<![CDATA[Mathematical representation ability is an essential skill that enables students to connect mathematical concepts with real-life contexts. However, many senior high school students still experience difficulties in representing mathematical ideas, particularly in modeling word problems, which is influenced by inappropriate instructional models and low levels of self-efficacy. This study aims to examine differences in students’ mathematical representation abilities between those taught using the Quantum Learning model and those taught through direct instruction, as well as to analyze the interaction between instructional models and self-efficacy on mathematical representation ability. This study employed a quantitative approach using a quasi-experimental 2×2 factorial design. The participants consisted of 101 eleventh-grade students selected through multistage random sampling. Data were collected using a mathematical representation ability test and a self-efficacy questionnaire, and were analyzed using descriptive statistics, prerequisite tests (normality and homogeneity), and two-way analysis of variance (2×2 ANOVA) followed by Tukey’s post hoc test. The results indicate that the Quantum Learning model is significantly more effective than direct instruction in improving students’ mathematical representation abilities. Furthermore, a significant interaction was found between the instructional model and students’ self-efficacy. Students with high self-efficacy demonstrated better mathematical representation abilities when taught using Quantum Learning, whereas students with low self-efficacy benefited more from direct instruction. This study concludes that the effectiveness of mathematics instructional models is strongly influenced by students’ affective characteristics, particularly self-efficacy, highlighting the importance of adaptive teaching strategies to optimize students’ mathematical representation abilities.]]>
