Raw Data for the Paper Titled "Can real-time feedback driven by artificial intelligence improve English learners' oral performance? Evidence from a controlled trial of 128 non-English majors"

Abstract

The rapid development of generative artificial intelligence technology provides a brand-new technical possibility to solve the common practical dilemma of "teaching in large classes and insufficient feedback" in college oral English teaching for a long time. In this study, 128 non-English major sophomores in a local university were taken as experimental subjects, and a 16-week randomized controlled teaching experiment was carried out. The subjects were randomly assigned to the experimental group and the control group. The experimental group was taught by real-time oral feedback teaching mode driven by artificial intelligence, while the control group followed the traditional teacher-led oral feedback teaching mode, and the AI was systematically tested by mixed research methods such as pre-and post-test comparison, three-dimensional coding of oral output CAF (complexity, accuracy and fluency) and semi-structured in-depth interviews. The results show that after 16 weeks of teaching intervention, the improvement of oral output accuracy and fluency in the experimental group is significantly higher than that in the control group, and there is no statistically significant difference in complexity between the two groups. Real-time feedback of AI can greatly improve the frequency of students' oral practice, and at the same time effectively reduce students' anxiety level of oral expression. This study confirms that artificial intelligence real-time feedback can be an effective supplement to traditional teacher feedback and provide solid empirical evidence for the reform of college oral English teaching mode.