Abstract
Recent work has revealed that the perceived expression of a face may depend on the perceived gender of that face — female faces are typically reported as sadder than male faces, and male faces are reported as angrier than female faces. An inherent confound with these studies, however, is that the faces being evaluated (male vs. female) are not equivalent and may contain important differences unrelated to gender. In the current experiments, we employ a psychophysical approach to examine whether the perceived gender of equivalent faces impacts the perceived intensity of a given emotion. To equate facial content, we created morph sequences of androgynous faces ranging from sad to angry. We then created two sets of feminized and masculinized faces by overlaying the hairstyles of the original female and male faces used to generate the androgynous face. Importantly, the facial content was identical in both sets. In a 2x2 design, observers (N=68) viewed either an individual face or a set of four faces composed of either ‘female’ or ‘male’ morphs, followed by a single test face (without any hair). Sets varied in emotional expression, which was chosen randomly on every trial. Observers were asked to identify whether the test face was sadder or angrier than the previous individual/set of faces. In contrast to prior research and our own hypothesis, observers had a small but significant bias to view the ‘male’ faces as sadder than the ‘female’ faces. There was no effect of set type (individual or ensemble). This research in no way challenges the notion that emotional stereotypes exist, only that at a perceptual level when facial content is equated, biases may manifest in unexpected and independent ways.