The Effect of the Deep Learning-Based Problem-Based Learning (PBL) Model on the Critical Thinking Skills and Mathematical Disposition of Elementary School Students

Abstract

This study aims to analyze the effect and effectiveness of the Deep Learning-based Problem-Based Learning (PBL) model on the critical thinking skills and mathematical dispositions of elementary school students. A quasi-experimental study with a non-equivalent control group design was employed. The population consisted of fifth-grade students in Gugus 3 Kuta, with a total sample of 168 students (84 in the experimental group and 84 in the control group) selected through purposive sampling. The experimental group was treated with the deep learning-based PBL model, while the control group followed conventional learning. Data were collected using a critical thinking skills essay test and a mathematical disposition questionnaire. Analysis included N-Gain tests, Mann-Whitney U tests, and correlation analysis using SPSS. The results indicated a significant difference in outcomes between the two groups. The experimental group achieved a significantly higher post-test average for critical thinking (83.85) compared to the control group (47.65). Statistical tests confirmed that the deep learning-based PBL model significantly improved both critical thinking skills and mathematical dispositions (Asymp. Sig. 0.001 less than 0.05). Furthermore, the model was found to be effective in enhancing both variables simultaneously. The integration of deep learning strategies into the PBL framework—emphasizing conceptual understanding, reflection, and knowledge transfer—creates a more interactive and reflective learning environment. This synergy effectively addresses students' difficulties with contextual problems and fosters a more persistent, confident attitude toward mathematics. In conclusion, the deep learning-based PBL model is a potent alternative for developing 21st-century skills at the elementary level.