Facial beauty is driven by a handful of biological signals that most people, across cultures and even from infancy, respond to in broadly similar ways. Symmetry, averageness of proportions, skin quality, and sex-typical features consistently rank among the strongest predictors of how attractive a face is judged to be. Yet the science also shows that these universal tendencies interact with personal experience, cultural exposure, and context in ways that make the full picture far messier than any single formula can capture.
Why Symmetry Keeps Showing Up
If you’ve ever noticed that a face looks “off” without being able to say why, asymmetry is a likely culprit. A large body of research treats bilateral symmetry as a marker of developmental stability. The idea is straightforward: a body that grew up under fewer stresses from infection, poor nutrition, or genetic disruption tends to end up more evenly balanced on both sides. Faces high in symmetry consistently receive higher ratings not just for attractiveness but also for perceived health and positive personality traits.1Personality and Individual Differences. Facial symmetry and judgements of attractiveness, health and personality A related hypothesis links symmetry specifically to parasite resistance: in environments where infectious disease was historically a major threat, a symmetric face may have been an honest advertisement that its owner had a robust immune system.2PubMed. Human facial beauty: Averageness, symmetry, and parasite resistance
Symmetry alone, though, is not the whole story. Faces that are closer to the population average in their proportions also tend to be rated as more attractive. “Average” here doesn’t mean bland. It means that the distances between facial landmarks, the size of features relative to each other, and the overall geometry cluster near the middle of the distribution rather than at the extremes. One explanation is that average proportions signal genetic diversity, which in evolutionary terms correlates with better immune function and overall fitness. Research across five distinct populations, including Americans, Brazilians, and Paraguayan indigenous groups, confirms that average features, along with youthfulness and (in women) feminine or neotenous traits, consistently play a role in attractiveness judgments.3Springer Link / Human Nature. Criteria of facial attractiveness in five populations
Sex Hormones Shape What We Find Attractive
Sexual dimorphism, the degree to which a face looks distinctly male or female, is another strong driver of attractiveness ratings. In women, estrogen-related features like fuller lips, a smaller lower jaw, and larger eyes relative to face size are generally preferred. These traits signal youth and reproductive health. In men, the picture is more complicated: testosterone-dependent features like a strong jaw and prominent brow ridges can signal dominance and immune competence, but exaggerating them too far often makes a face look aggressive rather than appealing.4PubMed. Effects of sexual dimorphism on facial attractiveness
For female faces specifically, it isn’t just estrogen levels in isolation that matter. Research has found that the ratio of testosterone to estradiol helps shape an attractive female face, meaning both hormones contribute interactively rather than one simply overriding the other.5PubMed. Testosterone-to-oestradiol ratio is associated with female facial attractiveness This makes intuitive sense: a face doesn’t look attractive just because it is maximally feminine. Some structural definition, which testosterone helps create, adds to the appeal. The balance matters more than the absolute level of either signal.
There’s a related dimension worth knowing about: baby-like facial traits. Features like large eyes, a round face, and a high forehead trigger a “cute” response in viewers. This isn’t limited to actual babies. When researchers manipulated adult faces to have more infant-like proportions, those faces were rated as cuter than their adult-proportioned counterparts.6Ethology. Manipulation of Infant‐Like Traits Affects Perceived Cuteness of Infant, Adult and Cat Faces This “neotenous” pull helps explain why youthful features remain so central to female facial beauty across cultures, as the five-population study also confirmed.
Skin Tells Its Own Story
Bone structure and proportions get most of the popular attention, but skin may be equally important. Evenness of tone, smoothness of texture, and color all send signals about health and age. One of the more striking findings in this area involves carotenoids, the yellow-red pigments you get from eating fruits and vegetables. Faces with higher carotenoid coloration in the skin are rated as healthier and more attractive in both Western and East Asian populations.7PubMed. Carotenoid skin colouration enhances face and body attractiveness: A cross-cultural study The preference held for face and body skin but not for non-face objects colored the same way, suggesting people are reading the pigment as a cue about the person’s diet and health rather than just liking the color in general.
