<![CDATA[Mathematics learning outcomes in Indonesian elementary schools remain a persistent challenge, with national assessments revealing that many third-grade students have not yet achieved minimum competency thresholds. Conventional instruction-dominated classrooms, combined with limited use of concrete learning media, contribute significantly to low conceptual understanding among early-grade learners. This study aimed to examine the effect of the Teams Games Tournament (TGT) cooperative learning model assisted by natural gravel manipulatives on the mathematics learning outcomes of third-grade elementary school students in Cluster 2, Tanjung Mutiara District, Agam Regency, West Sumatra. A quantitative approach with a quasi-experimental pretest-posttest control group design was employed. The experimental group comprised 9 students from SDN 17 Bukik Malintang, while the control group consisted of 12 students from SDN 06 Bukik Malintang, selected through purposive sampling based on the highest inter-school homogeneity. Data were collected through mathematics achievement tests and analyzed using Normalized Gain (N-Gain) scores, Independent Sample t-test, Mann-Whitney U test as a robustness check, and Cohen’s d effect size. Results indicated a statistically significant difference in learning improvement between the two groups (t = 2.609; p = 0.017). The experimental group achieved a mean N-Gain score of 67.58%, compared to 37.18% in the control group. The effect size analysis yielded Cohen’s d = 1.150 and Hedges' g = 1.104, both categorized as large effects. These findings demonstrate that the integration of TGT with natural gravel manipulatives represents a promising, resource-efficient pedagogical intervention for enhancing mathematics learning outcomes in concrete operational stage learners.]]>
