Other-race faces: Limitations of expert face processing

Abstract

Adults’ expert face recognition is better for the kinds of faces they encounter on a daily basis (typically upright human faces of the same race). Adults process own-race faces holistically (i.e., as a gestalt) and are sensitive to small differences among faces in the spacing of features and in the appearance of individual features. Various tasks are used to measure each of these components of face processing in the literature; for example, the composite face task and the part/whole task are used interchangeably as measures of holistic processing. To measure the specificity of adults’ expertise we administered a battery of five tasks comprised of Caucasian and Chinese faces to two groups of participants: Caucasian adults (n=31) living in rural Pennsylvania and Chinese adults (n=32) living in Guangzhou who had minimal exposure to other-race faces. The battery included one measure of memory for faces, two measures of holistic processing (composite face task, part/whole task), and two measures of sensitivity to spatial and featural cues (Jane/Ling task, scrambled/blurred task). The race of face x race of participant interaction was significant in three conditions: the memory task, one measure of featural processing (Jane/Ling task) and one measure of spatial processing (blurred faces task), all ps<.02. There was no interaction in either measure of holistic processing (composite face task, part/whole task) or in the other measures of featural (scrambled faces task) and spatial (Jane/Ling task) processing, ps > .10. These results indicate that adults process both own- and other-race faces holistically, but are less sensitive to the spacing of features and the appearance of individual features in other-race faces, at least under some conditions. Surprisingly, individual scores on the three pairs of tasks designed to measure the same aspect of face processing were not correlated (ps>.20), indicating that they may be tapping different processes.