Abstract
Adults recognize own-race faces more accurately than other-race faces, a pattern called the other-race effect. We previously reported that Caucasian adults were more sensitive to differences among faces in both feature shape (e.g., eyes) and feature spacing (e.g., the distance between the eyes) for Caucasian faces than for Chinese faces. However these effects were surprisingly small (M difference = 5.6% and 9.9%) given the difficulty adults experience in recognizing other-race faces on a daily basis. Here we tested whether storage is better for own-race than other-race faces by varying the delay (1s, 5s, or 10s) in a delayed match-to-sample task. In Study 1 (n=24) we used featural and spatial manipulations of a single identity per race and a blank screen was presented during the delay. There was an effect of delay for both face sets, ps ps [[gt]] .5. To more closely mimic the real world, in Study 2 (n=24) we presented altered versions of two identities per race and a screen comprised of multiple Chinese and Caucasian faces was presented during the delay. The effect of delay and the face race x delay interaction were significant only for the spacing set, ps [[lt]].02. The drop in accuracy in the 10s-delay condition relative to the 1s-delay condition was larger for other-race (12%) than for own-race faces (6%). Collectively, these results suggest that the own-race advantage for feature shape occurs at the encoding stage, whereas the own-race advantage for the spacing set may occur at both encoding and storage (see Freire et al., 2000 for similar analyses of the inversion effect). The own-race advantage may be small in lab studies because cues that adults might normally rely on when encoding other-race faces (e.g., hair, clothing) are removed.