When you meet someone for the first time, can you tell that person’s personality just by looking at their face? If you think you can, you are not alone: Recent research showed that people across different cultures from 41 countries make inferences about a person just by looking at their neutral-expression face—for example, how much the person looks like they could harm you and how dominant they seem.

Okay, people might think they are reading others’ faces, but are they really good at it?

Past research yielded mixed results, with some studies suggesting that we are pretty accurate while others showed that our predictions are not much better than random guesses. My colleagues and I aimed to replicate these studies and test the accuracy of these first impressions on large groups of participants from the U.S. and Turkey.

We focused on two sets of personality traits: The commonly studied “Big Five” (emotional stability, extraversion, openness, agreeableness, and conscientiousness) and three especially negative “Dark Triad” traits:

  • Narcissism – arrogant, vain, pompous, self-absorbed
  • Psychopathy – reckless, antagonistic, assertive
  • Machiavellianism – manipulative, scheming

We used a set of emotionally neutral facial photographs of people who had already been rated on these personality traits, both by themselves and by people who knew them. In our study, we used composite images, which were averages of multiple faces with a certain trait. For example, one composite image was the composite of people with high narcissism, while another was the composite of people with low narcissism. We presented both of these composite images side by side and asked people to estimate which one has a higher level of each personality trait.

We found that people’s guesses were notably above the chance level of 50% (almost all above 60%, sometimes as high as 74%) for all the Dark Triad personality traits in both American and Turkish samples. This suggests that people not only notice and care about the subtle cues in others’ faces but can also detect the level of dark traits to some extent, although not with 100% accuracy. We also found that people could guess extraversion, agreeableness, and conscientiousness although the results for the Big Five traits were more mixed and variable across different groups, compared to the Dark Triad.

These results suggest that our first impressions regarding the Dark Triad traits, which are all malevolent and socially undesirable personality traits, are better than random guesses, although they are not always 100% accurate. This might be the result of an evolutionary adaptation for staying away from potentially dangerous people. One could speculate that people who were more successful in their spontaneous detection of others’ ill intentions would have the upper hand throughout our evolutionary history.  

Although the results are interesting, there are also important caveats. Although in our study, participants were successful at distinguishing two composite images with different levels of Dark Triad traits, this does not necessarily mean that they would accurately assess all humans’ faces; maybe they were just lucky with these particular composites of faces. Also, in real life people do not always form their first impressions based only on facial appearance, since in reality, other factors can contribute: the context of the interaction, ethnic/racial biases, facial expressions, gender, and self-resemblance of the target are only some of the factors that can alter a first impression.

Our findings, similarly to some of the past research, showed that people are at least partially accurate in their perceptions of strangers’ faces. However, this is far from being conclusive, considering the potential limitations. Future research will keep trying to understand whether people can read others’ personalities from their faces in different contexts, and if yes, how exactly they can do that.

For Further Reading

Alper, S., Bayrak, F., & Yilmaz, O. (2021). All the Dark Triad and some of the Big Five traits are visible in the face. Personality and Individual Differences, 168, 110350. https://doi.org/10.1016/j.paid.2020.110350

DeBruine, L. (2020, January 31). Composite images. https://debruine.github.io/posts/composite-images/

Jones, B. C., DeBruine, L. M., Flake, J. K., Liuzza, M. T., Antfolk, J., Arinze, N. C., ... & Sirota, M. (2021). To which world regions does the valence–dominance model of social perception apply? Nature Human Behaviour, 5, 159-169. https://doi.org/10.1038/s41562-020-01007-2

Todorov, A., Olivola, C. Y., Dotsch, R., & Mende-Siedlecki, P. (2015). Social attributions from faces: Determinants, consequences, accuracy, and functional significance. Annual Review of Psychology, 66, 519–545. https://doi.org/10.1146/annurev-psych-113011-143831.

Sinan Alper is an Assistant Professor of Psychology at Yasar University in Turkey. His research interests include moral and political attitudes, social cognitive processes underlying these attitudes, and face perception.