Comparing Hong Kong secondary school students’ perceptions of ChatGPT-assisted EFL writing

This quantitative study investigates ChatGPT for English as a foreign language (EFL) writing, exploring 107 Hong Kong secondary school students’ perceptions. It compares three teaching approaches—process-based, genre-based, and prompt engineering only (no explicit writing approach)—assessing their impact on motivation, cognitive load, and satisfaction across academic achievement levels. The results show explicit writing approaches significantly enhanced motivation and reduced cognitive load compared to prompt engineering alone. Genre-based instruction proved superior to process-based, leading to significantly higher motivation and satisfaction, alongside lower cognitive load. Students from higher-achieving schools within explicit writing instruction groups demonstrated greater motivational improvements, lower cognitive load and higher satisfaction. From a cognitive load theory perspective, the structured schemas inherent in a genre-based approach effectively reduce cognitive burden, particularly benefiting students with stronger language proficiency. The study highlights the need for differentiated instruction, with genre-based approaches showing particular promise for ChatGPT-assisted EFL writing, especially when carefully scaffolded to support diverse learners.