Abstract
The study explores how personalisation, learner engagement and accuracy influences knowledge retention in a gamified virtual reality education context. By drawing on the spacing effect and Gamification Theory, the research uses secondary data from the Duolingo Spaced Repetition dataset for analytical purposes. The analysis focuses on whether personalised repetition intervals and engagement metrics predict recall performance.
Machine learning and clustering methods have been key to group learners with similar traits and engagement patterns to better tailor gamification elements to specific groups. By identifying shared interactions traits, educators can deliver more personalised learning content that will better align to the different types of learners. Based on these approaches, education is moving towards a more precise learner engagement with better content as it is more aligned to their desired learning outcomes. This can lead to better learner engagement and motivation on gamification systems.
A quantitative approach is applied, analysing over 12 million records by using descriptive statistics, correlation analysis and multiple regression modelling. The analysis shows that recall probability is extremely high as the mean is roughly 0.9 and the median is 1, suggesting strong short-term retention but limited variation among learners. Among the variables that were tested, accuracy was the most significant predictor of recall while spacing intervals and general engagement metrics reflect minimal effects.
The research results suggest that the effectiveness of learning through the gamification theory relies more on the quality of learner engagement and accuracy-driven feedback rather than on automated personalisation. This study contributes to theory by identifying boundary conditions for the spacing effect in large-scale digital environments. It further provides practical recommendations for improving adaptive feedback and learner engagement design in online learning systems within a digital framework.