The Future of Education
Beyond Accuracy and Speed: A Learning Analytics Approach to Performance Modeling in Digital Phonological Games for Brazilian Children
Tiago José Benedito Eugênio, Federal University of São Paulo (Brazil)
Ana Lucia Hennemann, Universidad Internacional Iberoamericana, US (United States)
Denise De Micheli, Federal University of São Paulo, Brazil (Brazil)
Abstract
The increasing use of digital technologies in education has expanded the possibilities for assessing cognitive processes related to learning. However, performance in digital tasks is still frequently analyzed using isolated metrics, such as accuracy or response time, which may fail to capture the multidimensional nature of cognitive performance. This study proposes a composite indicator, named the Game Progress Index (GPI), designed to integrate accuracy, speed, and efficiency into a standardized measure applied to digital phonological awareness games. A total of 339 Brazilian children in the early stages of literacy development participated in the study and completed seven digital game-based tasks. Different weighting models were tested, and the model showing the strongest association with performance indicators was selected. A regression-based approach was then employed to estimate expected performance, allowing the calculation of standardized residuals and the identification of discrepancies between observed and predicted outcomes. Results demonstrated a strong association between the GPI and performance rate (r = .90), supporting the coherence and sensitivity of the proposed index. The combined analysis of percentile scores and residuals revealed heterogeneous performance profiles, including children with high overall performance but lower-than-expected outcomes, as well as participants with low overall performance but outcomes above model predictions. Findings suggest that traditional normative metrics alone may be insufficient to capture individual variability in digital learning contexts. The proposed approach offers a more sensitive framework for modeling performance in educational games, with potential implications for educational assessment, learning analytics, and game-based cognitive screening.
Keywords: learning analytics; digital assessment; serious games; cognitive performance; phonological awareness
REFERENCES
[1] Shute, V. J., & Rahimi, S. (2021). Review of computer-based assessment for learning in elementary and secondary education. Journal of Computer Assisted Learning, 37(1), 19–32.
[2] Snowling, M. J., & Hulme, C. (2021). The Science of Reading: A Handbook. Oxford, UK: Blackwell Publishing. (pp. 1–608).
[3] Buckingham Shum, S., & Luckin, R. (2019). Learning analytics and AI: Politics, pedagogy and practices. British Journal of Educational Technology, 50(6), 2785–2793.
[4] Siemens, G., & Baker, R. S. (2012). Learning analytics and educational data mining: Towards communication and collaboration. In Proceedings of the 2nd International Conference on Learning Analytics and Knowledge (pp. 252–254). New York, NY, USA: ACM.
