Sheida Mirzaei Domabi¹, Ellie Leahey, Jason Haberman; ¹Rhodes college, ²University of Alabama, Birmingham

Recent studies show that the perceived emotion of a face may be influenced by its perceived gender — female faces are often seen as sadder than they actually are, while male faces are viewed as angrier. However, a major concern in how these studies are designed is that the stimuli (male and female faces) come from different identities, which introduces a confound: Are the differences in emotional perception driven by differences in gender, or by other unidentified differences between the stimuli. We address this concern by using physically identical, androgynous faces along with a rigorous psychophysical approach. We biased perceived gender by introducing non-face cues to androgynous faces (e.g., long hair). We generated morph sequences of androgynous faces (perceptually in-between male and female) from sad to angry, then feminized or masculinized them by modifying their hairstyles. Critically, the facial content was identical in both sets. In a 2x2 design, observers (N=46) viewed either an individual face or a set of four faces composed of either ‘female’ or ‘male’ morph(s), followed by a single, neutral test face (no hair). The mean expression of the sets was chosen randomly on every trial, and sets comprised faces varying in emotion. Observers adjusted the emotional intensity of the test face to match the perceived emotion of the previous set. There was a significant bias to view the ‘male’ faces as angrier and the ‘female’ faces as sadder, consistent with the stereotyped view of how perceived gender influences perceived emotionality. This, in spite of facial content being identical. There was no effect of set type (individual or ensemble). Although some of our previous work did not show an effect of perceived gender, the method-of-adjustment employed here provides a more sensitive measure to detect differences.