REDUCING AND MEASURING THE CROSS-RACE EFFECT

Abstract

Individuals recognise same-race faces better than different race faces, a robust phenomenon called the cross-race effect (CRE). The CRE can contribute to mistaken identifications, making it crucial to study it. This project included two studies. The first aimed to reduce the CRE using targeted interventions and to evaluate how individual differences, confidence, and response time predict eyewitness accuracy. The second focused on creating a reliable measure of eyewitness accuracy for White and Asian witnesses and targets. Both studies examined the confidence-accuracy relationship for White and Asian targets. The first study compared known interventions to reduce the CRE (discrimination training, individuation instructions) against no intervention and explored new variables— individual differences in working memory capacity, selective attention, and need for cognition—and extant variables—confidence and response time—predicting accuracy. Contrary to expectations, participants (White, n = 403) showed similar identification accuracy of Asian and White faces. A CRE was observed for target-absent accuracy. Discrimination training altered the CRE (cf. control group): it increased correct rejections for cross-race faces but decreased them for same-race faces. Working memory capacity, confidence, and response time reliably predicted identification accuracy. Confidence and response time explained unique variance, but the variance explained by working memory capacity overlapped with these. The second study developed the CRE Inventory, combining known predictors of the CRE to improve the prediction of same- and cross-race eyewitness accuracy. A CRE was observed for White participants (n = 202). However, Asian participants (n = 203) recognised White faces similar to Asian faces—potentially because of their minority status. Exploratory factor analysis produced reliable scales with expected factors that significantly predicted identification performance: general face recognition ability, race-specific face recognition ability, racial attitudes towards White individuals, quantity of contact with Asian individuals, motivation to individuate White individuals, and cognitive disregard of Asian individuals. Three scales that predicted identification accuracy also explained unique variance compared to that explained by the eyewitness’ confidence. This research advances the CRE literature theoretically—by identifying factors that relate to recognition—and practically—by testing multiple predictors of accuracy and developing a tool to enhance the reliability evaluations of White eyewitnesses.