Simulated dataset and analysis pipeline for "Effects of AI-Driven Multimodal Video Feedback on Aerobic Gymnastics Skill Acquisition and the Role of Self-Efficacy"

This deposit contains the simulated dataset and the analysis pipeline used in a methodological validation study of AI-driven multimodal video feedback for aerobic gymnastics skill acquisition. All values are computer-simulated; no human participants were recruited. The dataset comprises 60 simulated cases (30 per group; 57 analysed after simulated attrition) measured at baseline, week 6 and week 12, including expert composite scores from three simulated raters, lower-limb kinematic parameters, self-efficacy and feedback satisfaction. The analysis scripts reproduce every statistic reported in the manuscript (linear mixed models with Satterthwaite degrees of freedom, bias-corrected bootstrapped mediation and confirmatory factor models, and post hoc power analysis).