VR Avatar Gender Categorization Task

This virtual reality environment was developed to investigate how people categorize virtual avatars in terms of gender identity based on their external appearance and visual features. The paradigm presents participants with a set of 10 different avatars, each characterized by distinct combinations of outward features, allowing researchers to examine how variations in avatar appearance influence gender-related judgments and categorization processes.

The task follows a within-subject design in which each participant is exposed to all 10 avatars across repeated trials. After viewing each avatar, participants provide their evaluations directly inside the virtual environment using Likert-type response scales. This interaction system was specifically designed to avoid removing the headset or interrupting the experimental experience, thereby helping to preserve immersion and sense of presence throughout the entire task.

The avatars were created using MetaPerson, while participant interaction with the virtual interface is supported through Meta hand-tracking libraries. Responses can therefore be entered naturally within VR without relying on traditional external input devices.

The system also includes an integrated data-management procedure. At the beginning of the session, a unique participant ID can be entered and associated with all subsequent responses. Information about the experimental session and the in-VR questionnaire is automatically saved as a CSV file directly on the device storage. This makes it possible to match VR-generated data with responses collected through external survey platforms, such as Qualtrics, by using the same unique participant identifier.

The environment is currently available in Italian and is designed as a fully guided experimental experience, with participants led through the procedure without requiring intervention from the researcher during the task.

The paradigm is currently being used as a research tool within an ongoing study on gender-related social categorization in immersive virtual environments. It should therefore be considered an active research prototype whose structure, stimuli, and validation are still being refined as data collection progresses.