Improving Students' Mathematical Self-Regulated Learning and Mathematical Flexibility with a Deep Learning Approach

Abstract

The rapid development of science and technology demands a transformation in mathematics education to develop higher-order thinking skills. This study investigates the effectiveness of deep learning approaches in improving mathematical flexibility (MF) and mathematics self-regulated learning (MSRL) among grade 8 junior high school students in Indonesia. Employing a sequential explanatory mixed-methods design, this quasi-experimental study involved 99 students (68 experimental, 31 control) who completed pre-tests and post-tests measuring MF and MSRL. Quantitative data were analyzed using ANCOVA, t-tests, and regression analysis, followed by qualitative interviews and classroom observation to explain the findings. Results revealed that the deep learning approach significantly improved students' mathematical flexibility (t = 14.92, p 0.05), with an average gain of 4.27 points, but did not significantly improve MSRL (p = 0.165). Regression analysis demonstrated a significant positive relationship between MSRL and MF (r = 0.524, p 0.001), with MSRL accounting for 27.2% of the variance in mathematical flexibility. This study contributes to mathematics education research by providing empirical evidence that deep learning approaches effectively develop mathematical flexibility while highlighting the need for multifactorial interventions targeting affective and metacognitive dimensions to improve self-regulated learning.