Abstract
People can extract some kinds of statistical information from multiple items and faces. Previous study suggests that high-level ensemble representations (e.g., average facial expression) is independent from low-level ensemble representations (e.g., average color). This study investigates that how precisely people can perceive average of facial expressions. Participants were presented 12 faces expressing happy/angry and neutral, and asked to judge which expression face was majority. The results indicated that although accuracy was higher when many faces expressed the same expression, the evidence that participants perceived a precise average facial expression was not found. In contrast, participants can perceive a precise average color when color circles were presented. The current study suggests that ensemble representation of facial expression is not precise rather than that of simple feature when many faces were presented in short duration.
