Generative AI in Statistics Education: Investigating the Role of ChatGPT and English Proficiency in Improving Statistical Literacy

Abstract

The integration of generative Artificial Intelligence (AI), particularly ChatGPT, has transformed statistics education by providing students with instant explanations, problem-solving support, and personalized learning experiences. However, previous studies have primarily examined the educational benefits of ChatGPT without adequately considering the role of students' English language proficiency, despite the fact that most AI-generated resources and interactions are conducted in English. This gap raises the question of whether ChatGPT utilization alone is sufficient to improve students' statistical literacy or whether English language proficiency also plays a significant role. Therefore, this study aims to examine the influence of ChatGPT Utilization and English Language Proficiency on the Statistical Literacy of Information Systems students. Using a quantitative associative approach with a saturated sampling technique, data were collected from 45 students enrolled in the Statistics course at Institut Teknologi & Bisnis Bina Sarana Global during the 2025/2026 academic year through a Likert-scale questionnaire and analyzed using multiple linear regression after the classical assumption tests were satisfied. The results yielded the regression equation Y = 11.026 + 0.280X₁ + 0.465X₂. Simultaneously, ChatGPT Utilization and English Language Proficiency significantly affected Statistical Literacy (F = 22.359; p < 0.001), with a coefficient of determination (R²) of 0.516, indicating that 51.6% of the variance in statistical literacy was explained by the two predictors. Partially, ChatGPT Utilization had a positive and significant effect (t = 2.211; p = 0.032), as did English Language Proficiency (t = 4.016; p < 0.001), with the latter showing a relatively stronger contribution (β = 0.519 compared to β = 0.286). These findings indicate that the effectiveness of ChatGPT in supporting statistics learning depends not only on its utilization but also on students' English language proficiency. Accordingly, integrating generative AI with strategies to strengthen English proficiency may provide a more effective approach to improving statistical literacy in higher education.