Effect of AI‐Based Learning on Students' Computational Thinking Development: Evidence From a Meta‐Analysis
Gexing Cheng, De-Xin Hu · Computer Applications in Engineering Education · 2025
ABSTRACT AI‐based learning (AIBL) is becoming more and more popular. Is it effective for the development of students' computational thinking (CT)? This meta‐analysis explores the effects of AIBL on students' CT development based on 26 high‐quality experimental articles. The results suggest that AIBL has an upper‐medium positive effect on students' CT development (Hedges's g = 0.553, 95% CI [0.410, 0.708], z = 7.366, p < 0.001), indicating that AIBL can effectively promote students' CT development. Moreover, moderator analyses reveal that AIBL is more effective under the following conditions: (1) for AI intervention type that applies AI algorithms; (2) in earlier publication years. (3) among European students; (4) among senior secondary students; (5) with sample sizes between 30 and 50 students; (6) for interventions lasting more than 2 months or less than 1 week; (7) in traditional programming courses; (8) when using project‐based design; (9) for individual learning; (10) when assessed by test‐based measurement tools.