Social perception of faces
Research shows that people a) form instantaneous impressions of other people based on facial appearance; b) agree in these impressions; and c) often act on these impressions. Working with a computational framework, in which faces are represented as points in a multidimensional space, I describe two lines of research on identifying the basis of social impressions: configurations of facial features and statistical learning of face properties. Specifically in the first part of the talk, I will describe data-driven, reverse correlation methods that can be used to model evaluation of faces on any social dimension (e.g., trustworthiness) and to identify the perceptual basis of this evaluation. These methods provide an excellent discovery tool for mapping configurations of face features to specific social inferences. In the second part, I will describe research suggesting that statistical characteristics of the face – determined by the face’s position in a distribution of faces – can shape social perception. This research shows that exposure to different statistical distributions of faces leads to shifts not only in the perceived typicality of faces but also in social judgments. This statistical learning framework suggests a new parsimonious account of cultural, inter-group, and individual differences in social perception of faces.