ChatGPT-enhanced self-regulated learning in programming education : impacts on motivation, self-efficacy, and learning outcomes

Abstract

ChatGPT shows potential for enhancing self-regulated learning (SRL) in education. This study examined its role in programming instruction through a seven-week intervention with 83 sophomores. Students were assigned to a control group receiving traditional instruction (N = 27), an experimental group using ChatGPT (N = 30), or a group combining ChatGPT with SRL strategies (N = 26). Two-way ANCOVA results indicate that ChatGPT-supported groups reported higher motivation (p <.05, η2 =.052) and engagement (p <.01, η2 =.117) than the control group. Integrating SRL further improved self-efficacy (p <.001, η2 =.152) and motivation (p <.01, η2 =.094). However, no significant differences emerged in programming knowledge acquisition (p =.79), suggesting limitations of AI-based support for conceptual mastery. Possible explanations include the need for more interactive and scaffolded activities to promote in-depth learning. A gender imbalance (66 males, 17 females) also limits the generalizability of findings. Future research may investigate structured learning activities and qualitative approaches to better capture learners experiences. Overall, this study highlights the value of ChatGPT and SRL in promoting motivation, engagement, and self-efficacy while underscoring the challenges of advancing programming knowledge.