Designing Chatbots to Support Self-Regulated Learning and Systems Thinking

• In Event: AERA Roundtable Session Thursday 9:45 am - Gold 4
  In Roundtable Session: Roundtable: Informal Learning Practices in the Learning Sciences

Abstract

We investigate how self-regulated learning (SRL) scaffolds—implemented with large language model-based chatbots—support students’ SRL and systems thinking during scientific modeling. Students in grades 11-12 participated in a two-week curriculum on marine ecosystems, where they interacted with chatbots to refine their models. Students were randomly assigned to SRL-embedded (n=16) or non-SRL chatbot conditions (n=15). In this paper, we describe the chatbots’ prompt design following SRL theories to support forethought, performance, and reflection. While we did not find differences between groups in systems thinking or SRL strategies, findings highlight design considerations for future interventions. These include simplifying chatbots’ responses and making SRL prompts more prominent. This research contributes practical implications for designing and implementing SRL-embedded chatbots to foster systems thinking.

Authors

• Saerok Park, University of North Carolina - Chapel Hill

• Ha Nguyen, University of North Carolina - Chapel Hill

• Jie Cao, University of North Carolina - Chapel Hill