The Chicago Face Database
Led by Debbie Ma, we have developed a  of faces including images of more than 600 individuals, female and male, from a variety of racial and ethnic groups (Asian, Black, Latino, White). For many of the individuals, we have a variety of emotional expressions (the database is still being developed, and for some, the emotion images are not yet ready for distribution).Â
Critically, these faces have been painstakingly measured and extensively rated by large samples of participants, providing measures of physical dimensions as well as norming data on dimensions like attractiveness, prototypicality, age, babyfacedness, etc.
A model of face processing: Perceptual Enrichment, Expectancy, & Reciprocity (PEER)
We have recently advanced a general model of face processing intended to account for several race-related effects, such as better recognition of same-race faces, and faster classification of other-race faces. The model argues that perceivers engage in two parallel (but interactive) streams of processing. One stream rapidly compares a coarse representation of a novel face to the perceivers mental model of a "typical" face. The other stream synthesizes information about the face in a manner that prioritizes dimensions that have proven useful or diagnostic in previous experience.
Preferential Attention to Outgroups
We have conducted a number of studies on attention to race. We typically find that White participants preferentially attend to other-race faces (including Black and Latino faces). Our initial arguments involved the idea that preferential attention emerges because outgroups are often associated with threat. We have shown that this attention is flexible, depending on perceiver goals, and changing over the course of an experimental session. Steffanie Guillermo is leading an effort to tell a more complete story about the causes of race-based attention.
Hemispheric Asymmetries in Face Processing
Face recognition relies heavily on holistic processing, which tends to be lateralized in the brain -- it relies heavily on the left hemisphere. We have found evidence that this lateralization is more pronounced when processing same-race (as compared to other-race) faces. Critically, participants who report more interracial contact show a smaller effect of race.