This study aims to analyze the effect of the Think Pair Share learning model assisted by the traditional marble game on the mathematical numeracy skills of fifth-grade elementary school students. This study implemented a quasi-experimental method with a pretest-posttest control group design involving an experimental group and a control group. The experimental group received treatment through the TPS learning model assisted by the traditional marble game, while the control group applied conventional learning methods. The research instrument consisted of tests administered during the pretest and posttest to measure students’ numeracy skills. The findings revealed that the average pretest score of the experimental group was 58.81 and increased to 88.66 in the posttest, while the average posttest score of the control group was 69.12. The results of the hypothesis testing indicated that the Sig. (2-tailed) value was 0.000 < 0.05, showing a significant effect of the TPS learning model assisted by the traditional marble game on students’ numeracy skills. In addition, the N-gain calculation results showed that the experimental group obtained a score of 0.72 (high classification), whereas the control group obtained a score of 0.34 (moderate classification).
Copyrights © 2026