Opportunities and Challenges in Generative AI-Based Instruction for Formulating Statistical Investigative Questions

This study analyzed a classroom case in which students used generative AI to formulate statistical investigative questions, aiming to explore the implications of generative AI in teaching and learning. To address the research questions, a lesson was conducted with 20 first-year middle school students, where they used generative AI to formulate statistical investigative questions and design data collection plans. The analysis revealed that the questions formulated with generative AI were characterized by clearly defined variables and populations and required students to address the variability of data. However, students demonstrated limited understanding of these problems, often confusing samples with populations and failing to explain data variability. Students employed prompting strategies such as “Specify structured prompts” and “Split complex tasks into simpler subtasks” to interact productively with generative AI, though some students did not use these strategies effectively. These findings suggest that while generative AI can support students in formulating statistical investigative questions, it has limitations in fostering deeper student understanding. Therefore, it is important for teachers to engage in practices that foster deeper student understanding and to develop instructional approaches that support students’ effective prompt use with generative AI.