Overcoming the other-race effect in face recognition
The other-race effect refers to our reduced ability to recognise faces that belong to a race different from our own. This dissertation aimed to enhance our understanding of this phenomenon using three strands of research investigating different aspects of the other-race effect. Chapter 1 provides an overview of the other-race effect literature, with a particular emphasis on social-cognitive and perceptual expertise explanations of the other-race effect. In particular, the social-cognitive perspective suggests that the other-race effect is driven by our tendency to ‘individuate’ in-group members as heterogeneous individuals while ‘categorising’ out-group members as homogeneous entities (Hugenberg et al., 2007). Chapter 2 tests social-cognitive explanations of the other-race effect using a novel experimental manipulation designed to promote individuation encoding or categorisation encoding of faces. We found that categorisation encoding eliminated the other-race effect when participants viewed a small number of faces (Experiment 2), but not when they viewed a large number of faces (Experiment 1). However, inconsistent with social-cognitive predictions, we found that individuation encoding failed to moderate the other-race effect, regardless of how many faces were viewed.
Relevantly, Chapter 2 also found that Caucasian participants individuated Black faces less accurately than White faces during the encoding task itself. Chapter 3 further tested the perceptual basis of the other-race effect using a sequential matching task in which two faces were presented very rapidly (100ms) with minimal demands placed upon episodic memory. The sequential matching task was tested firstly with Caucasian participants (Experiment 1) and secondly with a fully crossed-over design involving Caucasian and Black participants (Experiment 2). We found that Caucasian participants produced more false alarms when viewing Black faces rather than White faces, whereas Black participants did not. We suggest that the increased difficulty with which Caucasian participants distinguish or individuate Black faces might explain why individuation encoding manipulations such as those used in Chapter 2 and other studies have failed to reduce the other-race effect in later recognition.
Finally, considering that the other-race effect literature focuses heavily on recognition sensitivity, the dissertation turns attention towards the relatively under-researched question of response biases. Given that people typically demonstrate a more liberal response bias and make more false alarm errors when attempting to recognise other-race faces, Chapter 4 tested whether people are willing to shift their response bias or ‘criterion’ equally for own-race and other-race faces when presented with base rate information promoting a more liberal or more conservative response bias. We found that participants shifted criterion significantly in line with base rate information, regardless of whether base rate information was valid (Experiment 1) or invalid (Experiment 2). Most importantly, participants shifted equally for own-race and other-race faces, suggesting that: (i) people do not adapt to their inferior sensitivity for other-race faces by making greater use of base rate information; and (ii) the provision of accurate base rate information could be a useful means of reducing false alarm errors for other-race faces. Chapter 5 discusses in more detail the theoretical and practical significance of these findings.