Facial contrast is another skin-level cue with a surprisingly large influence. Women naturally have higher luminance contrast than men, meaning their eyes and lips are darker relative to the surrounding skin. An androgynous face can be made to look female simply by increasing this contrast, or male by decreasing it.8PubMed. A sex difference in facial contrast and its exaggeration by cosmetics Because femininity in female faces is closely tied to attractiveness, this finding has direct implications for understanding why cosmetics work, a point we’ll return to shortly.
What Happens in the Brain
Looking at a beautiful face activates reward-related brain regions, and the response is not subtle. A meta-analysis of brain imaging studies found that perceiving facial beauty reliably activates the left ventral striatum, a structure at the core of the brain’s reward system.9PubMed Central. Seeking the “Beauty Center” in the Brain: A Meta-Analysis of fMRI Studies of Beautiful Human Faces and Visual Art This is the same circuitry that responds to food, money, and other rewards. The response in the orbitofrontal cortex, another reward-processing area, tracks linearly with how attractive the face is rated: the more beautiful, the stronger the signal.10PubMed. Brain systems for assessing facial attractiveness
But the relationship isn’t purely linear everywhere. In the nucleus accumbens, caudate, and several other reward regions, the neural response follows a nonlinear curve, jumping more sharply at the high and low ends of the attractiveness spectrum.11PubMed Central. Neural activation in the “reward circuit” shows a nonlinear response to facial attractiveness In practical terms, this means the brain doesn’t respond proportionally across the whole range. An extremely beautiful face generates a disproportionately intense reward signal compared to a moderately attractive one. This likely explains the outsized social impact of faces at the top of the attractiveness distribution.
Infants Already Know
One of the strongest arguments that facial beauty preferences have a biological root comes from research on newborns. Infants just days old spend more time looking at faces that adults rate as attractive than at faces rated unattractive.12PubMed Central. Preference for attractive faces in human infants extends beyond conspecifics These babies have had essentially zero cultural exposure to beauty standards, which makes it hard to argue that the preference is entirely learned.
Researchers have pushed this further. When newborns (average age about two days) were shown face pairs that had identical internal features but different external features, they showed no preference. But when the pairs had identical external features and different internal features, the babies looked longer at the attractive version.13PubMed. Newborn Infants’ Preference for Attractive Faces: The Role of Internal and External Facial Features Internal features like the eyes, nose, and mouth were already driving the preference, not just the overall shape of the head or hairstyle. This doesn’t mean culture is irrelevant. It means the starting point isn’t a blank slate. Babies enter the world with a template for certain configurations, and culture then layers additional preferences on top.
Cross-Cultural Agreement and Its Limits
People from very different cultural backgrounds agree on facial attractiveness more than you’d expect. White Scottish and Black South African observers showed significant agreement when rating faces, which is consistent with a biological baseline.14PubMed Central. Cross-cultural agreement in facial attractiveness preferences: the role of ethnicity and gender But the agreement was stronger for Scottish (White European) faces than for African faces, likely because both groups had more exposure to White European faces through media. South African observers relied more on color cues when judging African female faces, while Scottish observers relied more on shape cues. Familiarity with a particular face type seems to shift which features carry the most weight.
A large multi-ethnic study reinforced this nuance. While people generally agreed on how old a face looked, their agreement on attractiveness and health was weaker and depended on the combination of who was judging, whose face was being judged, and the sex of the assessor.15PLOS ONE. Cross-cultural perception of female facial appearance: A multi-ethnic and multi-centre study Age estimation is more universal because it is driven by concrete physical cues like wrinkles and skin laxity. Attractiveness sits in a more subjective space where cultural norms, personal experience, and aesthetic traditions shape the final judgment. The biological baseline gets you broad agreement on the extremes, but the middle of the distribution is where culture has the most room to operate.
