Enhance Learner Engagement ThroughExperiential Learning in a Gamified Simulation:A Longitudinal Study
Chen Li, Jeff K. T., Ye Jia*, Yufei Lu, Peter H. F. Ng, Laura Zhou, Jing Liu, Qing Li
18th International Conference on Blended Learning (2025)
Gamified simulation of Fitts’ Law in VR. (a) Overview of the simulation in VR; (b) UI panel displaying feedback based on Fitts’ Law; (c) Enhancement tools aiming to inspire improved interactions; (d) Third-person view of a participant using the simulation.
Abstract
Experiential learning in virtual reality (VR) has gained con-siderable attention from both practitioners and researchers in recentyears. Despite its potential, there is still a gap in effectively engagingstudents through experiential learning. Gamification has emerged as apromising strategy to bridge this gap by incorporating interactive andmotivational elements into learning experiences. To explore the poten-tial of combining gamification with experiential learning in VR, we con-ducted a longitudinal study with 68 undergraduate students enrolled ina human-computer interaction (HCI) course. The study focused on agamified simulation of Fitts’ Law, a fundamental concept for modellinghuman movement in HCI, to evaluate its impact on learner engagementand the translation of this engagement into knowledge gains over boththe short and long term. Our findings indicated that the gamified ex-periential learning approach significantly enhanced learner engagement,especially affective engagement, compared to traditional methods thatrely on texts and figures. The longitudinal data also suggested that theapproachledtobetterknowledgeretention.However,norelationshipwasfound between the observed learner engagement and knowledge change.This highlights the need for further research to study how to effectivelyconvert learner engagement into measurable knowledge gains throughexperiential learning in gamified simulations.
The Engagement Assumption in Edu-VR
A quiet assumption runs through much of the edu-metaverse literature: that engagement mediates the relationship between immersive technology and learning. More immersion → more engagement → more learning. It is an intuitive chain of reasoning, and it motivates a great deal of investment in VR for education. But it is rarely tested directly. This paper tests it — and finds that the second link in the chain (engagement → learning) does not hold in the way the field has assumed.
What the Study Did
Sixty-eight undergraduate students in a Human-Computer Interaction course participated in a longitudinal study comparing two ways of learning Fitts' Law — a fundamental HCI principle that models the relationship between target size, target distance, and movement time. The control condition used traditional text-and-figure materials. The experimental condition used a gamified VR simulation where students physically performed pointing tasks under varying parameters (different target widths, different distances) and received real-time feedback visualized through Fitts' Law calculations. The simulation included game elements: progressive difficulty levels, performance scores, and visual enhancement tools that encouraged experimentation ("what happens if I make the target smaller?").
The longitudinal design — measuring both immediate engagement and longer-term knowledge retention — is a methodological strength. Most edu-VR studies are single-session; this one tracked knowledge over time, which matters because the practical goal of education is retention, not immediate post-test scores.
The Central Finding (and the Tension It Creates)
Two results sit in productive tension:
Result 1: Gamified VR significantly increased engagement. Specifically, affective engagement — students reported enjoying the learning experience more, feeling more invested, and being less likely to disengage. This is not surprising (VR is novel, games are fun), but the magnitude and reliability of the effect, with 68 participants across multiple sessions, gives it weight beyond the typical small-sample VR demo.
Result 2: Higher engagement did not predict higher knowledge gains. The correlation between engagement measures and knowledge change was essentially zero. Students who were more engaged did not learn more than students who were less engaged. The gamified VR group did show better long-term retention overall, but the mechanism was not engagement — or at least, not the kind of engagement the study measured.
This is the paper's most important contribution: a clean, empirically grounded challenge to the "engagement drives learning" assumption. It suggests that engagement and learning may be parallel outputs of good instructional design rather than causally linked, or that the engagement measures used (which focused on affective engagement — enjoyment, interest, motivation) miss the forms of engagement that actually matter for knowledge acquisition (cognitive engagement — active processing, elaboration, self-explanation).
How to Interpret the Null Result
The null relationship between engagement and learning should not be read as "engagement doesn't matter." It should be read as "engagement is not sufficient, and we need to understand what converts engagement into learning." The paper is appropriately cautious here, but the implication is clear: the field needs to shift from demonstrating that VR increases engagement (which is increasingly well-established) to understanding how to design VR experiences where engagement leads to learning. This may require embedding explicit instructional strategies — self-explanation prompts, retrieval practice, error-based learning — within gamified VR, rather than assuming that engagement alone will carry the cognitive load.
The longitudinal retention advantage for the VR group, despite the null engagement-learning correlation, suggests there may be unmeasured mediators at work. Possibilities include: embodied cognition (physically performing Fitts' Law tasks may create stronger motor-memory traces than reading about them), variability of practice (the simulation exposes students to a wider range of parameter combinations than static examples), or test-enhanced learning (the simulation's feedback loop functions as a form of continuous self-testing). The study design doesn't isolate these, but identifying them is the natural next step.
Boundaries
Fitts' Law is a single concept — a well-defined, mathematically tractable principle. Whether these findings generalize to more complex, less quantitative topics (e.g., learning about design principles, ethical frameworks, or historical context) is unknown. The gamification elements were relatively simple (scores, levels, visual enhancements); the interaction between more sophisticated game mechanics and the engagement-learning relationship is unexplored.
The sample, while larger than many VR studies, is still a single course at a single university in a specific cultural context. The longitudinal component tracked retention within a single semester, not across semesters or years. And critically, the study measures the engagement-learning relationship at the between-subjects level; within-subjects analyses (do the same students show a correlation between their own engagement and their own learning across different topics?) might reveal patterns that the group-level analysis obscures.