Abstract

Objectives: To explore whether the own-race effect is sensitive to viewpoint and number of faces to remember (set size). It was hypothesised that recognition for own-race faces will be (i) higher than other race faces (ii) higher in upright than profile view and (iii) decreased as set size increased.
Design: A 2 (viewpoint: frontal, half-profile) x 3 (set size: 2, 4, 6 faces) x 3 (race: Caucasian, Black and Hispanic) repeated measure design. The dependent variables were: (i) d-prime and (ii) criterion
Methods: A total of 40 Caucasian participants completed a change detection task designed to characterise the relationship between N (set size) and d’ (a signal detection theory measure) for own and other race faces presented in frontal and profile viewpoints. The proportion of hits and false alarms were converted into d-prime (d’) scores in order to provide a more sensitive measure of signal detection. A separate 2 (viewpoint: frontal, profile) x 3 (race: Black, Caucasian, Hispanic) x 3 (set size: 2, 4, 6) repeated measures ANOVA was conducted with sensitivity d’ and criterion (C) as the repeated measures.
Results: Sensitivity decreased as set size increased from 2 to 6 faces and higher in profile than frontal viewpoint. Sensitivity was stronger for own than other-race faces independent of facial viewpoint. However, sensitivity for own-race faces decreased at larger set size (≥4 faces).
Conclusions: This study has revealed that own race advantage can survive changes in facial viewpoint. However, there is capacity limits to representing own-race faces in visual working memory. Implications for eyewitness facial identification is discussed.