The Halo Effect in the Age of Filters
Attractive faces don’t just look good. They get credited with intelligence, trustworthiness, sociability, and happiness by default. Psychologists call this the attractiveness halo effect, and it persists even in 2024 when the “attractive” version of a face is created by a beauty filter rather than by genetics. In a recent study, the same individuals were rated higher on all positive traits after their photos were digitally beautified.16PubMed Central. What is beautiful is still good: the attractiveness halo effect in the era of beauty filters The halo effect wasn’t diminished by people knowing filters exist. It operated automatically, before and after beautification, in the same direction.
This matters beyond dating profiles. The halo effect has been documented in hiring decisions, criminal sentencing, teacher evaluations of students, and medical interactions. When technology makes it easy for anyone to appear more attractive in photos, the social advantages of beauty get redistributed in unpredictable ways, and the disadvantages for those who don’t use filters may intensify.
The Golden Ratio Is a Myth
You may have seen claims that beautiful faces follow the golden ratio, a mathematical proportion of roughly 1.618 to 1. It’s one of the most persistent ideas in popular beauty science, and the evidence doesn’t support it. A systematic review looking at whether the golden ratio appears in naturally attractive faces found no significant association between the ratio and facial evaluation scores across all ethnicities studied.17PubMed. Assessment of the golden proportion in natural facial esthetics: A systematic review A separate review came to the same conclusion, stating that there is no convincing evidence linking the golden ratio to idealized human proportions or facial beauty.18PubMed Central. The golden ratio-dispelling the myth
The idea likely persists because it offers a tidy mathematical explanation for something that feels mysterious. Beautiful faces do tend to have proportions near the population average, as discussed earlier, but “average” and “golden ratio” are not the same thing. One is a statistical center, the other is a fixed mathematical constant. They overlap sometimes by coincidence, not because beauty obeys a formula.
How Cosmetics Exploit Biological Signals
Given what we know about skin contrast and sex-typical features, it’s no surprise that cosmetics work by amplifying the same cues the brain already responds to. Eye makeup and lipstick darken the eyes and lips, increasing luminance contrast and making a female face appear more feminine.19PubMed Central. Cosmetics alter biologically-based factors of beauty: evidence from facial contrast Foundation and concealer even out skin tone, mimicking the smooth, homogeneous complexion associated with youth and health. Blush adds redness that signals blood flow and vitality.
Research has dissected exactly what cosmetics change. Darkening the eyes and lips accounted for part of the effect, but not all of it. Changes in skin smoothness, redness, lip color, and shading that accentuated cheekbones also contributed, together modulating signals of youth, fertility, and femininity.20PLoS ONE. Cosmetics as a Feature of the Extended Human Phenotype: Modulation of the Perception of Biologically Important Facial Signals Cosmetics are, in effect, a technology for fine-tuning the biological signals your face already sends. They don’t create beauty from scratch; they turn up the volume on the traits the visual system is already wired to notice.
Your Phone Makes Your Face Look Wrong
If you’ve ever felt your nose looks too big in a selfie, it probably does, just not in real life. Camera lenses introduce predictable distortions at close range, and smartphones held at arm’s length are still close enough to warp your proportions. A study quantifying facial distortion found that smartphone photos taken at eight inches produced an 18% increase in vertical stretching of the midface, and photos at twelve inches still showed a 12% increase.21PubMed. Quantifying Facial Distortion in Modern Digital Photography Full-frame cameras showed similar distortion at short focal lengths, with a 12-19% increase at 24, 35, and 50 mm.
The nose, being the most forward-projecting feature, gets hit hardest. A mathematical model of the “selfie effect” demonstrated the increased perceived nasal size and calculated how it varies with distance.22PubMed Central. Nasal Distortion in Short-Distance Photographs: The Selfie Effect Plastic surgeons have noted that patients increasingly bring in selfies as evidence of features they want corrected, features that may appear entirely normal at conversational distance.23PubMed. Seeing Is Not Believing: Facial Distortion in Smartphone Photography If you’re judging your own facial beauty from phone photos, you’re evaluating a geometrically distorted version of yourself.
