학술논문

How Immersion and Self-Avatars in VR Affect Learning Programming and Computational Thinking in Middle School Education
Document Type
Periodical
Source
IEEE Transactions on Visualization and Computer Graphics IEEE Trans. Visual. Comput. Graphics Visualization and Computer Graphics, IEEE Transactions on. 29(8):3698-3713 Aug, 2023
Subject
Computing and Processing
Bioengineering
Signal Processing and Analysis
Programming profession
Cognition
Training
Task analysis
STEM
Solid modeling
Virtual environments
Computer science education
embodied cognition
immersion
self-avatars
VR in middle school education
virtual reality
Language
ISSN
1077-2626
1941-0506
2160-9306
Abstract
We present an empirical evaluation of immersion and self-avatars as compared to desktop viewing in Virtual Reality (VR) for learning computer programming and computational thinking in middle school education using an educational VR simulation. Students were asked to programmatically choreograph dance performances for virtual characters within an educational desktop application we built earlier called Virtual Environment Interactions (VEnvI). As part of a middle school science class, 90 students from the 6th and 7th grades participated in our study. All students first visually programmed dance choreography for a virtual character they created in VEnvI on a laptop. Then, they viewed and interacted with the resulting dance performance in a between-subjects design in one of the three conditions. We compared and contrasted the benefits of embodied immersive virtual reality (EVR) viewing utilizing a head-mounted display with a body-scaled and gender-matched self-avatar, immersive virtual reality only (IVR) viewing, and desktop VR (NVR) viewing with VEnvI on pedagogical outcomes, programming performance, presence, and attitudes towards STEM and computational thinking. Results from a cognition questionnaire showed that, in the learning dimensions of Knowledge and Understanding (Bloom's taxonomy) as well as Multistructural (SOLO taxonomy), participants in EVR and IVR scored significantly higher than NVR. Also, participants in EVR scored significantly higher than IVR. We also discovered similar results in objective programming performance and presence scores in VEnvI. Furthermore, students’ attitudes towards computer science, programming confidence, and impressions significantly improved to be the highest in EVR and then IVR as compared to NVR condition. Our work suggests that educators and developers of educational VR simulations, who want to enhance knowledge and understanding as well as simultaneous acquisition of multiple abstract concepts, can do so by employing immersion and self-avatars in VR learning experiences.