Spectacles, distinctiveness, and face recognition: a web-based experiment

This study sought to clarify previous research in the face recognition literature regarding memory for faces with spectacles. A second aim of this research was to further investigate Valentine's face-space model, a leading model of face recognition that predicts better performance on distinctive faces compared to typical faces. Prior to this experiment, independent observers provided distinctiveness ratings for faces with, and without, spectacles. Experimental participants then accessed the PsychExperiments website and completed a face recognition experiment. Based on the judgments of the independent observers, the face-space model predicts that memory for spectacled faces should be superior to memory for non-spectacled faces. An analysis of hit rate (percent correct) supported this notion, as a higher hit rate was observed for spectacled faces compared to non-spectacled faces. However, the analysis of false alarms (false identifications) did not support the predictions made by the face-space model, as participants demonstrated reliably higher false alarm rates for faces with spectacles. A further analysis of response bias suggests that the overall pattern of responding may have been largely due to changes in response criteria for trials where spectacled faces were presented. Implications for models of face recognition are discussed.