Developing Speaking Fluency through Metaverse-Assisted Learning in Engineering Students

Abstract

Speaking fluency is one of the important communication skills that engineering students must master in the era of globalization and Industrial Revolution 5.0. However, engineering students still experience difficulties in speaking English fluently because of limited vocabulary, lack of speaking practice, low confidence, and limited use of interactive learning media. This study aimed to improve students’ speaking fluency through metaverse-assisted learning. The research employed Classroom Action Research (CAR) using the Stephen Kemmis and Robin McTaggart model consisting of planning, action, observation, and reflection stages. The study was conducted in two cycles involving first-semester students at Bumigora University. The data were collected through observation, interviews, speaking tests, and field notes. The speaking assessment focused on pronunciation, vocabulary usage, grammatical accuracy, and fluency. In Cycle I, students became more active and confident during speaking activities through virtual self-introduction, role-play, group discussion, and project presentation using Roblox avatars and YouTube short-story videos. However, some students experienced difficulties in pronunciation and vocabulary usage. Therefore, the researcher revised the learning activities in Cycle II by adding pronunciation practice, vocabulary scaffolding, simulations, and collaborative discussion activities. The results of Cycle II showed significant improvement in students’ speaking performance. Students communicated more fluently, used more varied vocabulary, and actively participated in classroom interaction. The average speaking score improved from 62.4 in the preliminary study to 84.2 in Cycle II, categorized as good. The findings also indicated that metaverse-assisted learning created a more interactive, collaborative, and enjoyable learning environment.