Exploring the Key Affordances of AI Tutors and Their Influence on Learning Outcomes: Using Structural Equation Modeling and Fuzzy-Set QCA
- 주제(키워드) AI tutor , ITS , Affordance , Delphi Method , PLS-SEM , fsQCA , Affordance Process Model , SOR Theory
- 주제(DDC) 658
- 발행기관 아주대학교 일반대학원
- 지도교수 Juyoung Kang
- 발행년도 2024
- 학위수여년월 2024. 2
- 학위명 박사
- 학과 및 전공 일반대학원 비즈니스애널리틱스학과
- 실제URI http://www.dcollection.net/handler/ajou/000000033484
- 본문언어 한국어
- 저작권 아주대학교 논문은 저작권에 의해 보호받습니다.
초록/요약
This research investigates the evolving landscape of artificial intelligence (AI) in education, focusing on AI tutors and Intelligent Tutoring Systems (ITS). AI’s rapid advancement, particularly in generative AI and natural language processing (NLP), is reshaping educational methodologies, offering personalized, adaptive learning experiences tailored to individual needs. The study aims to understand the primary affordances of AI tutors and their impact on student performance and learning outcomes. Employing the Delphi method, insights from experts across technology, education, and business domains were amalgamated to identify key affordances intrinsic to AI tutors. The study integrates the Stimulus-Organization-Response (S-O-R) model and affordance process modeling to explore the influence of these affordances on learning outcomes, using structural equation modeling (PLS-SEM) for empirical validation. Additionally, fsQCA methodology was employed to examine the interplay between diverse affordances and student engagements, such as presence, motivation, and their contribution to learner performance. The research findings provide a comprehensive understanding of the role of AI tutors in educational environments. The identified affordances and their relationship with learner outcomes highlight the potential of AI tutors in enhancing educational experiences. This study contributes to the dialogue on AI tutors’ effectiveness, offering insights into their long-term impacts and opportunities and presenting recommendations for integrating AI advancements in education. The collective insights from the studies are anticipated to guide the future development of AI tutors, enhancing their applicability and effectiveness in learning environments.
more목차
1. Introduction 1
1.1. Overview 1
1.2. Overall Scope 3
1.3. Theoretical Foundations 4
1.3.1. What Is AI Tutors 4
1.3.2. AI Technologies in AI Tutors 6
1.3.3. Development of AI Tutors 8
1.3.4. Characteristics of AI Tutors 10
1.3.5. Affordances of AI Tutors 11
1.3.6. Exploring Key Affordances of AI Tutors 13
2. Study 1: Exploring Key Affordances of AI Tutors Using Delphi Method 14
2.1. Design of Delphi Study 14
2.1.1 The Delphi Method and Revised Delphi Method 14
2.1.2 Panel Selection 16
2.1.3 Analysis of Surveys 17
2.2. Result 19
2.3. Key Affordances 22
2.3.1 Adaptivity 22
2.3.2 Feedback 23
2.3.3 Multimodality 25
2.3.4 Scaffolding 26
2.3.5 Metacognitive Support 28
2.4. Implications and Limitations 29
3. Study 2: AI Tutors and Their Influence on Learning Outcomes: Using SEM 32
3.1. Introduction of Study 2 32
3.2. Theoretical Background 33
3.3. Research Model and Hypotheses Development 35
3.3.1. Research Model of Study 2 35
3.3.2. Characteristics of the AI Tutor Affordances 36
3.3.3. Learning Outcome 39
3.3.4. Mediation process 41
3.4. Research Methodology 44
3.4.1. Affordance Confluence as a Second-order 44
3.4.2. Measurements 46
3.4.3. Data Collection 48
3.4.4. Hierarchical Model 50
3.4.5. Hypothesis 52
3.5. Research Results 54
3.5.1. Measurement Model Analysis 54
3.5.2. Hypotheses Test Results 62
3.5.3. Multi-Group Analysis 65
3.5.4. Discussion of Results 71
3.6. Contributions to Research and Practice 73
3.6.1. Theoretical Implications 73
3.6.2. Practical Implications 74
3.6.3. Limitations and Future Research 76
3.7. Conclusion 77
4. Study 3: AI Tutors and Their Influence on Learning Outcomes: Using fsQCA 79
4.1. Introduction of Study 3 79
4.2. Theoretical Background 81
4.2.1. Complexity Theory and Configuration Theory 81
4.2.2. Benefits and Limitations of Configurational Analysis Applying in Education 82
4.3. Conceptual Model and Research Propositions 84
4.3.1. Conceptual Model 84
4.3.2. Research Proposition 85
4.4. Research Method 86
4.4.1. Data collection 86
4.4.2. Measurements 86
4.5. FsQCA 87
4.5.1. Data Calibration 87
4.5.2. Analysis of Necessary Conditions 88
4.5.3. Analysis of Sufficient Condition Sets 89
4.5.4. Testing Predictive Validity 93
4.6. Discussion and Implications 95
4.6.1. Implications 95
4.6.2. Limitations and Future Research 97
5. Conclusion 98
5.1. Key Findings of the Sub-study 98
5.2. Comprehensive Theoretical Contributions 100
5.3. Comprehensive Practical Contributions 101
5.4. Limitations of Comprehensive Research and Future Research Direction 102
5.5. Concluding Remarks 103
References 105
