Full version
Morphing of faces
Average faces
Reformed facial shapes
Characteristics of beautiful faces
Beautiful figure
Virtual attractiveness
Virtual Miss Germany
Social perception
Press review
Making of
Contact us


Das durchschnittliche Frauengesicht-Das durchschnittliche Männergesicht
Left: averaged female face, made of 64 female faces; right: averaged male face, made of 32 male faces.

What feature makes a face attractive? Is it the average proportions of that face? We examined this popular theory ("averageness hypothesis") by conducting an experiment: We asked subjects to rate the attractiveness of all original faces and composites, which were calculated from 2, 4, 8, 16, 32, or 64 original faces. The results are: The attractiveness ratings of the transformed faces depend on the number of original faces that have been used to create them. The more original images were used to create the composite, the more attractive it was rated.  (r = 0.57 ** for female faces, r = 0.64 ** for male faces). On the one hand this result tends to support the averageness hypothesis put forward by Langlois & Roggman (1990), on the other hand we could clearly show: The ratings of the morphed faces also depend on the attractiveness of the underlying original faces (r = 0.75 ** for female faces, r = 0.68 ** for male faces). 

This means: Average faces are attractive, but not all of them. It is crucial which faces are used to compute an average face. Composites made from unattractive faces remain rather unattractive and average faces from attractive faces remain attractive. This clearly contradicts the averageness hypothesis, which holds that merely the number of faces contained in a composite alone accounts for the attractiveness ratings. 

Surprisingly, especially male faces benefit from being blended together with respect to their attractiveness. This does not support older findings that found positive effects only for women. Bad image quality of the faces, especially blurred contours, may be the cause for it. 

Why are the resulting average faces generally beautiful? One reason might be the fact that by calculating average proportions unpleasant asymmetries and irregularities become levelled out. Moreover, by blending together several faces wrinkles and pimples gradually disappear. As a consequence, the skin looks younger and perfectly smooth. 

By conducting another experiment we could show that it is these skin-smoothing side effects rather than the averaged proportions that account for the increased attractiveness ratings of the composites (see reformed facial proportions). 



Last modified: 01-07-2002, Webmaster
german version