Students’ low mathematical literacy remains a critical issue in mathematics learning because it may hinder their ability to understand contextual problems, construct mathematical models, perform procedures, and interpret solutions meaningfully. This study aimed to analyze the role of mathematical literacy as a predictor of students’ mathematics learning difficulties in the topic of Two-Variable Linear Equation Systems within the context of deep learning. A quantitative explanatory design was employed. The participants were 35 eighth-grade students from SMP Negeri 2 Sepatan, Tangerang Regency, Indonesia, selected through purposive sampling. Data were collected using a mathematical literacy essay test and a mathematics learning difficulties questionnaire. The mathematical literacy test measured three processes, namely formulating, employing, and interpreting, while the questionnaire assessed difficulties in understanding problems, performing arithmetic operations, and solving problems. The data were analyzed using descriptive statistics, assumption testing, and simple linear regression. The results showed that students’ mathematical literacy was at a moderate level, with a mean score of 20.40 out of 32. The regression analysis indicated that mathematical literacy significantly and negatively predicted mathematics learning difficulties (B = -0.742, t = -4.188, p < 0.05). The coefficient of determination showed that mathematical literacy explained 34.7% of the variance in students’ learning difficulties. These findings suggest that strengthening students’ ability to formulate, employ, and interpret mathematical ideas may reduce learning difficulties and support more meaningful mathematics learning in deep learning-oriented classrooms. The study offers preliminary evidence for targeted literacy-based algebra instruction design.
Copyrights © 2026