| The role of similarity and typicality in mediating the effect of attractiveness on face recognition was explored. In Experiment I the interrelationships between perceived frequency, typicality, similarity, and attractiveness was investigated. Attractive female faces were perceived to be more common, typical, and similar than average or unattractive female faces. Attractive faces were found to be narrower and have longer, darker hair than other faces. In Experiment II the role of context specific similarity as produced by the prevalence of attractive, average, and unattractive faces was tested. Results indicated that recognition accuracy is best for unattractive faces regardless of their prevalence. These faces were also recognized with most confidence. Attractive faces received less false alarms than average faces. In Experiment III recognition of attractive, average, and unattractive female faces was tested by pairing faces at each level of attractiveness with faces of only one other level of attractiveness. This procedure afforded the opportunity to test the recognizability of each type of face relative to each other. Results indicated that all faces were recognized more accurately, and received less false alarms when they were tested against attractive faces.; Taken together, the results of these studies indicate that unattractive and attractive faces are recognized better than average faces. Unattractive faces seem to be recognized well due to their dissimilarity while attractive faces may be recognized well because they command greater attention than other faces. Results of these studies are discussed in terms of Gibson's perceptual learning theory. |
