Attractiveness Rating: Inside the Science and Beauty

Attractiveness ratings, when studied scientifically, turn out to be far more consistent across people than most of us expect and far more malleable than the consistency suggests. Your brain begins sorting faces by attractiveness within roughly 100 milliseconds of seeing them, often before you are consciously aware a face is there at all. Yet the number you would assign to a face on a 1-to-7 scale can shift depending on the face you saw five seconds earlier, the personality traits you associate with the person, and even your own hormone levels that day. The science of beauty ratings sits at the intersection of evolutionary biology, neuroscience, and social psychology, and its findings regularly surprise people who assume beauty is either universal or entirely subjective.

Your Brain Rates Faces Before You Know It

One of the most striking findings in this field is how early attractiveness judgments form. Experiments using masked face presentations, where a face flashes on screen so briefly that participants report they could not see it, show that people still rate the briefly glimpsed attractive faces higher than unattractive ones. Participants in these studies typically say they were guessing, yet their ratings reliably distinguish between attractive and unattractive faces, suggesting the process is largely unconscious.¹

Electrophysiology work confirms this timeline. Attractive faces generate a larger brain response at the P1 component, which peaks at around 100 milliseconds after the face appears on screen.² A separate line of research found that faces closer to the population average evoke a smaller neural response at the N170 component, suggesting they require fewer processing resources. The researchers interpret this as evidence that we find average-looking faces attractive in part because they are easier for our visual system to process.³ In other words, “processing fluency” may be one reason composite, average faces reliably score well on attractiveness scales.

The Features That Drive Facial Attractiveness

Three broad properties come up repeatedly in the literature on faces: symmetry, averageness, and sexual dimorphism. They do not all work the same way, and they interact in ways researchers are still untangling.

Symmetry has long been proposed as a signal of developmental stability. Studies consistently find that more symmetric faces are rated as more attractive, but the picture has a wrinkle: symmetric faces also appear to possess other attractive qualities that are independent of symmetry itself, qualities that researchers have not yet pinned down.⁴ An early study found that facial attractiveness correlated negatively with fluctuating asymmetry in men but not in women, hinting that symmetry carries different weight depending on sex.⁵ And in young women, while facial beauty did track with physical fitness, asymmetry did not explain the connection between the two.⁶

Sexual dimorphism adds another layer. You might expect that the most masculine male faces and the most feminine female faces would be rated highest. That holds for women: feminized female faces are preferred across populations. But for men, the story is more complicated. A landmark cross-cultural study found that subjects preferred feminized male faces over masculinized ones, even though masculinized faces were rated higher in perceived dominance. The more masculine a male face looked, the more observers attributed negative personality traits to it, such as coldness and dishonesty.⁷ A more recent analysis across world populations found that only one specific component of male masculinity, the kind associated with taller body size, reliably boosted male attractiveness.⁸

Skin Tells Its Own Story

Face shape gets most of the popular attention, but skin quality is a powerful driver of attractiveness ratings in its own right. Skin homogeneity, meaning even texture without blemishes or discoloration, correlates positively with how attractive male raters find women’s faces.⁹

Color matters too, and in a surprisingly specific way. Skin with higher carotenoid coloration (the warm yellowish tones that come from eating fruits and vegetables rich in carotenoids) is rated more attractive across populations, including both Caucasian and Hong Kong Chinese raters.¹⁰ The preference is not just an aesthetic response to the color itself. When researchers scrambled face images so the colors and contrasts were preserved but the facial structure was gone, the carotenoid color boost to attractiveness disappeared. Only when the color appeared in a recognizable face did it increase ratings, consistent with carotenoid skin color functioning as a health signal rather than simply being a prettier hue.¹¹

Body Proportions and the Waist-to-Hip Ratio

The waist-to-hip ratio has been one of the most studied body metrics in attractiveness research. Women with a lower ratio are consistently rated as more attractive, a finding replicated across diverse populations and methods, including in historically diverse, clothed depictions of women in pre-contemporary European portraits.¹² When men choose the most attractive female silhouette from a range, low to average values are picked most frequently, and values outside the normal range in either direction are rarely selected.¹³

However, a recent study suggests that overall body curvature may matter more than the ratio alone. When researchers compared how well width, waist-to-hip ratio, and curvature predicted attractiveness ratings, curvature accounted for about 65 percent of the variance, while the ratio and width each accounted for less than 30 percent.¹⁴ The waist-to-hip ratio may have been doing some of its predictive work by roughly capturing curvature, rather than being the key signal itself.

