A Robust Method of Measuring Other-Race and Other-Ethnicity Effects: The Cambridge Face Memory Test Format

Elinor Mckone,Sarah Cohan,Sacha Stokes,Mary Broughton,Jia Liu,Galit Yovel,Chiara Fiorentini,Madeleine Pidcock,Michel Pelleg,Manos Tsakiris

doi:10.1371/journal.pone.0047956

Abstract

Other-race and other-ethnicity effects on face memory have remained a topic of consistent research interest over several decades, across fields including face perception, social psychology, and forensic psychology (eyewitness testimony). Here we demonstrate that the Cambridge Face Memory Test format provides a robust method for measuring these effects. Testing the Cambridge Face Memory Test original version (CFMT-original; European-ancestry faces from Boston USA) and a new Cambridge Face Memory Test Chinese (CFMT-Chinese), with European and Asian observers, we report a race-of-face by race-of-observer interaction that was highly significant despite modest sample size and despite observers who had quite high exposure to the other race. We attribute this to high statistical power arising from the very high internal reliability of the tasks. This power also allows us to demonstrate a much smaller within-race other ethnicity effect, based on differences in European physiognomy between Boston faces/observers and Australian faces/observers (using the CFMT-Australian).

Highlights

In this article, we use the term race of a face to refer to the relatively large physical differences in faces with ancestry from different major world regions, such as Europe, Asia, or Africa
A two-way ANOVA revealed no main effect of Test Version, F(1,42) = 0.261, MSE = .006, p..6, indicating that the new Cambridge Face Memory Test (CFMT)-Chinese is well matched in overall difficulty to the CFMT-original
There was a just-significant effect of participant race, with Asians performing better overall than Europeans, F(1,42) = 4.246, MSE = .031, p = .046; the reason for this is unclear, but one speculation is that individuals willing and able to live in another country to study differ from local students on some variable relevant to face recognition

Summary

Introduction

We use the term race of a face to refer to the relatively large physical differences in faces with ancestry from different major world regions, such as Europe, Asia, or Africa. We use the term ethnicity of a face to refer to the smaller physical differences that exist within a race, such as with ancestry from Norway versus Greece within Europe, or China versus Japan within Asia, or Nigeria versus Ethiopia within Africa. The other-race effect, known as the own-race bias or other-race deficit, is well established (e.g., [1], for review). Individuals of another race are remembered more poorly that those of one’s own race, as demonstrated in a two-way interaction between race of observer (e.g., Asian, European), and race of face (Asian, European). Only one of these ethnicity studies tested the full crossover design (both observer groups crossed with both types of faces) and reported a significant two-way interaction (in Turkish/ German participants crossed with Turkish/German faces [4])

Objectives

Methods

Results

Conclusion