Face similarity and the other -race effect in face perception

• Main Author: Yamashita, Jill A
• Format: Dissertation
• Language: English
• Published: ProQuest Dissertations & Theses 01.01.2003
• Subjects: Cognitive psychology; Cognitive therapy; Experimental psychology; Experiments; Psychology
• ISBN: 0496472941; 9780496472949

The other-race-effect (ORE) is when a person has more difficulty recognizing or perceiving someone from a different race. The ORE presumably occurs because we have less frequent exposure to other-race faces, because we experience a non-uniform distribution of faces with most faces clustered around the distribution mean. This type of theoretical face space is similar to the distribution of colors in our environment, and my major goal is to investigate if the ORE shows parallels to properties of color-coding. Color studies show that the color system is very sensitive to colors around the average (white) and not as sensitive to colors at the extremes- or colors not often seen in the natural environment. The neural response follows an ideal sigmoidal curve that is matched to this distribution so that response changes are steepest where the color signals are most dense. This predicts that equal physical differences should appear less similar, better remembered, and more discriminable when the differences are near the mean and thus fall on steeper parts of the response curve. These experiments test for analogous difference in perceptual performance in face recognition by way of 3 experiments that compare responses to same and cross-race faces. Experiment 1 investigates perceptual similarity judgments, experiment 2 examines face recognition and experiment 3 looks at face discrimination. Experiment 1 showed that faces from one race were perceived as more similar compared to cross-race faces but did not reveal a difference between the perceived similarity of same and other-race faces. Experiment 2 showed that faces that appeared similar to the study faces were more likely to be falsely recognized but again did not reveal an ORE. Finally, experiment 3 showed that upright faces are more discriminable than inverted, but did not reveal discrimination differences for Japanese and Caucasian faces. All 3 experiments therefore failed to show a robust ORE. This could mean that faces are encoded in a way that is not analogous to color, or that the students tested were already exposed and thus acclimated to different-race faces.