Cross-Cultural Agreement Is Real but Not Uniform

A common question is whether beauty standards are universal or culturally constructed. The research points to something in between. A study comparing attractiveness judgments by African and Scottish observers found significant cross-cultural agreement overall. But the level of agreement depended on whose faces were being rated: agreement was stronger when both groups judged Scottish faces than when both judged African faces.¹⁵ The cross-cultural study on dimorphism preferences mentioned earlier also found that while feminized female faces were preferred in both UK and Japanese samples, the preference was stronger for within-population judgments, suggesting that some attractiveness cues are learned or culturally amplified.⁷

The Brain’s Reward Circuitry and Attractive Faces

Neuroimaging studies help explain why attractiveness ratings feel less like analytical judgments and more like gut reactions. The orbitofrontal cortex, a brain region involved in reward processing, shows a linear increase in activity as faces increase in attractiveness. Meanwhile, the amygdala shows a different pattern: it responds most strongly to faces at both extremes, the highly attractive and the highly unattractive, rather than to faces in the middle.¹⁶ The same medial orbitofrontal region also increases activity as a function of moral “goodness” ratings, which may help explain why people intuitively associate beauty with virtue.¹⁷

Bonding chemistry layers onto these reward circuits. When men received oxytocin and then viewed photos of their romantic partner alongside photos of unfamiliar women, their brain’s reward regions, including the ventral tegmental area and nucleus accumbens, responded more strongly to the partner’s face. This boost was partner-specific, not just familiarity-driven.¹⁸ Romantic attachment, in effect, rewires how attractive a specific face feels.

Beyond the Face and Body

Attractiveness is not only visual. Research on multisensory attractiveness asks whether someone who looks attractive also tends to sound attractive or smell attractive. For women, face, voice, and body odor attractiveness are weakly correlated, offering modest support for a “general attractiveness factor” across senses.¹⁹ For men, those cross-modal correlations largely vanish, suggesting that male attractiveness is more modality-specific. Another study found a positive association between how attractive someone’s body movements looked and how attractive their body odor smelled, adding movement to the mix of signals that can align across senses.²⁰

Immune activation may even alter these channels independently. After vaccination, participants’ body odor was rated as more attractive and their faces as less attractive, while vocal attractiveness did not change.²¹ Each sensory channel appears to carry partly independent information about health and condition.

Personality Rewrites Physical Attractiveness

One of the most practically relevant findings in this field is that knowing someone’s personality genuinely changes how physically attractive they look to you. In controlled experiments, providing positive personality information alongside a photograph significantly increased attractiveness ratings for faces at all starting levels, whether the face was initially rated as attractive, neutral, or unattractive. The effect ran the other way too: negative personality information dragged ratings down.²² Brain-level evidence supports this. When personality trait words that match attractiveness levels are paired with faces, they speed up perceptual processing, as if the brain integrates character and appearance into a single evaluation.²³

The personality effect extends to body attractiveness as well. When men were given positive personality descriptions alongside photos of women’s bodies, they rated a wider range of body sizes as attractive than the control group did. Negative personality information narrowed the range.²⁴ This matters because it means attractiveness ratings gathered under the sterile conditions of a lab, where subjects see photos stripped of context, personality, and movement, systematically underestimate how much real-world attraction varies between individuals.

How Your Own Hormones Shift What You Find Attractive

The person doing the rating matters as much as the person being rated. Women’s preferences for facial attractiveness and sexual dimorphism change with their hormone levels. When women had higher testosterone levels in a given test session, they showed greater motivational salience for attractive faces and feminine female faces. When they had higher estradiol-to-progesterone ratios, they found sexually dimorphic and attractive female faces more rewarding.²⁵ These are within-person fluctuations, meaning the same woman shifts her ratings depending on where she is in her cycle or how her hormones happen to be running that day.

The Face You Just Saw Changes the Face You Rate Next

When people rate faces in a sequence, their judgment of the current face is pulled toward the attractiveness of the face they just saw. If you just saw a very attractive face, you tend to rate the next face a bit higher than you would have otherwise. This assimilation effect is well documented, though a competing “contrast” effect (where the next face seems less attractive by comparison) can also emerge under certain conditions.²⁶ Research has shown that both forces operate simultaneously: a response-level bias pulls ratings toward the previous rating, while a stimulus-level bias pushes perception away from the previous face’s attractiveness. Increasing how long the previous face is viewed strengthens the contrast effect but not the assimilation effect, suggesting the contrast effect has a perceptual root.²⁷

Category membership modulates these biases. When raters perceive two faces in a sequence as belonging to the same group, assimilation dominates. When they see the faces as belonging to different categories, assimilation weakens or disappears.²⁸ This has practical implications for anyone interpreting attractiveness ratings from studies: the order in which faces are shown can systematically inflate or deflate specific ratings.

