Abstract
Although a large body of evidence in the face recognition literature indicates that we are better at recognizing faces from our own race than faces from other races (i.e., the Other-Race Effect or ORE), the precise mechanisms mediating the ORE are not well understood. It has been speculated that people encode in-group, own-race faces as individuals and classify out-group, other-race faces in terms of race. According to the perceptual expertise hypothesis, the more specific, subordinate level of recognition of own-race faces demands a finer grain of perceptual analysis. It has been suggested that this type of expert perceptual analysis is selectively tuned to the recognition of within-race faces, but does not apply to the recognition of between-race faces. ORE research is complicated by the fact that participants differ in the degree and kind of own- and other-race familiarity. In the current study, we addressed the familiarity issue by creating two novel races, the Thutmosians and the Guansians. Members of the Thutmosian race differed in the upper eye region of the face where members of the Guanshian race varied in the lower mouth region. As a direct test of the perceptual expertise hypothesis, participants were trained to either individuate Thutmosian faces and categorize Guanshian faces or vice versa. Recognition performance was assessed with a new set of Thutmosian and Guanshian faces in an old/new recognition test administered before and after training. The main finding was that individuation training produced reliable gains in recognition of faces from the individuated race whereas categorization training had little effect on the recognition of faces from the categorized race. Consistent with the perceptual expertise hypothesis, our results suggest that it is the kind, not the amount of racial experience that determines the ORE.
Meeting abstract presented at VSS 2012