Asymmetric Normalization in Social Virtual Reality Studies

Abstract

We introduce the concept of asymmetric normalization, which refers to decoupling sensory self-perception from the perception of others in a shared virtual environment to present each user with a normalized version of the other users. This concept can be ap-plied to various avatar-related elements such as appearance, location, or non-verbal communication. For example, each participant in a polyadic virtual reality study can see other participants at an average height of the respective test population, while individual participants continue to see themselves embodied according to their actual height. We demonstrate in a pilot experiment how asymmetric normalization enables the acquisition of new information about social interactions and promises to reduce bias to promote replicability and external validity.