Adaptive Game-Based Learning for Artistic Anatomy: Effects on Learning Achievement and Creative Self-Efficacy Among First-Year University Art Students

This study investigates whether interpretable adaptive difficulty regulation in game-based learning enhances learning achievement and creative self-efficacy in artistic anatomy education. An Adaptive Game-Based Learning(AGBL)system was developed using the ADDIE model and grounded in constructivist theory, the adaptive learning framework, and social cognitive theory. The system implements a rule-based, transparent difficulty adjustment mechanism driven by a performance index derived from accuracy and response time. A quasi-experimental design was conducted with 109 first-year university art students, who were assigned to either theAGBLcondition or a Non-Adaptive Game-Based Learning(N-AGBL)condition. Analysis of covariance (ANCOVA) controlling for pretest scores revealed that theAGBLgroup significantly outperformed theN-AGBLgroup in both learning achievement (F(1,106) = 42.56,ppr= .587), and the association remained significant after controlling for group assignment (partialr= .370), providing convergent validity evidence. Offline benchmarking against fixed, random, and Bayesian knowledge tracing scheduling strategies demonstrated that the rule-based mechanism achieved more stable difficulty trajectories while maintaining comparable accuracy. The findings suggest that structured and interpretable adaptive control is associated with improvements in cognitive mastery and motivational belief. This study contributes empirical evidence for rule-based adaptive game-based learning in visual-spatial domains and offers replicable design principles for integrating adaptive regulation into creative learning environments.