Faces in Motion
Almost all attractiveness research uses still photographs, but real life involves moving, expressive faces. Expression turns out to matter. People whose natural expressions trend positive are rated as more attractive, and the positivity of the expression itself independently boosts how good-looking the face appears.24American Journal of Applied Psychology. Role of Dynamic Facial Expressions in Judgement of Attractiveness: AI and Human Analysis This held for both male and female faces and was confirmed by both human raters and AI analysis. A genuine smile, in other words, doesn’t just make you seem friendlier. It makes your underlying features look better.
Beyond expression, attractiveness may extend across sensory channels. An analysis of how visual, auditory, and olfactory attractiveness ratings relate to each other found a significant positive association between how attractive someone’s voice sounded and how attractive their face looked.25PubMed Central. Attraction in every sense: How looks, voice, movement and scent draw us to future lovers and friends The link between body scent and facial attractiveness was weaker and only suggestive. Still, the voice-face connection hints that attractiveness is not a purely visual trait. Whatever biological qualities make a face look good may also shape how a voice sounds, giving observers converging signals across senses.
When AI Tries to Score Beauty
Machine learning systems trained to predict facial attractiveness have become common in apps and research tools, but they carry significant bias problems. Even when trained on seemingly balanced datasets, deep learning models exhibit significant prediction disparities across ethnic groups.26arXiv. Analysis of Bias in Deep Learning Facial Beauty Regressors Worse, cross-dataset testing shows that these models amplify societal beauty biases rather than mitigating them. The systems don’t discover some objective beauty standard. They absorb and concentrate whatever biases exist in their training data, then present the output as a number, which feels authoritative.
This is worth keeping in mind as AI-driven beauty tools become more common in social media filters, cosmetic surgery planning, and even hiring algorithms that analyze video interviews. A score generated by software isn’t measuring an objective property of your face. It’s reflecting the preferences and prejudices baked into whatever dataset the software learned from.
How Aging Reshapes the Signals
The biological cues underlying facial beauty don’t stay constant over a lifetime. Aging involves coordinated changes across every tissue layer. The facial skeleton itself remodels, with the eye sockets widening and the jaw losing volume. Fat pads atrophy or shift downward. Muscles lose tone and thickness. The skin thins, loosens, and loses the evenness of color and texture that signals youth.27PubMed Central. The Facial Aging Process From the “Inside Out” Each of these changes degrades the specific cues that the attractiveness system responds to: symmetry becomes harder to maintain as structures shift, skin contrast drops, sex-typical features soften.
Understanding aging as a multi-layer process explains why no single cosmetic intervention fully turns back the clock. A facelift tightens skin but doesn’t restore bone volume. Fillers replace lost fat but don’t address skin texture. The attractiveness system reads the whole face as an integrated set of signals, and addressing only one layer can sometimes produce a result that looks unnatural precisely because the other layers don’t match.
Beauty Standards in Art Across Centuries
If you look at faces in Western paintings from the 13th century through the 20th, you don’t see a fixed ideal. Researchers who measured the averageness, symmetry, and orientation of depicted faces found that the representation of human faces shifted substantially over time. Faces painted between the 15th and 18th centuries scored higher on conventional attractiveness metrics than those from either the 13th or 20th centuries.28Faculty and Staff Works. A Quantitative Approach to Beauty: Perceived Attractiveness of Human Faces in World Painting The 20th-century drop reflects the rise of modernism and greater artistic freedom, which led to the depiction of diverse face types rather than idealized ones.
Renaissance artists often idealized facial features to align with the cultural and aesthetic values of their era, prioritizing symmetry and balanced proportions, traits that remain central to contemporary beauty standards.29PubMed. Plastic Surgeons and Portraits Before the Invention of Photography The continuity is interesting: artists working centuries before evolutionary psychology existed were intuitively emphasizing the same signals (symmetry, averageness, youthful proportions) that modern research identifies as biologically driven preferences. What changed over time wasn’t the underlying wiring so much as how freely artists chose to follow or diverge from it.