DIFFERENCES IN FOREIGN VOCABULARY LEARNING OUTCOMES BETWEEN VIRTUAL ENVIRONMENT IMMERSION-BASED, TEXT-BASED, AND PICTURE-BASED LEARNING
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The main focus of the present study was to compare foreign vocabulary learning outcomes between immersion-based, text-based, and picture-based training within a virtual environment. Researchers have yet to quantitatively compare outcomes of foreign vocabulary learning between students who use virtual environments as immersive tools and more traditional text-based and picture-based training methods. The present study explored differences across time between groups assigned to one of these training methods on quizzes testing generalization of foreign vocabulary to real-world pictures. A 3D virtual foreign vocabulary learning environment created by ESLI, a language learning company, was utilized. All three groups learned material within the game in order to minimize computer-based group differences, but only the immersion group was able to explore the world and see the physical objects, while the other two were limited to learning Spanish phonology from either 1) English text translations or 2) picture presentations, both within a classroom area. Each group completed all 3 sections of material on 4 separate occasions, and took quizzes on vocabulary knowledge and generalization after every section. Analyses were conducted on a final sample of 32 participants. T-tests revealed no differences between individual times, and immersion-based learning to have lower overall accuracy than either text-based or picture-based learning, with no significant differences between text-based or picture based learning. A 3x4 mixed-measures ANCOVA was conducted, comparing participants of different foreign vocabulary learning methods (either immersion, text, or picture) on quiz accuracy over a time period of 4 learning and testing sessions, while controlling for prior foreign (Spanish) vocabulary knowledge. While Spanish vocabulary and time each significantly predicted accuracy (p’s < 0.05), neither the main effect of condition nor the interaction of condition and time significantly predicted accuracy (p’s > 0.1). Implications and potential future directions are explored.