Abstract
Studies of the perception of face gender often assume that androgynous faces lie near the midpoint of a continuum that includes “femininity” and “masculinity” on either extreme. For example, androgynous face stimuli are often generated by morphing male and female faces with equal weighting. However, the validity of using morphs to represent all androgynous faces has yet to be tested; the current study addresses that gap. Undergraduate students (n=185) used a 5-point scale to independently rate faces’ levels of femininity, masculinity, and androgyny. A subset of these faces were rated, on average, at least “somewhat androgynous” (i.e., mean rating >3). We classified these faces as natural (non-morphed) androgynous faces. We recruited an additional 25 subjects to rate the same set of non-androgynous and androgynous natural faces, along with a new set of morphed androgynous faces that were created from faces previously rated as strongly feminine or strongly masculine (i.e., faces with mean femininity or masculinity ratings >4). We found that the two types of androgynous faces were given similar androgyny ratings and masculinity ratings, but were given significantly different femininity ratings. The natural androgynous faces were perceived as more feminine than masculine, whereas the morphed androgynous faces were perceived as more masculine than feminine. This result suggests that morphing male and female faces may create androgynous faces that differ in important ways from natural androgynous faces. Further, the fact that the two groups of androgynous faces had unequal masculinity and femininity ratings suggests that androgynous faces do not all fall at the midpoint of a masculinity-femininity continuum. We are currently investigating the stimulus characteristics that distinguish natural and morphed androgynous faces, and that are correlated with various gender judgments.