基於知識翻新理論之AI素養導向程式教學對國小高年級學生AI學習及程式設計自我效能之影響;The Impact of Knowledge-Building-Oriented AI Literacy Programming Instruction on Elementary School Students’ AI Learning and Programming Self-Efficacy

NCU Institutional Repository > 資訊電機學院 > 網路學習科技研究所 > 博碩士論文 > Item 987654321/98383

請使用永久網址來引用或連結此文件: https://ir.lib.ncu.edu.tw/handle/987654321/98383

題名:	基於知識翻新理論之AI素養導向程式教學對國小高年級學生AI學習及程式設計自我效能之影響;The Impact of Knowledge-Building-Oriented AI Literacy Programming Instruction on Elementary School Students’ AI Learning and Programming Self-Efficacy
作者:	官慶惠;Kuan, Ching-Hui
貢獻者:	網路學習科技研究所
關鍵詞:	知識翻新;素養導向教學;AI素養;程式教學;生成式AI;人機協作;knowledge building;competency-based instruction;AI literacy;programming;generative AI;human-AI collaboration
日期:	2025-07-16
上傳時間:	2025-10-17 12:42:42 (UTC+8)
出版者:	國立中央大學
摘要:	本研究旨在探討基於知識翻新理論之AI素養導向程式教學，結合不同人機協作模式（「一人一機」與「兩人一機」），對國小高年級學生在「AI學習」與「程式設計自我效能」的影響。研究採準實驗設計，以新北市某國小五年級四個班級做為研究對象，依協作模式分為兩組，實施為期六週之生成式AI輔助網頁遊戲設計課程。課程結束後，運用AI學習量表、程式設計自我效能量表與兒童遊戲設計創意評量Rubric量表進行學習成效分析，並透過成對t檢定（paired sample t test）與共變數分析（ANCOVA）檢驗本教學方法與不同協作模式對學習成果之差異。為補充量化結果之詮釋與理解，亦針對焦慮明顯上升或下降的學生進行半結構式訪談，探討其焦慮形成原因、調節策略與未來態度。研究結果顯示，「兩人一機」協作模式的學生在「AI學習」與「程式設計自我效能」上皆有明顯進步，凸顯合作歷程在促進理解、強化信心與提升創造力方面的重要性；相較之下，「一人一機」組在部分面向成效有限。此外，本研究亦發現學生於課程中能透過同儕討論協調輸入內容，並針對AI回應進行判斷，接著逐步建立與AI互動的語言表達策略。惟部分學生在初期操作過程中感受到壓力，對AI的焦慮感略有提升，學生需在理解AI生成邏輯以及人類智慧與人工智慧協作上獲得更多支持。質性資料進一步指出，學生對AI的焦慮感受主要來自於擔心被取代、操作與提問困難、生成錯誤回應、對AI的未知與不確定性以及來自學習負擔或其他額外壓力。其針對壓力的調節策略又可分為自我調節與尋求他人協助，多數學生仍展現對未來使用AI之正向態度，展現出積極的學習意圖與應用想像。本研究結果肯定「兩人一機」的人機協作模式在AI教育中的應用價值，並提供教師於設計AI素養導向課程與規劃協作模式時之實務參考。關鍵字：知識翻新、素養導向教學、AI素養、程式教學、生成式AI、人機協作;This study aims to explore the impact of a knowledge-building-based AI literacy-oriented programming curriculum—integrated with different human-computer interaction modes (one student per device vs. two students per device)—on upper elementary school students’ AI learning and programming self-efficacy. A quasi-experimental design was adopted, involving four fifth-grade classes from an elementary school in New Taipei City. Students were divided into two groups based on collaboration mode and participated in a six-week generative AI-assisted web game design course. After the course, learning outcomes were assessed using the AI Learning Scale, Programming Self-Efficacy Scale, and a Game Design Creativity Rubric. Paired sample t-tests and ANCOVA were conducted to examine the effects of the instructional model and collaboration modes on student outcomes. To complement the quantitative findings, semi-structured interviews were conducted with students whose AI anxiety scores increased or decreased significantly, to explore the causes of their anxiety, regulation strategies, and future attitudes toward AI. The results showed that students in the “two students per device” collaboration group made significant progress in both AI learning and programming self-efficacy, highlighting the importance of collaboration in promoting comprehension, building confidence, and enhancing creativity. In contrast, the “one student per device” group showed limited improvement in several aspects. Additionally, students were found to engage in peer discussion to coordinate input and evaluate AI responses, gradually developing strategies for communicating effectively with AI. However, some students experienced pressure during the initial stages of the course, and their anxiety about AI increased slightly, suggesting a need for more support in understanding AI’s generative logic and the collaboration between human and artificial intelligence. Qualitative data further revealed that students’ AI-related anxiety stemmed mainly from fears of being replaced, difficulties with operation and questioning, incorrect AI responses, uncertainty about AI, and extra burdens related to learning or collaboration. Their regulation strategies included both self-regulation and seeking assistance from others. Nevertheless, most students expressed a positive attitude toward future AI use, demonstrating strong learning intentions and imaginative applications. The findings affirm the value of the “two students per device” human-computer collaboration model in AI education and provide practical implications for teachers designing AI literacy-oriented curricula and collaborative learning models. Keywords: knowledge building, competency-based instruction, AI literacy, programming, generative AI, human-AI collaboration
顯示於類別:	[網路學習科技研究所 ] 博碩士論文

文件中的檔案:

檔案	描述	大小	格式	瀏覽次數
index.html		0Kb	HTML	19	檢視/開啟

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....