Components of Facial Attractiveness

Abstract

Terry and Brady (1 ) reported that the relative contribution of a facial component to female undergraduates' self-ratings of their over-all facial attractiveness varied with the type of visual corrective worn. Averaging across this variable, the oral region was the most influential component, followed in order by the eyes, hair, nose, and facial structure (shape and complexion). However, since this order was generated by selfratings, a judge's ratings of the attractiveness of her several components were likely to have been contaminated by each other. A better estimate of the relative influence of a component in over-all facial attractiveness would be afforded from judges' ratings of isolated components of other persons' faces. Subjects, 25 male and 25 female Caucasian undergraduates, were shown two-color, 4X G-in. portrait photographs of 5 males and 5 females of che same race and approximate age as the subjects. One photograph of each pair was removed from its background and glued to a neutral grey file card. The other pair member was dissected into the separate facial components, viz . , mouth, eyes, hair, nose, and facial structure, which were pasted onto similar file cards and coded on the reverse side to permit later matching. The deck of 60 stimuli was presented to the subjects in a shuffled order and subjects checked an appropriate level of a 10-point scale on which 1 represented the lowest level of attractiveness and represented the highest level. Each subject's 10 over-all ratings of attractiveness were correlated with his ratings of the respective components. The higher a coefficient, the more important a component was assumed to be in determining a given subject's rating of over-all attractiveness; hence, the sign or direction of a relationship was ignored. The absolute values of these correlations were then transformed to zs, and an analysis of variance, sex of subject (rnale/female) X component (mouth/eyes/hair/ nose/facial structure), was computed. The F ratios for the sex of subject and the interaction were less than 1.00, whereas the components were significant (F = 5.53, df = 4/192, p < .01). Transforming the average zs back to ss resulted in the following order of importance for the components: mouth ( r = .53), eyes ( r = .44), structure ( r = .43), hair ( r = . 3 4 ) , and nose (s = .31). Except for facial structure, this is precisely the order obtained by averaging Terry and Brady's (1) results. However, facial structure in that study explicitly referred to complexion and shape, whereas here structure was all that remained of the face when the other components were removed. Therefore, this component was not treated comparably in the two studies. This study replicates and extends Terry and Brady's by demonstrating that their order can be generated by perceptions of the isolated components of other persons' faces as well as by self-perceptions.