Conversational Artificial Intelligence as a Statistics Learning Scaffold
Statistics courses provide students with skills that can help them critically consume information. Despite the value of these skills, some students do not effectively engage with statistics. For some, this lack of engagement may stem from statistics anxiety or a lack of statistics self-efficacy. Little research has examined the effectiveness of artificial intelligence (AI) chatbots as an educational tool for undergraduate students. Conversational AI, like ChatGPT, could be a beneficial learning tool as it can provide immediate assistance to students struggling to solve problems. However, such technologies may hinder learning, as they can potentially prompt students to further disengage from instructional content. The current study examines the relationship between ChatGPT use and students’ comprehension of statistics in an introductory statistics course for social science students. Students were assigned to complete assignments as usual, use ChatGPT freely, or use scaffolded ChatGPT prompts. We will examine how students’ statistics self-efficacy and anxiety relate to their ChatGPT use and statistics comprehension.