Exploring How to Improve Programming Learning Experience with Generative AI Tools: An Integrated Means-End Chain and fsQCA Approach

Abstract

Recent advancements in generative artificial intelligence (AI) are poised to transform education, but doubts persist about their effectiveness in enhancing the learning experience. Previous studies were more “technology-driven” rather than “learner-driven” and largely neglected how different configurations (specific combinations of variables) affect generative AI-assisted learning experience. Based on the means-end chain theory, the research identifies functional and psychological factors that influence programming learning experiences with generative AI tools, including tool ease of use, interactivity, quality of the result, knowledge level, and self-efficacy. Through an analysis of the survey results of 221 university students and fuzzy-set qualitative comparative analysis (fsQCA), the findings reveal two types of generative AI-assisted programming learning enhancement patterns: (1) the dual-driven pattern, which combines functional and psychological consequences; (2) the psychological consequence-dominated pattern. In addition, although the five factors cannot alone constitute the necessary conditions for a learning experience, tool ease of use, interactivity, and knowledge level play an indispensable role in improving the generative AI-assisted learning experience. The findings offer valuable insights for programming educators, deepening their understanding of how generative AI tools influence students’ programming learning experiences and informing the development of more effective strategies for integrating them into computer science education.