Reflective learning is crucial in programming education, fostering critical thinking and problem-solving skills. However, traditional guided reflection often falls short in addressing individual challenges and providing personalized feedback. Recent studies exploring large language models like ChatGPT for personalized feedback have primarily focused on immediate solutions, potentially hindering the development of higher-order thinking skills (HOTS). This study investigates the efficacy of ChatGPT-generated personalized feedback reflective reports on undergraduate programming students’ self-efficacy, HOTS, and project implementation skills. Using a quasi-experimental design, 79 students in a Python programming course were divided into experimental and control groups over two semesters. The experimental group received ChatGPT-generated reports based on their error codes, while the control group engaged in traditional guided reflection. Results showed that the ChatGPT-facilitated approach significantly enhanced students’ self-efficacy in logical thinking, algorithm design, and debugging. Moreover, students demonstrated improved critical thinking and problem-solving skills. The experimental group also produced more logical, useful, and well-crafted projects, as evaluated using the Creative Product Analysis Matrix (CPAM). This research underscores the potential of AI-supported personalized feedback in enhancing reflective learning practices and HOTS development in undergraduate programming education. It offers valuable insights for educators and researchers in higher education contexts, potentially transforming how universities prepare students for the challenges of the digital age and contributing to the development of a more skilled and adaptable workforce. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.
Author Keywords
Index Keywords
