We all know that you shouldn’t judge a book by its cover. But we all do it anyway. If I were to show you some faces, you’d find it pretty easy to make a snap judgment of –say – how clever they look. But this would just be prejudice, right? You couldn’t, just by looking, guess people’s actual intelligence. Could you?
Let’s try. Below are three men. Rank them in order of intelligence from most to least (these aren’t real people, but composites created – in each case – by averaging across lots of difference faces).
Scroll down to find out the rankings that most people give.
For the male faces, most people rate the man on the right as the most intelligent, and the man on the left as the least. For the female faces, most people rate the woman on the left as the most intelligent, and the woman on the right as the least (in both cases, the middle faces are in between).
But these are just stereotypes, right? No. When researchers in the Czech Republic asked people to estimate the intelligence of 80 biology students on the basis of facial photographs, a significant correlation was observed between perceived and actual intelligence (as measured by an IQ test); but only for men.
Admittedly, the relationship was small. Statisticians express the relationship between two measures using a number called a correlation coefficient (sometimes abbreviated to r). This number ranges between 0 (i.e., no relationship between face-ratings and IQ-test scores) and 1 (i.e., you can perfectly predict anyone’s IQ on the basis of the face-rating). On this scale, the correlation between perceived and actual intelligence for men was tiny[i] – r=0.06 – though easily statistically significant. The concept of statistical significance is explained in detail in the book version of Psy-Q (look for the section The Tea Test), but the take-home message is this: The likelihood of researchers finding an apparent relationship between perceived and actual intelligence of this size (i.e., r=0.06 on the 0-1 scale) by chance alone, if none actually existed, is less than 1 in 1,000.
Two unsolved puzzles remain. First, why is it that we can read intelligence in the faces of men (at least to a modest degree), but not women? Perhaps, in evolutionary terms, giving off signals of intelligence is more important for men than women. Indeed, cross cultural studies tell us that – pretty much universally – women value intelligence much more than do men when seeking a partner (You can take this “Perfect Partner” test yourself in the book version of Psy-Q, as well as measuring your own IQ, and completing other face-rating tasks looking at trustworthiness, aggression and attractiveness).
Another possibility is that, when asked to rate intelligence in women’s faces, the raters (both men and women) couldn’t help being swayed by attractiveness (what researchers call a “halo effect”). Indeed, the relationship between rated attractiveness and rated intelligence for female faces was r=0.9 on the 0-1 scale, meaning that the two are almost interchangeable (the corresponding figure for male faces was r=0.5).
The second remaining question is just what features of men’s faces raters were using to predict intelligence (which, remember, they were able to do at way above chance levels). The researchers found that, on the whole, faces that were rated as intelligent tended to be long and thin with larger noses, while those that were rated as less intelligent tended to be rounder and fatter with small noses. But, confusingly, when they fed these measurements into the computer, they did not predict IQ-test scores, just perceived intelligence.
So, for now, this second puzzle remains unsolved: We don’t know exactly what features people use to correctly predict intelligence in male faces. The researchers speculate that it may be “particular configurations of eyes or gaze, colour of eyes, hair and skin, or skin texture”. What we need now is studies that tease apart these possibilities (for any psychology students reading this, this might be an interesting idea for a 3rd year project, if you can find a suitable supervisor).
In the meantime, the findings of this study suggest that people not only do judge a book by its cover but – to at least some extent – are justified in doing so. So, whilst we don’t yet know exactly what this entails, then the lesson – particularly if you are a man – is clear: Look sharp!
This is a sample chapter in the style of Psy-Q by Ben Ambridge, forthcoming from Profile Books (UK) and Penguin (US). For details click here
Kleisner, K., Chvátalová, V., & Flegr, J. (2014). Perceived Intelligence Is Associated with Measured Intelligence in Men but Not Women. PloS one, 9(3), e81237.
Buss, D. M. (1989). Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioral and Brain Sciences, 12, 1–14.
[i] One reason that the effect is tiny is that the researchers had to control for ratings of attractiveness, which is highly correlated with perceived intelligence (r=0.5 for men).