Theories of Other-Race Face Identification

Five hypotheses have been offered to explain the "other-race" effect in face recognition (Ayuk, 1990; Chance & Goldstein, 1996):

1. The first hypothesis proposes an inherent difficulty between races. Members of individual racial groups are more difficult to differentiate from one another than from another racial group, hence the saying, "They all look alike." Shared similarities within racial groups create and help maintain this effect. The few experiments that investigated this proposition have shown mixed results. This hypothesis is difficult to test because difficulty in discrimination may not be due to physical sameness, but to inappropriate cue utilization. The hypothesis cannot be isolated from other effects enough to eliminate other possible factors in the other-race effect.
2. The second hypothesis proposed is that prejudicial attitudes may influence other-race identification. However, no correlation between identification and attitudes has been found (Lavrakas, Buri, & Mayzner, 1976, cf. Ayuk). Yarmey and Kent found no evidence of racial or attitudinal biases in race focus. Prejudicial attitudes toward infants, similar to the effects of the other-race effect in that "all babies look alike," showed no effect. Subjects in their study reported having positive regard toward infants. However, Brigham and Williamson (1979) found that a same-race bias did affect recognition memory of African Americans and Caucasians (cited in Yarmey, 1996).
3. Prior experience or knowledge of another race may also influence processing of a face. Previous exposure to other races is dependent on cultural issues which have been explored by Lindsay, Jack, and Christian (1991). They hypothesized that perceptual expertise was required for facial recognition. Quantity of experience affects ability to distinguish one face from another. Representations of similar faces can interfere with one another because faces appear to share the same storage space. Ng and Lindsay (1994) examined the contact hypothesis with Caucasians and Orientals from Canada and Orientals from Singapore. Contact of the other-race for Orientals from Singapore was severely limited. They were able to replicate the other-race effect in that Orientals recognized Oriental faces better than Caucasians, and Caucasians recognized Caucasian faces better than Orientals. However, differences between Canadian Orientals and Singapore Orientals for Caucasian faces was not significant. Thus, they concluded that the race issue was not related to country of origin.
4. A fourth hypothesis is that encoding strategies an individual uses for own-race recognition are also employed for recognition of other-race faces. These strategies are often less than perfect when viewing faces of another race. Although there is evidence that individuals use different cues in asking someone to describe faces of own versus other race faces, there is no evidence that links cue utilization and the other-race effect.
5. A fifth hypothesis is differential processing. Subjects use differential processing due to inferences and judgments made during initial viewing of own- and other-race faces. Related to this hypothesis is levels of processing (Craik & Lockhart, 1972). Faces of one's own race are processed deeply for possible character traits, whereas other-race faces are processed only superficially. The human face, however, is rich in perceptual information regardless of race. Part of the problem lies in perceived similarity of stimuli and elusiveness of a precise feature list used to describe a particular face. There appears to be no specific feature, but a group of features that defines a face of a given race, as has been noted in face recognition.

Traditionally, a face recognition task is used to examine the other-race effect (reviewed in Bothwell, Brigham, & Malpass, 1989). An initial set of target pictures of African American and Caucasian faces are shown to African American and Caucasian subjects. Then subjects are shown another set of faces including target pictures mixed randomly with a set of distracter pictures. Subjects are required to make an old judgment for pictures that are included in the target set and a new judgment for those faces that are not part of the target set. Subject responses are scored correct or incorrect, and recognition ability is assessed, for both correct and incorrect responses. Own-race bias exists if the recognition ability is greater for own-race pictures than for other-race pictures.

REFERENCES

Ayuk, R.E. (1990). Cross-racial identification of transformed, untransformed, and mixed- race faces. International Journal of Psychology, 25, 509-527.

Bothwell, R.K., Brigham, J.C., and Malpass, R.S. (1989). Cross-racial identification. Personality and Social Psychology Bulletin, 15, 19-25.

Chance, J.E., & Goldstein, A.G. (1996). The other-race effect and eyewitness identification. Psychological Issues in Eyewitness Identification (153-176) Sporer, S.L., Malpass, R.S., & Koehnken, G. eds. NJ: Lawrence Erlbaum.

Ng, W., & Lindsay, R.C.L. (1994). Cross-race facial recognition: Failure of the contact hypothesis. Journal of Cross-Cultural Psychology, 25, 217-232.

Yarmey, A.D., & Kent, J. (1980). Eyewitness identification by elderly and young adults.
Law and Behavior, 4, 359-371.