Exploring Primary School Students’ Computational Thinking Process in an AI-Enhanced STEAM Course: A Lag Sequential Analysis Approach

STEAM education requires learners to be able to utilize interdisciplinary knowledge to draw wisdom from different disciplines to solve complex real-world problems, from which computational thinking capacity is necessary for decomposing problems, abstracting problems and designing algorithms. The personalized support capabilities of artificial intelligence can facilitate effective STEAM courses. Analyzing students' computational thinking behaviors in AI-enhanced STEAM classrooms can help teachers understand patterns of learning behavior, diagnose potential challenges in fostering computational thinking, and adapt instructional strategies to improve learning outcomes. In this study, the computational thinking behavioral patterns of teachers and students in an AI-enhanced school-based STEAM course was developed for a voluntary teaching program for an under-represented primary school was explored. This course lasted for six weeks and a total of 798 behavioral samples were collected from students and teachers. Three behavioral sequences, which were decompose problems, design algorithms, and summarize, were derived from lagged sequence analysis.