Individual differences in estimating the gender composition of groups of famous faces

Abstract

In this experiment, the performance of male and female participants as they were instructed to estimate the gender composition of arrays containing 16 simultaneously presented faces was studied. The proportion of male to female faces present in the array was systematically varied. Explicit enumeration was prevented by continuously shuffling the position of the faces in the array using a 4 x 4 moving grid. In addition, any effect that familiarity might have on the decisions was looked at. Cumulative normal distributions were fitted to individual response data in order to measure sensitivity and the presence of any bias. Results showed that group estimates were made rather accurately across all participants, which is consistent with recent research on “ensemble” face processing. However, in contrast to recent literature, participant gender did not seem to have an effect on the decisions made. The main finding that resulted from this research is that familiarity did have an effect on the estimates. Specifically, when presented with famous faces participants weighted the females more heavily whereas when presented with unfamiliar faces, the males dominated.