바이브코딩을 활용한 멀티모달 영어 발음 교육 웹 프로그램 개발 및 효과 검증 = Development and Effectiveness Evaluation of a Multimodal English Pronunciation Education Web Program using Vibe Coding

다국어 초록 (Multilingual Abstract)

This study aims to develop EduLip, a multimodal English pronunciation education web program that integrates visual articulatory information with auditory input, using vibe coding and to explore its effects on learners’ affective domain through application in an actual high school classroom. The study departs from the recognition that conventional English pronunciation instruction has been predominantly centered on auditory-based methods.
The study was designed based on the ADDIE instructional design model. The researcher developed EduLip over approximately one month through a total of 83 natural language prompt exchanges using Cursor AI. EduLip was structured as a sequential multimodal learning system comprising four core activities: listening to a native speaker’s TTS audio, recording and comparing one’s own pronunciation, watching a native speaker’s lip-movement video, and recording and comparing one’s own video. A tongue twister module was additionally developed in response to findings from a pre-instructional learner needs survey. The development process consisted of six phases: overall structural design, loading of textbook vocabulary, implementation of the multimodal learning module, integration of native speaker lip-movement videos, completion of the target phoneme review module, and deployment stabilization. The possibilities and limitations of the vibe coding approach were concretely identified at each phase.
The EduLip was implemented in a 100-minute class session with 15 high school students. A post-instruction survey based on ARCS motivational model was administered to measure affective outcomes. The results indicated an overall ARCS mean of 3.70 out of 5.00, with subscale scores as follows: Confidence （M = 4.04）, Satisfaction （M = 3.80）, Relevance （M = 3.65）, and Attention （M = 3.32）. Relatively high scores in the Confidence and Satisfaction subscales suggest that the sequential multimodal learning structure and step by step completion design contributed positively to learners’ sense of confidence and achievement in pronunciation learning. However, decreased attentional engagement due to repetitive activities and insufficient guidance on real life application contexts were identified as areas requiring future improvement.
The significance of this study lies in its empirical demonstration that vibe coding can serve as a practical pathway for non-programmer English teachers to independently design and implement multimodal pronunciation education tools, and in its confirmation that the resulting program can positively influence learners’ confidence and satisfaction in pronunciation learning.

목차 (Table of Contents)

• Ⅰ. 서론 1
• 1. 연구의 필요성 및 목적 1
• 2. 연구 문제 4
• Ⅱ. 이론적 배경 5
• 1. 영어 발음 교육의 흐름 5
• 2. 우리나라의 영어 발음 교육 8
• 3. CAPT와 AI 기반 발음 교육 연구 정리 15
• 3.1 CAPT 기반 발음 교육 15
• 3.2 AI 기반 발음 교육 17
• 4. 멀티모달 발음 학습 설계의 이론적 기반 18
• 4.1 멀티미디어 학습 인지 이론 20
• 4.2 시각적 조음 정보와 멀티모달 발음 학습 21
• 4.3 자기 점검 기반 발음 학습 22
• 5. 바이브코딩 24
• Ⅲ. ADDIE 모형을 활용한 영어 발음 교육 프로그램의 설계 및 개발 과정 27
• 1. 분석 28
• 2. 설계 29
• 2.1 학습 목표 설정 29
• 2.2 어플리케이션 구조 설계 29
• 2.3 바이브코딩 프로그램 선정 30
• 3. 개발 31
• 3.1 1차 개발 33
• 3.2 2차 개발 35
• 3.3 3차 개발 37
• 3.4 4차 개발 41
• 3.5 5차 개발 48
• 3.6 6차 개발 51
• 4. 학습자 요구 조사 및 추가 개발 55
• 4.1 사전 설문지 55
• 4.2 텅 트위스터 모듈 개발 57
• 5. 실행 61
• 6. 평가 63
• Ⅳ. 논의 68
• 1. 바이브 코딩을 활용한 멀티모달 영어 발음 교육 웹 프로그램 개발 과정과 결과 68
• 2. 멀티모달 영어 발음 교육 웹 프로그램이 학습자의 정의적 영역에 미친 영향 74
• Ⅴ. 결론 및 제언 79
• 1. 결론 79
• 2. 제언 81
• 3. 연구의 제한점 83
• 참고문헌 85
• Appendix 94