A Meta-analysis of Pedagogical Agents promoting Self-Regulated Learning

As generative artificial intelligence evolves, pedagogical agents are increasingly used to support self-regulated learning, but their effectiveness remains controversial. To fill this research gap, we conducted a meta-analysis of 32 international studies. Our results show a moderate positive impact of pedagogical agents on self-regulated learning, influenced by publication time, region, experimental duration, agent type, and role. Further exploration of their educational applications, refinement of design features, and optimization of impact evaluation are essential. This study offers valuable insights for designing and implementing pedagogical agents to promote self-regulated learning.