Flow in ChatGPT-based logic learning and its influences on logic and self-efficacy in English argumentative writing
Abstract
Flow is a state of full engagement in an activity. Learning environments featured by Skill-challenge balance, Clear goal, Feedback, and Playability — collectively known as flow antecedents – can induce flow experiences and improve learning outcomes. ChatGPT-based environment seems to encourage a flow in learners: By customising tasks to match students' abilities, aligning materials with clear objectives, providing instant feedback, and ensuring ease of use, ChatGPT can help learners enter a flow state, which, in turn, leads to improved learning. However, there hasn't been much research on flow in ChatGPT-based learning. To bridge the gap, we developed a ChatGPT-based environment for developing logic in English argumentative writing. We studied 40 Chinese university English-as-a-foreign-language (EFL) students in the learning using questionnaires, eye-tracking data, knowledge tests, essay writing tasks, and semi-structured interviews to understand how they experienced flow and how it affected their learning. Our findings showed that the ChatGPT-based environment strongly supports flow antecedents. Skill-challenge balance and Playability were particularly influential for inducing flow experiences. Students who experienced a deeper flow showed better understanding of argumentative writing logic, although their writing self-efficacy became lower. Drawing from the findings, our study highlights how AI like ChatGPT can influence experiences and outcomes of logic learning and language learning, which may be applicable across various domains and disciplines.
Link to publication in Science Direct
