A data-driven comparison of game based and problem based learning in primary education

Abstract

This study aims to compare the effectiveness of Game-Based Learning (GBL) and Problem-Based Learning (PBL) in elementary education and to identify the factors that influence student learning outcomes through a data-driven approach. The method used is a quantitative analysis based on secondary data from 48 empirical articles reporting on the implementation of GBL and PBL at the elementary school level. The variables analyzed include pretest and posttest scores, sample size, intervention duration, grade level, and subject area. Data analysis was conducted using descriptive statistics, independent samples t-tests, effect size calculations, and regression analysis. Additionally, machine learning techniques were used to identify the primary predictors contributing to learning outcomes. The results indicate that PBL yields significantly higher posttest performance compared to GBL, with a large effect size. Intervention duration and students’ prior ability were identified as the most influential factors on learning outcomes. These findings suggest that while both approaches make positive contributions, PBL is more effective in improving elementary school students’ learning achievements. This study offers practical implications for educators in selecting appropriate learning strategies and underscores the importance of data-driven approaches in supporting more adaptive and evidence-based educational decision-making.