Analysing Game Data to Evaluate Children's Learning Progress with the EDURINO App

The EDURINO study examines gameplay log data from the EDURINO app to evaluate whether young children show measurable learning progress within educational minigames. Drawing on data from 66,463 players across 18 Literacy and Maths minigames, the authors apply generalised linear mixed models to track changes in error counts and play durations. The analysis finds that children generally improved in game performance, with fewer errors and shorter completion times appearing across later rounds and repeated play. Age and gender emerged as the most consistent predictors, with older children and female players performing better, while higher difficulty levels reduced performance. Two episodes are analysed separately, one focused on literacy and one on numeracy, lending the work a dual-domain scope.

The authors position the study as an exploratory, first-of-its-kind investigation of real-world EDURINO gameplay data, grounded in learning analytics and educational data mining methods. Their central conclusion is that the observed gains reflect within-game learning progress and support the app's developmental appropriateness for its target age range. However, they stress that these short-term improvements are confined to specific skills and do not yet demonstrate broader or transferable learning outcomes. The findings therefore motivate longitudinal efficacy research before stronger claims about educational impact can be made. The study's methodological framing draws on prior work on educational apps, game-based assessment, and preschool literacy and numeracy outcomes.