People are more accurate at recognising individuals from their own race than those who come from other racial groups, the Cross-Race Effect (CRE). We explored whether this effect was still evident for faces that had been seen multiple times. Black and White participants viewed five blocks of Black and White faces presented in a continuous series with 30 target faces presented in the first block. In Blocks 2–5, target faces were repeated and intermixed with new faces. The impact of this familiarisation was markedly different for hits or false alarms. For instance, the false alarm rate data, as well as the d' data, showed a CRE across all 5 blocks, and repetition had no reliable effect on the magnitude of the CRE. In contrast, the hit rate data showed no CRE at all. Consistent with prior findings, though, our CRE pattern was asymmetric – plainly evident with White participants, but not Black participants. Overall, our results suggest that in the realm of face recognition, hits and false alarms are responsive to different mechanisms and are dissociable in contexts such as ours. These results have important implications for identifications and misidentifications of familiar same- and cross-race faces in the real world.
Marsh et al. (Fri,) studied this question.