Show a stranger a photograph of a teenager and ask how old she is, and the odds are good they will guess too high. Hand them a picture of a woman in her sixties and the guess tends to drift the other way, shaving off a couple of years. Somewhere in the middle, around the early forties, the eye gets it roughly right. This is not a quirk of one careless observer. It is a pattern that held across more than sixteen thousand judgements made by 308 women looking at one another.
The work, led by Yuanyuan Diao and colleagues at the Beiersdorf Innovation Center Shanghai, set out to do something the cosmetics industry has wanted for years: turn the slippery idea of “looking your age” into a number you can actually measure. And measure reliably, with ordinary people doing the looking rather than a trained dermatologist or a machine.
Perceived age, as the field calls it, is an old notion. Francis Galton was poking at it back in the 1880s, noticing that some people simply read as older or younger than their birth certificates would suggest. For a long time it stayed a curiosity, a thing everyone noticed and nobody quantified. Only with standardised photography, and then digital imaging, did it harden into something a study could pin down. The Shanghai team wanted to take the next step and validate a version of it that leans on lay judgement, the kind of judgement that actually operates in the world when someone glances at your face across a table.
How You Build a Number Out of a Glance
The setup was meticulous, almost clinical. Chinese women aged 15 to 65, evenly spread across ten five-year bands, were photographed under controlled light with their hair tucked away and every accessory removed. A dermatology technician scored their faces against the Asian Skin Aging Atlas, grading wrinkles and folds and sagging on fixed scales. Then, on a separate visit, the women sat a metre from a 27-inch screen and guessed the ages of strangers from the photo set, working only from faces in age bands adjacent to their own.
Why neighbours rather than a single panel of experts? Because people read faces of their own generation more accurately, and stacking sixty-odd independent guesses per face and taking the median washes out the wild outliers. The average woman, it turned out, was judged 1.6 years older than her actual age. A small number on its own. But the average hides the interesting bit.
The young got the worst of it. Teenagers between 15 and 20 were pegged, on average, more than five years older than they were. The oldest group, 61 to 65, went the other way and were read as 2.4 years younger. The researchers describe perceived age as an integrated, holistic measure rather than a tally of separate flaws, which is part of why it tracks biological aging so well.
The Folds That Give Us Away
So what is the eye actually catching? Across every group, two features dominated: the nasolabial fold, the crease running from nose to mouth corner, and the marionette fold below it. These topped the list whether a face was being aged up or down. After them came the softer, less obviously “wrinkle” cues, the tightness of the facial contour, the evenness of skin tone, and radiance. The things, in other words, that no single line on a face can account for.
The statistics carry a sting for the young. In the teenage group, overall facial sagging and under-eye lines more than doubled the odds of being judged older, presumably because any blemish on an otherwise smooth, even, firm young face stands out as something gone wrong, an anomaly the eye reads as premature aging. For older faces the logic flips: assessors lean on whether the prominent markers are present or absent, so a sixty-something who has kept her skin even and bright can read as unusually well preserved. Dullness and yellowness, meanwhile, pushed older faces older still, in the oldest band raising the odds of overestimation by more than two-thirds.
Crow’s feet, oddly, barely registered, ranking 17th of 25 features here, even though a separate multi-ethnic study had flagged them as a major tell in Chinese women. Faces, clearly, do not hand over their secrets in quite the same way to every set of eyes.
There are caveats worth keeping in view, and the authors are candid about them. The whole cohort came from a single city, Guangzhou, with one climate and one broad skin phenotype, so the numbers may not travel to other regions, never mind other ethnicities. There is also the awkward fact that the assessors knew they were in an aging study, which could have nudged them to scrutinise faces more harshly than they would in a café. And the study was paid for by Beiersdorf, the company behind Nivea and Eucerin, whose interest in a quick, cheap, consumer-friendly aging yardstick is not exactly disinterested.
Still, the practical pull is obvious. A trained grader needs roughly 30 seconds per facial parameter; a perceived-age guess takes about three. Push the photos online and you can gather a crowd of assessors in an afternoon rather than booking a clinic for three months. For a company wanting to know whether a cream actually makes someone look younger, rather than whether it nudged a wrinkle score by a decimal, that is a tempting shortcut.
What lingers, though, is the smaller, stranger finding buried in the data. We are hardest on the young, quickest to mistake a single early line for the onset of age, and oddly generous to the old. The face we present to the world is read not against some absolute scale but against what a stranger expects a face like ours to look like. Whether the next wave of skincare can actually shift that verdict, or merely measure it more cheaply, is the question the industry has yet to answer.
DOI: 10.1016/j.jdsct.2026.100148
Frequently Asked Questions
Why are young people judged as older than they are?
On a smooth, even, youthful face, any small sign of aging stands out sharply against what observers expect, so the eye over-weights it. In this study, facial sagging and under-eye lines more than doubled the odds that a teenager would be guessed older than her real age. The same features on a sixty-year-old barely move the needle, because they are already expected.
What facial features most strongly drive how old we look?
Two folds dominate across all ages: the nasolabial fold from nose to mouth, and the marionette fold beneath it. After those, observers rely on the tightness of the facial contour, the evenness of skin tone, and skin radiance, all features that no single wrinkle can capture. Crow’s feet, interestingly, ranked low here despite being prominent in some other studies.
Is perceived age actually a reliable measurement?
In this cohort it tracked closely with expert-graded aging features, and pooling more than sixty independent guesses per face and taking the median smooths out individual error. That said, the assessors all came from one city and knew they were in an aging study, so some bias is hard to rule out entirely. It is best treated as a practical complement to clinical grading rather than a replacement.
Could this change how anti-aging products are tested?
That is the explicit hope of the team behind it. A perceived-age judgement takes about three seconds against roughly thirty for a single clinical parameter, and photos can be assessed by large online panels rather than in a clinic. It also measures what consumers actually care about, looking younger overall, rather than an abstract wrinkle score.
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.