How Should Learners be Socially Embedded in Virtual Learning Simulations? Effects on Problem-Solving Performance, Perceived Authenticity, Enjoyment, and Social Presence
摘要
Virtual learning simulations (VLS) are increasingly employed in education for their potential to replicate real-world environments and support authentic, action-oriented learning. Given that learning is fundamentally socially constructed, an important question concerns how varying degrees of social embedding within VLS, particularly through the inclusion of artificial players, affect learning outcomes. This study investigates how varying degrees of social embedding within VLS affect problem-solving performance and learners’ perceptions of authenticity, enjoyment, and social presence. A problem-solving task was adapted into four variants, each representing a distinct level of social embedding, based on a predefined taxonomy. In a randomized controlled experiment, 296 postsecondary learners from academic and vocational education tracks were randomly assigned to one of the four variants. To examine the effects of social embedding, a one-way MANOVA was conducted, followed by ANOVAs with planned contrasts. In addition, learners evaluated all four variants based on their preferences. Higher levels of social embedding significantly increased perceived authenticity, enjoyment, and social presence, whereas no significant effects were observed for problem-solving performance. Notably, while learners viewed social embedding as essential for enhancing authenticity, enjoyment, and social presence, a subset preferred individual work in socially enriched yet non-collaborative environments. The findings emphasize the importance of designing socially enriched VLS to enhance learner experience and engagement. While problem-solving performance may remain unaffected, the results highlight the pedagogical value of social embedding for fostering affective learning outcomes. These insights offer practical implications for educators aiming to design more engaging simulation-based environments.