It’s a Hollywood stereotype: Men prefer to partner up with feminine-looking women, and women favor masculine men. But even when you allow for same-gender couples and variations in personal preference, plenty of research suggests that the proposition is generally true. “It’s been replicated many times across different cultures,” says Isabel Scott, a psychologist at Brunel University in Uxbridge, on the outskirts of London, “so people tend to assume it’s universal.” A new study in Proceedings of the National Academy of Sciences challenges that thinking, however.
Historically, human studies have shown that women with more feminine faces tend to have higher estrogen levels, which are in turn associated with reproductive health. In men, the argument is that masculine-looking faces are associated with stronger immune systems—always a good thing in a mate, especially if that trait is passed on to the kids. Masculine appearance may also a sign of a dominant and aggressive personality, but our distant female ancestors might plausibly have gravitated toward these men anyway, for the sake of their children’s health.
These theories fall under the rubric of evolutionary psychology—the idea that many of our fundamental behaviors have evolved, just as our bodies did, to maximize reproductive success. But as in many cases with evolutionary psychology, it’s easier to come up with a plausible explanation than to demonstrate that it’s correct. In this case, says Scott, “the assumptions people were making weren’t crazy. They just weren’t fully tested.”
To correct that, Scott and the 21 colleagues who put together the new study used computer simulations to merge photos of men’s and women’s faces into composite, “average” faces of five different ethnicities. Then they twirled some virtual dials to make more and less masculine-looking male faces and more or less feminine female versions. (“More masculine” in this case means that they calculated the specific differences between the average man’s face and the average woman’s for each ethnicity, then exaggerated the differences. “Less masculine” means they minimized the differences. Same goes, in reverse, for the women’s faces.)
Then they showed the images to city-dwellers in several countries and also to rural populations in Malaysia, Fiji, Ecuador, Central America, Central Asia and more—a total of 962 subjects. “We asked, ‘What face is the most attractive’ and ‘What face is the most aggressive looking,'” says Scott.
The answers from urban subjects more or less confirmed the scientists’ expectations, but the others were all over the place. “This came as a big surprise to us,” Scott says. “In South America,” for example, “women preferred feminine-looking men. It was quite unexpected.”
If these preferences had an evolutionary basis, you’d expect them to be strongest in societies most similar to the ones early humans lived in. “These are clearly modern preferences, though,” Scott says, which raises the question of why they arose.
One idea, which she calls “extremely speculative at this point,” is that when you pack lots of people together, as you do in a city, stereotyping of facial characteristics might be a way of making snap judgements. “In urban settings,” she says, “you encounter far more strangers, so you have a stronger motive to figure out their personalities on zero acquaintance.”
Read next: Wide-Faced Men: Good Guys or Bad?