Join us at Lumivero’s virtual conference as we present our research, “ChatGPT in Statistics Education: A Learning Activity to Foster Statistical Thinking.”

Abstract: This presentation proposes a learning activity using ChatGPT, a large language model (LLM), to foster statistical thinking. Despite controversies surrounding ChatGPT, there is emerging evidence of its utility in education. Drawing from existing research, we developed a learning activity that leverages ChatGPT to foster statistical thinking throughout each stage of the statistical investigative cycle. This activity is aligned with recommendations from the Guidelines for Assessment and Instruction in Statistics Education and involves a four-stage investigative cycle, in which ChatGPT aids students in formulating questions, collecting data, analyzing data, and interpreting results. Students critically evaluate ChatGPT's outputs and conduct independent analyses, thereby cultivating their statistical thinking skills and understanding of statistical concepts. Our implementation throughout the development of this activity suggests ChatGPT's proficiency in assisting the learning process despite some discrepancies and inaccuracies in its outputs. This exploration holds promise for developing novel teaching methodologies in statistics and beyond, with the caveat that the evolving nature of such technologies necessitates ongoing monitoring and training for their effective classroom use.