Research on the Computational Thinking Microgenesis Based on Generative Advanced Prompts

Generative artificial intelligence (GenAI) can provide personalized real-time feedback for programming learning, while prompt engineering can standardize GenAI-generated feedback and foster the development of computational thinking (CT) development. However, there is a lack of research on generative advanced prompts that integrate both general prompts and advanced prompts at present. Accordingly, this paper constructs a generative advanced prompt model and conducts a quasi-experimental study supported by generative advanced prompts. Through holistic analysis of changes in CT and microgenetic analysis of CT among learners in different clusters, the results show that generative advanced prompts can improve learners’ CT. Based on the paths, rates, and source characteristics of CT microgenesis, the experimental group can be classified into two clusters of the thinking collaboration type and the task agency type. Learners from different clusters exhibit distinct microgenetic characteristics of CT. Specifically, for thinking collaboration learners, their CT evolves along a well-ordered sequential paths with a steadily increasing change rate; while for task agency learners, their CT proceeds alongside repetitive fixation cycles with fluctuating and stagnant change rates. This paper reveals the influencing mechanism of generative advanced prompts on CT from a microgenetic perspective and can provide theoretical guidance for the practice of programming teaching empowered by GenAI.