Increasingly popular social virtual reality (VR) platforms like VRChat created new ways for people to interact with each other, generating dedicated user communities with unique idioms of socializing in an alternative world. In VRChat, users frequently gather in front of mirrors en masse during online interactions. Understanding how user communities deal with the mirror’s unique interactions can generate insights for supporting communication in social VR. In this study, we investigated the mirror's synergistic efect with avatars on behaviors and dedicated user conversational performance.
In social VR, people appear as avatars, expressing their true identity. This study aims to explore the differences
between mirrors in social VR and real life, examining adaptations made by expert users. While prior research has
studied real-life mirrors and the Proteus effect in avatars, none have investigated the interactions occurring in
front of social VR mirrors.
So the research gaps motivate us to investigate the following two research questions:
RQ1: How do users behave and perceive themselves in the context of mirrors in social VR?
RQ2: How do mirrors influence user conversational performance in social VR?
We conducted in-depth interviews to explore user perceptions of mirrors in social VR, addressing RQ1. As communication mainly occurs through conversation, we investigated RQ2 by examining the influence of mirrors on interpersonal communication. To answer RQ2, we conducted a controlled experiment with conversational tasks (speaking performance) in mirror and non-mirror settings using a VR headset. Follow-up interviews were conducted, and performance was evaluated using the Public Speaking Competence Rubric scored by external scorers.
Mixed method experiment workflow
Overhead view of the pre-designed experimental and mirror on/off from participants’ view
Answering RQ1 using Grounded Theory, we identified key themes in social VR mirror gatherings. Users follow an unwritten social rule, sitting in circles for mutual support and enjoying instant feedback from mirrors. Mirrors provide contextual information, wider views, and help users ensure accurate avatar synchronization and appearance. Mirrors enhance anonymity, encouraging confident expressions and unique social behaviors, such as the phantom sense and transgressive acts. Participants also expressed their obsession with ideal embodiment and self-awareness in the virtual world through mirrors.
In answering RQ2, experiment results indicate that participants perform worse in two areas, including
'listener's needs awareness,' but feel more confident with mirrors on. Mirrors boost confidence and reduce
anxiety by reflecting body language without direct gaze, yet they decrease conversational performance,
possibly due to avatar distraction.
Based on previous literature and our research results, we find mirrors in social VR, similar to self-view in
teleconferencing, increase cognitive load and reduce performance due to avatar synchronization and appearance
concerns. Both real-life and social VR mirrors enhance self-awareness, but social VR mirrors also affect embodiment,
body-ownership, and the Proteus effect, influencing behavior. In conclusion, VR doesn't always parallel physical
reality, but people adjust and develop new metaphors for objects like mirrors in immersive settings.
MPSCR score of conversational task
Self-perception level of conversational task
VRChat offers unique interactions, influencing social behavior and perception through avatars. Mirrors in social VR emphasize ideal body illusions, enhance self-awareness, encourage bold behaviors, provide consistency feedback, and create a fun interaction culture. Our study highlights communication and behavioral patterns in social VR, positive self-perception impact, and negative verbal performance effects. We explain this discrepancy using mirror theory and avatar distraction and explore unique strategies and adaptations in virtual versus physical spaces.
Fu, K., Chen, Y., Cao, J., Tong, X., & LC, Ray. (2023).
CHI '23: Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems April 2023. No.: 385, 1–19.