Gaze and task performance in shared virtual environments

Abstract
Non‐verbal behaviour, particularly gaze direction, plays a crucial function in regulating conversations and providing critical social information. In the current set of studies, we represented interactants in a shared immersive virtual environment. Interactants sat in physically remote rooms, entered a common virtual room and played games of 20 questions. The interactants were represented by one of three types of avatars: (1) human forms with head movements rendered in real time; (2) human forms without head movements rendered; or (3) human voice only (i.e., a conference call). The data demonstrated that interactants in the rendered head movement condition rated a higher level of co‐presence, liked each other more, looked at each other's heads more, and spoke for a lower percentage of time during the game, compared to the other two conditions. We discuss implications for the design of shared virtual environments, the study of non‐verbal behaviour and the goal of facilitating efficient task performance. Copyright © 2002 John Wiley & Sons, Ltd.