Masculinity and symmetry showed no predictive value whatsoever.
The occurrence of two independent samples revealing the same pattern is significantly more intriguing than a single instance.
Both teams also acknowledge that the effects are minor. Averageness accounts for only a small fraction of why one face is rated higher than another, with the rest influenced by skin, color, expression, body weight, and a multitude of other factors yet to be thoroughly measured.
## What average is actually indicating
Averageness is an unusual term to use as a compliment. The concept originates from [Judith Langlois and Lori Roggman](https://journals.sagepub.com/doi/10.1111/j.1467-9280.1990.tb00079.x), who in 1990 digitized student photographs, combined them pixel by pixel, and discovered that the composites scored better than nearly every individual face included. The more faces in the composite, the higher the rating rose.
Two explanations have been vying for dominance ever since. One posits that a prototypical face signals a strong immune system and good genetics. The other suggests that brains rapidly process familiar forms, mistaking this ease for enjoyment. Neither theory has triumphed, and both studies here recognize the inquiry as