National Geographic News
A composite photo of digitally-created faces used to judge trustworthiness.

In the study, subjects judged digitally created faces for their trustworthiness and other characteristics.

DIGITAL ILLUSTRATION COURTESY OF EMILY COGSDILL, HARVARD UNIVERSITY  

Marc Silver

National Geographic

Published April 1, 2014

We've all looked at someone's face and thought: "Now there's someone I can really trust." Or perhaps: "I wouldn't trust him with a wooden nickel." To the surprise of social scientists, children as young as three make the same sort of judgments based on nothing more than facial features. That's what researchers found in a new study published in Psychological Science.

Mahzarin Banaji, Emily Cogsdill, and Elizabeth Spelke of Harvard and Alexander Todorov of Princeton showed pairs of faces to 99 adults and 141 children ages three to 10. Each pair of faces was designed to connect to one of three adjectives: trustworthy, dominant, or competent. (For kids, the terms used were mean/nice, strong/not strong, smart/not smart.)

The researchers expected the adults to be pretty much in agreement in picking who was trustworthy, dominant, and competent or not. And they were. The rate of consensus was in the 80 to 95 percent range. But they didn't expect a similar trend among the kids.

"We were dumbstruck when we found very young children have these preferences by age three," says psychology professor Banaji. Children aren't born with such biases, and the researchers imagined that "it was a slow process" to develop them.

But in the study, the kids basically reached the same conclusions as the adults, and, depending on the category, 65 to 97 percent of the children made the same calls. The lower range was for faces that were judged as strong/not strong. The upper number was for whether faces seemed mean or nice. "We would not have expected to see that data showing that [preferences] are present in near-adultlike forms pretty early," Banaji says.

The faces were computer-generated, designed by Todorov to remove distractions and to reflect what we already know about how faces are perceived. They were all men, to avoid gender bias, and they were all bald so that hair could not sway a subject's choice. From previous studies, researchers know that a person with a more feminine or childlike face, with eyes tilted upward, is considered more trustworthy than a person with a thick neck, jutting jaw, and heavy brow. So they used such traits to stack the deck.

Cute Equals Nice?

"It's hard to know exactly where this is coming from," says psychology graduate student Cogsdill of the kids' biases. Perhaps the appearances of characters on TV shows for kids, where cute equals nice, is a factor. "But we don't have a fully clear answer."

Of course, the problem with this kind of snap judgment is that you can't judge character by the face. Yet we all do.

The lesson for parents, Cogsdill says, "is to start a conversation with their children about making judgments based on appearances." Parents might also look for storybooks where the ugly duckling turns out to be the hero.

More research needs to be done to figure out the origins of this kind of facial bias. One way might be to look at whether primates have the same types of biases about other primates' faces. "You look at a bunch of monkey faces on the web and you might think, 'That's one mean monkey,'" says Banaji.

What about humans who have a face that, even at rest, looks unhappy? Cogsdill notes that the Internet snarkily discusses something called "resting bitch face"—people who, without trying to express any emotion, look angry or mean all the time. Would it help to make an effort to smile? "There's no data speaking to that," says Cogsdill (presumably with a smile), "but it seems reasonable to expect."

In the study, children and adults were shown pairs of computer-generated faces and asked to identify which of the pair was nicer, stronger, or smarter.

Click on the faces below to see how your preferences compare to those of the study participants.

18 comments
Jose Antonio V.
Jose Antonio V.

I think there is a database in each person's brain, which records the muscles and nerves that are activated with every feeling or learning. And that is why from 3 years of age you can infer what a face expresses.

L Bongers
L Bongers

The left one looks smarter because he looks into the camera, the other one is looking into the air as if he is thinking and unsecure.


The left one , to me is stronger mainly because of his neck that looks as if there are more muscles therefore i asume his body is also stronger.

Zoë Davis
Zoë Davis

I don't judge by face, so I couldn't choose

Sheri S.
Sheri S.

The first one looks smarter because the image is MEETING YOUR EYES; it has nothing to do with symmetry.  

The last one looks stronger because of the aggressive eyebrow slant, not because of a larger chin or nose.


I am not impressed with the "science" here.

Daniel Watson
Daniel Watson

"Of course, the problem with this kind of snap judgment is that you can't judge character by the face. Yet we all do."


Why cant we? I mean in general. You can generally tell a nice, or a dodgy person can't you? 

Charlie Debenham
Charlie Debenham

I found the first to be the most interesting...is it possibly because the man on the left is more Asian-looking rather than it being related to symmetry?

Andris Secis
Andris Secis

try refresh this page and click to different faces, the same....

Chryssanthi Ventouratos
Chryssanthi Ventouratos

The problem with these studies is that they ask participants to be biased and then they judge their bias. What I am trying to say is that if asked the question, "Here are two faces, pick which one is smart and which isn't", even participants that do not engage in that type of thinking feel that they just have to do it. However that does not mean that that is how they make decisions in their everyday life. 

Kids too. They try to please and to prove themselves.When asked a question in this context they will try to find references from the way they see smart people portrayed in cartoons or other TV programs or media and try to give the "correct" answer. 

Very few people and even fewer kids would say "I don't know" because there doesn't seem to be an option and most would feel "stupid" to do so.

Another bias perhaps?

Fiona Nicholls
Fiona Nicholls

Very interesting- I had no idea that we made these calls at such a young age. It's strange that 'dominant' is among the choices- I don't often consciously see someone and think of how dominant they are. Perhaps it's just something we do on a very base level?

fifijane.blogspot.co.uk

Kate Barsotti
Kate Barsotti

This is more about expression than features.

Jim Henry
Jim Henry

Yes, someone smiling is interpreted to be more happy/trustworthy than the guy staring with a stern face.  This is a sign that people can accurately interpret emotions based on faces and doesn't seem to prove that facial features have any impact.  Maybe this synopsis didn't explain this portion properly, but just my observation based on the test I took.  I'm sure dogs would be do pretty well at this test as well.

KENNETH LANE
KENNETH LANE

Then we are somewhat shallow in our decisions of others?

hi ho
hi ho

very interesting indeed

Charlie Debenham
Charlie Debenham

@Andris Secis  No it just says the same statement regardless of what you click, in all three the person on the left was the most commonly chosen.

David Wa
David Wa

@Chryssanthi Ventouratos  Surely that benefits the study, in that it assumes the importance of a snap decision being the most likely outcome when demanded. Therefore the analysis across age groups is likely to be more valid. 

Share

Featured Article

Latest From Nat Geo

See more photo galleries »

The Future of Food Series

See more food news, photos, and videos »