Scholarlink

<p>This study explores the potential of generative AI (GenAI) to facilitate elementary students’ epistemic understanding in scientific inquiry. Utilizing the Korean version of Views About Scientific Inquiry for Elementary school students (VASI-E) instrument, we collected 560 responses from 80 fourth-grade students in Korea and implemented a series of prompt engineering strategies within ChatGPT-4o. We developed GenAI prompts for scoring and feedback aligned with established epistemic frameworks, categorizing responses as naive, mixed, or informed. GenAI’s scoring was compared to human raters, showing high agreement (κ = 0.825), with kappa values ranging from 0.606 to 0.923 across items. GenAI-generated feedback was analyzed for learner-centered feedback quality based on three criteria: future impact, sensemaking, and student agency. The feedback was found to be generally appropriate (average score = 2.75/3), with notable strength in promoting student agency through personalized and reflective guidance. However, some linguistic challenges were observed in Korean phrasing, affecting clarity. The findings suggest that GenAI, when properly designed and prompted, can function as a dialogic partner for facilitating students’ epistemic understanding. This study offers new pathways for integrating GenAI into formative assessment and teacher education, especially for supporting student understanding of the nature of scientific inquiry.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Harnessing Generative AI to Facilitate Epistemic Understanding of Students in Scientific Inquiry

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月10日

查看全文
  • Jina Chang,
  • Jisun Park,
  • Ju Yeon Sim

摘要

This study explores the potential of generative AI (GenAI) to facilitate elementary students’ epistemic understanding in scientific inquiry. Utilizing the Korean version of Views About Scientific Inquiry for Elementary school students (VASI-E) instrument, we collected 560 responses from 80 fourth-grade students in Korea and implemented a series of prompt engineering strategies within ChatGPT-4o. We developed GenAI prompts for scoring and feedback aligned with established epistemic frameworks, categorizing responses as naive, mixed, or informed. GenAI’s scoring was compared to human raters, showing high agreement (κ = 0.825), with kappa values ranging from 0.606 to 0.923 across items. GenAI-generated feedback was analyzed for learner-centered feedback quality based on three criteria: future impact, sensemaking, and student agency. The feedback was found to be generally appropriate (average score = 2.75/3), with notable strength in promoting student agency through personalized and reflective guidance. However, some linguistic challenges were observed in Korean phrasing, affecting clarity. The findings suggest that GenAI, when properly designed and prompted, can function as a dialogic partner for facilitating students’ epistemic understanding. This study offers new pathways for integrating GenAI into formative assessment and teacher education, especially for supporting student understanding of the nature of scientific inquiry.