In a study published today in the Royal Society’s journal Biology Letters, my colleagues and I ask if there is any validity in judgements of sexual faithfulness made from the faces of unfamiliar men and women. The answer? Well, read on …

We know, without having to test it, that we make snap judgements about strangers based on their appearance – even though, as sentient beings, we strive not to.

Axioms such as “never judge a book by its cover” and “looks can be deceiving” underscore the point. But is there any validity in the judgements we try so hard not to make?

What we did

We used colour photographs of 189 Caucasian adult faces (101 males and 88 females) from a previous study, in which we had explored the relationship between the lifetime number of self-reported sexual partners and facial attractiveness.

In that study we asked men and women to report on sexual unfaithfulness — how frequently they had been involved in a relationship with someone who already had a stable partner (poaching), or a relationship with a third party when they themselves were in a stable relationship (cheating).

For our new study, 68 self-reported heterosexual, adult Caucasians (34 males and 34 females) from the University of Western Australia participated in the study for course credit. The men were aged between 17 and 48, the women between 17 and 45.

In the photos participants were shown, a mask hid most of the subjects' hair, but left the face contour and inner hairline visible.

Our study participants were shown these opposite sex images and asked, on a scale of 1 (not very likely), to 10 (very likely): “How likely is this person to be unfaithful?” in a sexual context.

Women were able to assess mens' unfaithfulness with a modest degree of accuracy (incorrectly rating an “unfaithful” face as “faithful” on 38% of occasions). Facial masculinity was found to be correlated both with womens' ratings of unfaithfulness and the extent to which the rated man had actually engaged in sexual cheating and poaching.

In contrast, men appeared unable to rate womens' unfaithfulness with any validity (incorrectly rating an “unfaithful” face as “faithful” on 77% of occasions).

Attractiveness and femininity were highly correlated with unfaithfulness ratings, and each other, indicating that men perceived attractive, feminine women as likely to be unfaithful. But there was no evidence that they were. Attractive women were rated as more trustworthy, as we’ll discuss shortly.

If you’re curious, the full results table can be seen here in the Biology Letters paper.

A question of trust

We also asked a second group of raters how generally trustworthy the individuals in the photographs appeared.

Interestingly, there were no associations between ratings of untrustworthiness and unfaithfulness for either men rating women’s faces or women rating men’s faces. Nor were ratings of trustworthiness related to reported sexual behaviour.

This suggests that our impressions of faithfulness are distinct from our impressions of trustworthiness.

A picky past

We can only speculate on the reasons why women seem able to assess unfaithfulness with greater accuracy than men. It may be that in our evolutionary past selection for accuracy in mate choice for women has been stronger, perhaps due to the greater costs of making poor choices.

For most animals with biparental care, including humans, females invest more heavily in reproduction and have more to lose should their partner desert.

Our work builds on other recent studies that suggest our immediate impressions of strangers may contain a kernel of truth.

In a 2008 study in the Canadian Journal of Behavioural Science, the faces of Nobel Peace Prize winners and humanitarians were rated as trustworthy at slightly above chance, while faces of America’s most wanted were rated as untrustworthy.

Where we’re at

In our study, accuracy in determining unfaithfulness was certainly modest and may be limited to women. Nevertheless, our results seem to demonstrate that accurate judgements of unfaithfulness can be made from the face alone, in the absence of behavioural cues.

There is certainly much more to be done in this area of research.

The models and raters we used were exclusively Caucasian and from a university environment. Further research will determine whether our findings generalise beyond this limited sample group.

Likewise, our paper – by necessity – relies on self-reports of infidelity, and these must always be interpreted with caution.

All that aside, our findings could well suggest that greater credence should be given to womens' intuition.

This work was conducted in association with the ARC Centre of Excellence in Cognition and its Disorders.