Despite the injunction "Don't judge a book by its cover," considerable research demonstrates that people have a strong tendency to use facial appearance when forming first impressions of others' psychological traits, and these impressions show considerable consensus across perceivers. Ongoing research in my social perception laboratory addresses three questions concerning this phenomenon: (1) What are the facial qualities that influence trait impressions? (2) Why do perceivers respond as they do to these particular facial qualities? (3) What are the social and psychological consequences of judging others by their appearance?

This research takes a functional approach to social perception, and it is assumed that social perceptions based on appearance either should be accurate or should reflect perceptual biases that serve some general adaptive function. A working hypothesis is that the evolutionary importance of detecting attributes like emotion, age, or genetic fitness has produced such a strong tendency to respond to the facial qualities that reveal these attributes that these responses are overgeneralized to people whose faces merely resemble a particular emotion, age, or level of fitness.

Most of the research in my laboratory has focused on the "babyface overgeneralization effect." This work has established that people of all ages whose facial qualities resemble those of infants (e.g., large eyes, round faces, small chins) are perceived to have childlike traits and are treated differently from the mature-faced in real-world venues, such as employment and the justice system. The psychological consequences of the babyface stereotype and other appearance stereotypes are currently being investigated through the use of longitudinal data archives that enable us to determine the effects of appearance on social outcomes and psychological development across the lifespan.

Also under investigation is the validity of the babyface overgeneralization hypothesis as well as other overgeneralization hypotheses. To this end, we are using connectionist modeling of face perception to determine, for example, whether the tendency to attribute childlike traits to particular adults can be predicted from a neural network's tendency to confuse the faces of those adults with the faces of babies. We also are testing the overgeneralization hypotheses through fMRI research.