Journal of Higher Education Research

This study addresses the persistent challenges in vocational English writing instruction, such as varying student proficiency, low engagement, and subjective assessments, by developing a data-driven iterative teaching model using the CSMS+ system. Guided by the Flanders Interaction Analysis System (FIAS), Kolb’s experiential learning theory, and the 4-Wh questioning framework, the research explores how real-time multimodal data can enhance teaching methodologies. Conducted over two cycles of action research at a secondary vocational school in China, the study focuses on 36 first-year Early Childhood Care students. Results reveal significant improvements in classroom interaction quality, writing performance, and student participation, particularly through the use of targeted questioning techniques and style-sensitive task designs. Notably, teacher talk time decreased, while complex questioning sequences and student engagement increased. Writing proficiency improved markedly in structural completeness, formality, vocabulary precision, grammatical accuracy, and workplace relevance. These findings underscore the effectiveness of integrating educational analytics with pedagogical theories to achieve precision teaching and foster professional language competence in vocational education.