Appearance and Task Success in Novel Avatars

Novel avatar bodies are ones that are not controlled in a one-to-one relationship between the user’s body and the avatar body; for example, when the avatar’s arms are controlled by the user’s legs, or, when the avatar has a third arm. People have been shown to complete tasks more successfully when controlling novel avatar bodies than when controlling avatars that conform to the normal human configurations, when those novel avatars are better suited to the task (Won, Bailenson, Lee, & Lanier, 2015). However, the novel avatars in such studies tend to follow two conventions. First, the novel avatars still resemble biological forms, and second, the novel extensions of the avatar are connected to the avatar body. In the following study, participants operated bodies with three arms. We examined the interaction between biological appearance of the third arm and whether or not it was attached to the body. There was a significant effect of biological appearance on performance, such that participants inhabiting an avatar with a biological appearance did worse overall. There was also an interaction with biological appearance and an extension that appeared detached from the participant’s body such that participants in this condition performed most poorly overall. We propose a relationship between self-reported presence and task success, and discuss the implications of these findings for the design, implementation, and use of novel avatars. When users inhabit an avatar of a different race, gender, or age than their own, they may change their attitudes and behavior (Yee & Bailenson, 2007). But what happens when users inhabit novel avatars, or avatars that diverge from the normal human template (Lanier, 2006)? The ability of humans to adapt to novel avatar bodies is important both as it informs a theoretical understanding of body schema and as it applies to the practical implications of avatar use. For example, changing the relationship between an avatar’s movements and the movements of the user may be used in the treatment of pain. However, since mimicking the user’s actual appearance is not possible in avatars that deviate from the normal human template, the question of how such novel bodies should appear remains open. The effects of avatar appearance in other settings, especially virtual environments designed for therapeutic purposes, are beginning to be explored. This includes the effects on reported pain of changing the color of an avatar’s hand (Martini, Perez-Marcos, & SanchezVives, 2013), and how the way an avatar’s movements are rendered may increase or decrease an injured patient’s willingness to move (Chen, Ponto, Sesto, & Radwin, 2014; Harvie et al., 2014; Won et al., 2015). Recent research in virtual reality has demonstrated that people can adapt to novel avatar bodies and develop a sense of body ownership for these bodies. In a 2012 study by Kilteni, Normand, Sanchez-Vives, and Slater, participants inhabited avatars with arms of varying length, and reacted to threats to even very long arms with increased arousal. In a second study Steptoe, Steed, and Slater (2013) allowed participants to inhabit an avatar with a long tail. Participants whose avatars’ tail movements were synchronous with their own body movements reacted more strongly to threats to the tail than participants who inhabited avatars where the tail movements