THE IMPACT OF AI FEEDBACK SYSTEMS ON FRUSTRATION TOLERANCE AND LEARNING PERSISTENCE AMONG CHINESE UNIVERSITY STUDENTS

Abstract

 This study tested whether an AI‐driven feedback system—engineered for immediacy, specificity, and supportiveness—enhances Chinese undergraduates’ frustration tolerance and learning persistence. The study conducted a six-week, parallel-group quasi-experimental study at a comprehensive university in China (N = 150). Students were randomly assigned to an experimental group receiving real-time, process-oriented AI feedback during twice-weekly standardised tasks or a control group receiving delayed, summary human feedback without process guidance. Outcomes were measured at baseline and post-test. A brief manipulation check verified perceived feedback immediacy, specificity, and supportiveness. Manipulation checks confirmed clear experiential differences between conditions. Relative to controls, students receiving AI feedback showed significantly greater gains in both frustration tolerance and learning persistence across the intervention period (significant Group × Time interactions for both outcomes). Results identify three actionable design principles for scalable formative feedback in large classes: short feedback latency, actionable next steps, and supportive tone. These features strengthened non-cognitive capacities that underwrite sustained engagement.