31 subjects classified 16 schematic faces on six dimensions. Attractiv
eness was generally associated with widely separated eyes, a short nos
e, and high feature placement. When features were considered in combin
ation, faces with low placement were judged attractive, although the e
ffect was mitigated by a narrow mouth. Previous findings concerning pe
rceived age and gender were partially replicated, but folk wisdom fink
ing intelligence to low placement (high forehead) was refuted. It is s
uggested that the relationship between individual features and attribu
tions is nonlinear and interactive, and that researchers should invest
igate these hypotheses by exercising finer control over feature variat
ions.