Effects of Avatar Appearance on User Perception and Behavior: Role of Labels and Cognitive Mediation in the Proteus Effect

Abstract

Previous studies suggest avatar appearance can influence user cognition and

behavior (i.e., the Proteus Effect), though the mechanism responsible is debated. This study examined whether the mechanisms proposed by two explanations of the Proteus Effect could be experimentally induced and subsequently measured via assessing user response time when rating their avatar’s traits. This study also explored whether in-game labels used to describe an avatar can bias users’ interpretations of their avatars in label- consistent ways, and potentially lead to similar changes in users’ offline behaviors. It was predicted that users generating their own evaluations of their avatars would do so more quickly than users asked to generate evaluations from the perspective of imagined others, and that users would rate their avatars in a manner stereotypically consistent with the avatar label. Participants were brought into a laboratory and played as an avatar in the desktop computer game The Sims before rating their avatar and squeezing a handgrip apparatus intended to measure potential changes in participants’ offline behavior. General support was found for the effect of label on avatar assessments; these labels biased users toward rating their avatars as having higher levels of 3 of 5 label-consistent traits. No support was found for any effect of label on users’ offline behaviors. Finally, results indicate response time may hold potential as a proxy measure for detecting the proposed mechanisms of the Proteus Effect. Implications of these results are discussed.