面向系统思维培养的智能化多主体建模学习工具设计与实证研究OA

Systems thinking requires students to recognize complex relationships among multiple variables in complex systems and to engage in nonlinear and cross-level reasoning to make sense of complex systemic phenomena.However,secondary school stu-dents often encounter cognitive difficulties in constructing causal reasoning chains and developing a systemic understanding when learning about complex systems.Although multi-agent-based modeling(MABM)offers advantages in representing dynamic interac-tions,students still struggle to develop systematic reasoning and mechanism-based scientific explanations within such environments.Guided by the principles of"structuring inquiry processes,dynamizing model representations,hierarchically organizing causal reason-ing,and intelligentizing feedback mechanisms",this study designed and developed an intelligent multi-agent-based modeling learning tool and empirically validated its effectiveness.A quasi-experimental design with three conditions was implemented among 104 eighth-grade students:an unsupported multi-agent modeling condition,a condition with hierarchical prompting scaffolds only,and a condition combining hierarchical prompting scaffolds with a chatbot.Group performance was compared in terms of content accuracy,reasoning complexity,and overall reasoning performance.The results showed that both conditions with hierarchical prompting scaffolds significantly outperformed the unsupported condition on all three indicators,whereas adding the chatbot on top of the scaffolds yielded slight but non-significant improvements.These findings suggest that embedding question-chain-based hierarchical prompting scaf-folds in a multi-agent-based modeling environment constitutes an important pathway for fostering middle school students' systems thinking,while the chatbot primarily plays a supplementary role in providing targeted knowledge support and conceptual clarification.The study offers empirical evidence and design implications for the further optimization and application of intelligent simulation-based learning environments.