대학 교양 토익 수업에서 ChatGPT 활용의 학습 효과 및 학습자 인식 분석: 학습자 활동분석을 중심으로
Analysis of the learning effects and learner perceptions of ChatGPT use in a college TOEIC course: Focusing on learning activity patterns
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This study examined the effectiveness of ChatGPT-based learning and students’ perceptions in acollege English course. A 9-week experiment was conducted with 42 students in the experimentalgroup using ChatGPT-3.5 and 40 in the comparison group without ChatGPT. The study analyzed preand post-test scores and student survey responses. While the experimental group showed significantimprovement, no statistically significant difference was found between the two groups. ChatGPTactivities were classified into five types: answer checking, question generation, explanation,translation, and follow-up questioning. Among these, 70% involved basic functions, while 30%utilized dynamic interactive features. Higher satisfaction was reported for dynamic interactions likequestion generation. Despite occasional hallucinations, students expressed overall satisfaction andwillingness to continue using ChatGPT. To improve ChatGPT-based learning, three strategies aresuggested: increasing participation, enhancing activity quality, and implementing prompt literacyeducation. First, instructors can promote active participation by setting minimum interactionrequirements. Second, encouraging dynamic features, such as question generation and follow-upquestioning, can deepen learning engagement. Lastly, prompt literacy education is crucial foroptimizing AI-assisted learning. Providing structured guidance and refining prompts can helpstudents maximize ChatGPT’s potential.
Abstract
This study examined the effectiveness of ChatGPT-based learning and students’ perceptions in acollege English course. A 9-week experiment was conducted with 42 students in the experimentalgroup using ChatGPT-3.5 and 40 in the comparison group without ChatGPT. The study analyzed preand post-test scores and student survey responses. While the experimental group showed significantimprovement, no statistically significant difference was found between the two groups. ChatGPTactivities were classified into five types: answer checking, question generation, explanation,translation, and follow-up questioning. Among these, 70% involved basic functions, while 30%utilized dynamic interactive features. Higher satisfaction was reported for dynamic interactions likequestion generation. Despite occasional hallucinations, students expressed overall satisfaction andwillingness to continue using ChatGPT. To improve ChatGPT-based learning, three strategies aresuggested: increasing participation, enhancing activity quality, and implementing prompt literacyeducation. First, instructors can promote active participation by setting minimum interactionrequirements. Second, encouraging dynamic features, such as question generation and follow-upquestioning, can deepen learning engagement. Lastly, prompt literacy education is crucial foroptimizing AI-assisted learning. Providing structured guidance and refining prompts can helpstudents maximize ChatGPT’s potential.