­­­­­The other-race effect in face recognition: Do people shift criterion equally for own- and other-race faces?

Abstract

ABSTRACT People are better at recognizing own-race faces than other-race faces. This other-race effect in face recognition typically manifests in sensitivity (i.e., better discrimination). However, research also consistently finds that people respond more liberally (and thus produce more false alarm errors) when recognizing other-race faces, rather than own-race faces. In an old/new recognition paradigm, we examined whether providing White participants with base rate information could attenuate the differences that they demonstrate, both in terms of response bias and sensitivity, when attempting to recognize Black and White faces. During the test phase of an old/new recognition task, participants viewed own-race and other-race faces surrounded by a red or green border. Participants were informed that these borders indicated a 25 percent or 75 percent likelihood of the face having been presented earlier. Border cues were either valid (Experiment 1) or invalid (actual probability was 50%; Experiment 2). Across both experiments, participants shifted their response criterion in line with border cues but did so equally for own-race and other-race faces. This finding suggests that the liberal bias for other-race faces is likely to originate early during encoding and is unlikely to be moderated by manipulations that target later, decision-making processes operating during retrieval.