The relationship between instructional feedback and academic performance: the role of artificial intelligence perception

Abstract

Instructional feedback can effectively improve academic performance of students. This research aims to explore the impact of instructional feedback on academic performance and to assess the indirect association of AI perception in this relationship. A cross-sectional study design was employed, conducted at Jining Medical University. The independent variable was instructional feedback, the dependent variable was the academic performance, and the mediating variable was the AI perception level. First, univariate and multiple regression analysis were conducted. Then, the bootstrap method was used to test the significance of the mediating effect. Finally, a restricted cubic spline (RCS) model was used to explore potential non-linear associations between AI perception and instructional feedback. A total of 470 medical undergraduate students were included in the analysis. Among the participants, 349(74.255%) respondents indicated that lecturers used AI to provide timely instructional feedback. Multivariate regression analysis revealed that academic performance was higher among female students (P = 0.017), urban students (P < 0.001), and those reporting instructional feedback (P = 0.002). Mediation analysis showed that instructional feedback was associated with AI perception (β = 0.465, P = 0.041), but AI perception was not significantly associated with academic performance after adjustment (β = 0.227, P = 0.110). RCS analysis suggested an overall association between AI perception and instructional feedback with no evidence of non-linearity (P-overall = 0.044; P-non-linear = 0.648). Instructional feedback was associated with academic performance among medical undergraduates, and it was also associated with AI perception. However, mediation analysis did not support a statistically significant indirect association via AI perception. RCS analysis suggested an overall association between AI perception and the odds of reporting receipt of instructional feedback, with no evidence of non-linearity. These findings suggest that integrating AI-enabled instructional feedback into course delivery may be a useful approach to support learning, while the role of AI perception as an explanatory pathway warrants further investigation.

Ähnliche Arbeiten