An investigation into the cross-race effect for composite construction

Abstract

The cross-race effect is a well established phenomenon in face recognition research (see Meissner & Brigham, 2001; Sporer, 2001 for reviews). It is possible that the own-race advantage might be transferable to the construction of face composites For example, Ellis et al. (1979) found an advantage of white Scots over black South Africans on white composites as predicted by the cross-race effect while black South Africans did not show an own-race advantage. Ellis et al attributed the latter to the more limited range of black features available in the composite system used. Even though the range of features in commonly used systems have expanded over the years, Davies and Valentine (in press) argue that even in these systems features for non-Caucasian faces are still not well represented.