<![CDATA[This study aims to compare the effectiveness of Deep Learning-based mathematics instruction and conventional instruction in developing seventh-grade students’ critical reasoning skills. The method used was a qualitative comparative approach with a multiple case study design involving two classes and two mathematics teachers selected purposively, with data collected through observation, interviews, and documentation and analyzed using Miles and Huberman’s thematic analysis model. The findings indicate that Deep Learning-based instruction is more effective in promoting students’ reflective thinking, strategy evaluation, and concept generalization compared to conventional instruction, which tends to emphasize procedural mastery. In conclusion, Deep Learning-based instruction contributes more significantly to the development of students’ critical reasoning skills; however, its effectiveness is influenced by the quality of teacher facilitation and task design, therefore the integration of both approaches is recommended to achieve optimal learning outcomes. Keywords: Critical Reasoning; Deep Learning; Mathematics Learning; Conventional Learning.]]>
Copyrights © 2026