How Attractiveness Rating Scales Actually Perform

Researchers use several scales to collect attractiveness ratings, most commonly Likert scales of varying lengths (1-to-5, 1-to-7, 1-to-10) and visual analog scales where participants mark a point along a continuous line. A systematic comparison found little evidence that any particular Likert scale length outperforms another.²⁹ Likert and visual analog scales tend to agree well through most of the range, but diverge at the top: faces scored as “very pleasant” on a Likert scale received lower scores on the visual analog scale than their Likert position would predict.³⁰ And the viewing angle matters: ratings from front, three-quarter, and profile views differ significantly for the same face in about 80 percent of cases.³¹ All of this means that a single attractiveness “score” is tied to the instrument, the angle, and the context in which it was collected.

When the Rating System Breaks Down

For most people, self-rated attractiveness roughly tracks with observer-rated attractiveness, but the gap between the two is predicted by self-esteem and gender role orientation. People with higher esteem and more traditional gender roles tend to rate themselves as more attractive relative to how observers rate them.³²

Body dysmorphic disorder represents a more severe distortion. People with BDD rate their own attractiveness as significantly lower than independent evaluators do, while simultaneously rating conventionally attractive photographs as significantly more attractive than other groups do. This combination, holding others to an elevated standard while devaluing oneself, fits the cognitive-behavioral model of BDD, which emphasizes perfectionism and distorted beliefs about the importance of appearance.³³ A meta-analysis found that people with BDD show measurable differences in selective attention, interpretive biases, and memory for facial and body stimuli.³⁴ Interestingly, one area where BDD does not seem to differ from healthy controls is in the ability to detect facial symmetry. A study comparing BDD patients and healthy subjects found that perceived symmetry, but not objective symmetry, predicted attractiveness ratings in both groups equally, undermining the idea that people with BDD are hyper-detectors of facial asymmetry.³⁵

Neotenous Features and Why “Cute” Is Not the Same as “Beautiful”

Research on baby schema, the set of infantile facial features including large eyes, a round face, and a small nose, reveals a useful distinction. These features affect ratings of cuteness, attractiveness, and beauty, but not identically. One study found that baby schema features in adult women increased all three types of ratings, yet participants interpreted “beauty” as more closely tied to sexual attraction, while “attractiveness” tracked with a broader, non-sexual appeal.³⁶ Children are sensitive to baby schema too; the response to infantile facial configurations appears early in development and extends beyond human faces to animal faces.³⁷ Neotenous facial features also influence judgments that seem unrelated to cuteness, such as perceived fertility and healthiness, while body shape metrics like waist-to-hip ratio primarily predicted only whether the person appeared to have been pregnant.³⁸

AI Attractiveness Prediction and the Hiring Premium

Facial beauty prediction has become a growing subfield of artificial intelligence. Deep learning models trained on large datasets of rated faces can now predict average human ratings with moderate accuracy. One system using machine learning on 5,500 frontal facial images achieved a correlation of about 0.78 between its predicted scores and the actual human consensus ratings.³⁹ These tools are marketed for applications from cosmetic surgery planning to social media filtering, though the ethical terrain is fraught. A comprehensive review noted that deep learning methods are well suited to capturing the complex features involved in facial beauty prediction, but the field still grapples with dataset biases and cross-cultural generalization.⁴⁰

These tools exist in a world where attractiveness already carries real economic weight. Hiring research consistently finds that managers prefer more physically attractive candidates, with the preference stronger at larger and more prestigious firms.⁴¹ The mechanism seems to be that employers attribute better social skills and greater likelihood of career success to better-looking people, which then influences interview and hiring decisions.⁴² If AI systems learn to replicate the same biases in the human training data, they risk automating and scaling a form of discrimination that already permeates face-to-face interactions.

What Orthodontic and Cosmetic Interventions Actually Change

For people considering whether clinical interventions can shift how their face is rated, the evidence says yes, within limits. Orthodontic treatment increased facial attractiveness ratings in a retrospective study, with the strongest gains seen in profile views. Extraction of premolars and orthognathic surgery produced the largest changes.⁴³ A study of the underlying geometry found that the relationship between facial measurements and attractiveness was not linear but curved: there were sweet spots where attractiveness peaked, and laypeople tended to prefer a slightly more protrusive lower lip and a jaw relationship slightly different from the clinical norm.⁴⁴ The clinical ideal and the layperson’s beauty ideal, in other words, do not always coincide.