远程教育杂志2025，Vol.43Issue(6)：22-32,11.DOI:10.15881/j.cnki.cn33-1304/g4.2025.06.004
像科学家一样思考:基于大语言模型的科学教育思想实验设计
Thinking Like a Scientist:Designing Thought Experiments Based on Large Language Models in Science Education
摘要
Abstract
Emerging technologies such as large language models(LLMs),with their advanced capabilities in thought simula-tion,logical reasoning,and conceptual analysis,provide new possibilities for the digital transformation of thought experiment instruc-tion in science education.Building on this potential,this study integrates LLMs into thought experiment-based teaching to investigate how such models can foster students'scientific thinking,critical analysis,and innovative capacities.Drawing on Reiner's structure of thought experiments,the study develops an LLM-supported instructional framework consisting of five sequential stages:problem intro-duction,hypothesis formulation,simulated experimentation,analytical discussion,and reflective summarization.An empirical case study was conducted using the"Earth's Gravity"lesson in a fifth-grade science class at H primary school in Shanghai.The findings indicate that the framework enhances students'conceptual understanding and scientific inquiry abilities,supports the development of scientific thinking,and helps students approach problems in ways similar to scientists.By systematizing thought experiment instruc-tional activities and embedding them into digital tools,this study provides both theoretical justification and practical guidance for the digital transformation of thought experiment pedagogy in science education.It also demonstrates the potential value of such approaches in strengthening scientific inquiry and informing instructional improvement.关键词
思想实验/科学教育/大语言模型/教学设计Key words
Thought experiments/Science education/Large language models/Instructional design分类
社会科学引用本文复制引用
靳旭莹,刘泽民..像科学家一样思考:基于大语言模型的科学教育思想实验设计[J].远程教育杂志,2025,43(6):22-32,11.基金项目
本文系2025年全国教育科学规划博士生项目"大模型时代思想实验促进中学生科学素养发展的理论框架与技术路径"(项目编号:FTA250469)的研究成果. （项目编号:FTA250469）
