Your Avatar Seems Hesitant to Share About Yourself: How People Perceive Others' Avatars in the Transparent System

Title of the Paper

Your Avatar Seems Hesitant to Share About Yourself: How People Perceive Others’ Avatars in the Transparent System

Paper Information

• Research Area: Human-Computer Interaction, Virtual Avatar Design, Social Virtual Reality
• Keywords: Avatar-mediated communication, transparent system, self-disclosure, social perception, customization, personalization, anonymity, interaction design, virtual reality, metaverse

Research Background and Problem

• Problems and Challenges:

  • In social virtual environments, users often find it difficult to discern how others' avatars are created, even though such information is crucial for forming first impressions and evaluating others.
  • Although personalized and highly realistic avatars are believed to enhance interaction experiences, little research has explored how "transparent systems" (i.e., systems that explicitly display how avatars are created) influence perceptions of users’ self-disclosure.
  • Bridging the cognitive gap between senders (avatar creators) and receivers (observers) regarding self-disclosure in virtual social environments remains an unaddressed research gap.

• Significance of the Research:

  • In virtual social platforms like the metaverse, avatar design and the transparency of related information could significantly impact user experiences.
  • Investigating these issues can help developers understand the pros and cons of transparency in design and optimize user experiences.

• Motivation and Existing Work:

  • Current research on AMC (Avatar-mediated communication) primarily focuses on how avatars affect their users (e.g., the Proteus Effect), with limited exploration of how others evaluate users and their behavior through avatars.
  • Existing theories (e.g., hyperpersonal model, theory of mind) suggest that receivers’ overreliance on limited cues (e.g., avatars) may lead to inaccurate judgments.
  • This study aims to fill this gap by examining how transparent systems alter receivers’ perceptions of senders and their self-disclosure.

Solution

• Research Methodology:

  • Designed and conducted a 2×2 mixed-design experiment.
    • Independent Variable 1: System design (transparent vs. non-transparent)
    • Independent Variable 2: Avatar creation method (personalized vs. customized)
  • The experiment involved 60 participants watching pre-recorded self-introduction videos and completing questionnaires about the senders.
  • The videos and self-introductions clearly marked the avatar creation methods with icons (e.g., camera icon = personalized, hand icon = customized).

• Innovations:

  • First integration of personalized avatars and transparent system design in AMC research to explore how these factors jointly influence social behavior and self-disclosure.
  • Systematically quantified receivers’ evaluations of senders across multiple dimensions, including anonymity, trustworthiness, and social attractiveness.
  • Provided practical design recommendations for transparency in avatar design on metaverse platforms.

• Research Steps and Key Techniques:

  • Data Collection: Users created and showcased different types of avatars on a virtual platform (Spatial.io).
  • Experimental Validation: Tested receivers’ evaluations of senders across dimensions such as anonymity, trustworthiness, social attractiveness, and self-disclosure in both transparent and non-transparent systems.
  • Data Analysis: Employed mixed-design ANOVA and generalized linear models for statistical analysis of experimental data.

Research Findings

• Specific Findings:

  • In transparent systems, receivers rated senders using personalized avatars lower:
    • Perceived these senders as less broad, comfortable, and honest in self-disclosure.
    • Rated their competence lower as well.
  • Senders in non-transparent systems attracted more attention from receivers (e.g., willingness to follow or engage with them).
  • Senders using customized avatars were perceived as more comfortable and trustworthy compared to those using personalized avatars.

• Experimental Results and Advantages:

  • The experiment confirmed that transparent system designs can lead to unfair judgments of avatars by receivers, especially based on surface-level information from personalized avatars.
  • In non-transparent systems, receivers’ evaluations of senders were closer to reality and more likely to foster future interactions.
  • The study highlighted the "unknown attractiveness" effect that may arise from concealing avatar creation information.

• Limitations and Future Directions:

  • Limitations:
    • Potential differences in video content could influence the interpretation of results.
    • The experimental setting involved one-way interactions (watching recorded videos), which may not fully reflect the dynamics of real-world two-way social interactions.
    • Lack of exploration into interaction patterns in pre-existing relationships (e.g., acquaintances).
  • Future Directions:
    • Further testing of these findings in real-world, two-way interactive metaverse scenarios.
    • Exploring optimal design solutions to balance transparency and user autonomy.
    • Considering the influence of variables such as users’ familiarity with AMC platforms, gender, and age on research outcomes.

Practical Recommendations

• Transparent systems should be designed cautiously to avoid unfair evaluations of users due to transparency.
• Provide users with the autonomy to choose whether to disclose avatar creation methods, while increasing awareness of the impact of transparency.
• When developing metaverse platforms, prioritize supporting diverse avatar creation methods (including personalization and customization) to meet varying needs for trust and anonymity in different scenarios.