A model for measuring the contribution of form to perception of biological sex in point-light actors

People are able to judge the sex of actors represented in point-light displays. Much of the previous literature has focused on the contribution of motion information to this judgment, or has focused specifically on information gathered from the shoulders and hips. The purpose of the current study is to model how form information, based upon the distance between the dots representing a point-light actor’s key joints, the shoulder, hip, elbow, knee, wrist, and ankle points, contributes to the perception of biological sex. In a pilot study, 6 naïve observers responded to 63 computer generated stimuli that ranged from extremely female to extremely male (i.e., from -3 to +3 standard deviations for biological sex) and also varied on a set of form and motion-affecting conditions (i.e., the actors were represented as being happy, sad, calm, nervous, heavy, light, or neutral). The results of the pilot study suggested that observers were less accurate at judging the sex of point-light walkers that were expressing emotions. In order to control for this, emotion conditions and motion variation between stimuli were not included for the thesis study. The model in the thesis study was developed based upon how 8 naïve observers judged the sex (i.e., male or female) of 72 computer-generated clips displaying a walking pointlight actor. Nine (9) of these clips displayed an actor whose sex ranged from extremely female to extremely male and the remaining 63 actors based upon a sex neutral actor with different combinations of characteristics from the extremely female actor (i.e., elbow distance, hip distance, and ankle distance) and the extremely male actor (i.e., shoulder distance, wrist distance, and ankle distance). The final model developed in the thesis study suggests that judgments of sex from form cues in point-light displays rely most heavily on the distance between the target actor's ankles, the distance between target actor's elbows, and the target actor's shoulder-to-hip ratio. However, observers also utilized other form cues, suggesting that the entire process of perceiving biological sex depends on information that is distributed throughout the actor. A MODEL FOR MEASURING THE CONTRIBUTION OF FORM TO PERCEPTION OF BIOLOGICAL SEX IN POINT-LIGHT ACTORS A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree Master of Arts