Purchase this article with an account.
William G. Hayward, Gillian Rhodes, Chris Winkler, Adrian Schwaninger; Own-race face effects in processing of configural and component information by Chinese observers. Journal of Vision 2005;5(8):534. doi: 10.1167/5.8.534.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
People are generally better at recognizing own-race than other-race faces. Many current accounts of face recognition argue that this phenomenon occurs because recognition of own-race faces is based on configural processing, which requires perceptual expertise. Without such expertise, recognition of other-race faces must rely on individual components. We tested this hypothesis in Hong Kong Chinese participants, using photographic stimuli of Chinese and Caucasian faces. In Experiment 1, sets of faces were created using Photoshop such that either configural information (distances between eyes, nose, and mouth) or component information (darkness of eyebrows and lips; shape of nose) was different between two exemplars of the same face. Participants viewed a target image, then were shown two images and were asked to identify the target. Faces could be Chinese or Caucasian, and upright or inverted. Participants showed better detection of changes to Chinese than Caucasian faces, and to upright than inverted faces; however, neither the size of the own-race detection advantage nor the inversion decrement varied between configural or component changes. In Experiment 2, participants initially learned 10 intact Chinese faces and 10 intact Caucasian faces. An old-new recognition paradigm was used at test, with both target and distractor faces of each race being either cut into components and then scrambled (requiring recognition based on features) or intact but blurred (requiring recognition based on configuration). Participants showed an own-race advantage, and better recognition of blurred than scrambled faces; however, the size of the own-race advantage did not vary across change type. In conclusion, both experiments found own-race advantages for both configural and component information. These results suggest that own-race expertise can be sensitive to both types of facial information.
This PDF is available to Subscribers Only