Abstract
The facial width-to-height ratio (FWHR), a size-independent sexually dimorphic property of the human face, is correlated with aggressive behaviour in men. Furthermore, observers' estimates of aggression from emotionally neutral faces are accurate and are highly correlated with the FWHR. In a series of experiments we tested if the FWHR is the basis of observers' accuracy in estimating aggressive propensity from emotionally neutral faces. In Experiments 1a-c, estimates of aggression remained accurate when faces were blurred or cropped, manipulations that reduce featural cues but maintain FWHR. Accuracy decreased when faces were scrambled, a manipulation that retains featural information but disrupts the FWHR. The estimates of aggression were highly consistent across observers for all conditions except the scrambled condition. Overall, estimates of aggression were most accurate when all facial features (even if blurred) were presented in their canonical arrangement, allowing for perception of the FWHR, with at most a small contribution from the appearance of individual features. There was no explicit use of the FWHR; 84% of participants indicated that “the eyes” were the basis for their judgement. No participant reported using any kind of configural information, including the FWHR. Nonetheless, in Experiment 1d, participants given instruction about the FWHR were able to accurately estimate the FWHR of faces presented for 39 msec. In Experiment 2, computer-modeling software (FACEGEN) identified eight facial metrics that correlated with estimates of aggression; regression analyses revealed that FWHR was the only metric that uniquely predicted these estimates. In Experiment 3, faces were manipulated to create pairs that differed only in FWHR. Participants' judgement of which individual of the pair was more aggressive was biased towards faces with the higher FWHR. Together, these experiments support the hypothesis that the FWHR is an honest signal of propensity for aggressive behaviour.