<![CDATA[This study aims to analyze the effect of mathematical reasoning ability through a Deep Learning approach using bowling games in Data Presentation material for Grade III Elementary School students. The research purposes were to examine the significance of the difference between pre-test and post-test scores, and to measure the improvement level of students' mathematical reasoning abilities. The research methods employed a pre-experimental design with a one-group pretest-posttest model involving 32 Grade III students. The research results demonstrated significant improvement: the pre-test mean score was 62.50 and the post-test mean score rose to 84.07, indicating an increase of 21.57 points. The normality test (Shapiro-Wilk) showed both data distributions were normal (sig. = 0.554 and 0.958, p > 0.05). The homogeneity test (Levene) confirmed homogeneous variance (sig. = 0.545, p > 0.05). The One-Sample T-Test yielded a significant result (t = 85.358, sig. = 0.000 < 0.05). The N-Gain score reached 0.58 (58.43%), categorized as moderate improvement. These findings confirm that the Deep Learning approach integrated with bowling game activities effectively enhances students' mathematical reasoning abilities across all six NCTM indicators, particularly in recognizing mathematical arguments, formulating and investigating conjectures, and selecting appropriate reasoning methods.]]>
Copyrights © 2026
