Scholarlink

This study investigates learner engagement behaviors in two distinct computing courses delivered online through Chulalongkorn University’s MOOC platform: Python Programming and Pre-Calculus. By analyzing learner 0engagement data, we identified key trends and challenges in both courses. In the Python Programming course, learners demonstrated selective engagement by skipping certain sections and focusing only on topics of interest. Interestingly, several sections had lower video-viewing rates compared to the number of assessments taken, suggesting that some learners possessed prior knowledge of the material and primarily enrolled to obtain certification. Specific challenges emerged in the ‘Variable’ and ‘String’ sections, which appeared to discourage learners. To address these issues, we employed three Large Language Models (ChatGPT, Gemini, and Claude) to analyze the course’s table of contents. Their evaluations offered consistent and actionable recommendations for reorganizing and grouping content. In contrast, the Pre-Calculus course revealed a steady decline in participation over time rather than selective engagement. The evaluation of the course by the same LLMs highlighted a significant issue: the lack of coherence and connection between lessons. These findings underscore the utility of LLMs not only as analytical tools for content evaluation but also as partners in iterative course design.

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

Enhancing Learner Engagement in Chulalongkorn University MOOC Computing Courses: Insights from Behavioral Trends and Analyses of Multiple Large Language Models

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Karin Huangsuwan,
  • Nagul Cooharojananone,
  • Proadpran Punyabukkana,
  • Atiwong Suchato,
  • Yupin Puangngam,
  • Pattamon Bunram

摘要

This study investigates learner engagement behaviors in two distinct computing courses delivered online through Chulalongkorn University’s MOOC platform: Python Programming and Pre-Calculus. By analyzing learner 0engagement data, we identified key trends and challenges in both courses. In the Python Programming course, learners demonstrated selective engagement by skipping certain sections and focusing only on topics of interest. Interestingly, several sections had lower video-viewing rates compared to the number of assessments taken, suggesting that some learners possessed prior knowledge of the material and primarily enrolled to obtain certification. Specific challenges emerged in the ‘Variable’ and ‘String’ sections, which appeared to discourage learners. To address these issues, we employed three Large Language Models (ChatGPT, Gemini, and Claude) to analyze the course’s table of contents. Their evaluations offered consistent and actionable recommendations for reorganizing and grouping content. In contrast, the Pre-Calculus course revealed a steady decline in participation over time rather than selective engagement. The evaluation of the course by the same LLMs highlighted a significant issue: the lack of coherence and connection between lessons. These findings underscore the utility of LLMs not only as analytical tools for content evaluation but also as partners in iterative course design.