Virtual Reality Serious Games and Language Learning: A Quasi-Experimental Study in an Ivorian School Setting

This quasi-experimental study examines whether a virtual reality serious game enhanced with adaptive artificial intelligence can improve foreign language performance among secondary school pupils in Abidjan, Côte d'Ivoire. The sample consisted of 68 pupils aged roughly 15 to 17 at the Collège Moderne Mami Faitai, who completed a roughly three-month intervention using the Mondly Languages VR application between February and April 2025. Outcomes were assessed through a pre-test and post-test language instrument, self-assessment questionnaires, and a follow-up transfer measure, structured around Kirkpatrick's four-level evaluation model. The absence of a control group and the reliance on a single convenience sample are acknowledged as significant limitations.

The principal quantitative result is a rise in median scores from 52% to 67%, a statistically significant gain (Wilcoxon W = 189, p < 0.001) with a Cohen's d of 0.75. Learner satisfaction was high, averaging 4.45 out of 5, with the enjoyment of gameplay and the usefulness of AI feedback rated most positively. The only satisfaction item falling below the 4/5 threshold concerned the appropriateness of difficulty levels, suggesting that the adaptive algorithm has room for refinement. Self-assessed gains were notably stronger for oral comprehension and speaking than for written expression, which received the lowest ratings.

The study also identifies a clear gap between engagement and behavioural transfer. Students reported strong motivation to continue using the game, yet their frequency of independent use and application of the language in everyday contexts remained modest, with most transfer items scoring below the neutral midpoint. The authors interpret this as evidence that engagement may depend largely on the institutional setting rather than being intrinsically sustained, and they call for longitudinal research with an active control group and objective transfer measures.