Technology-enhanced and AI-supported instruction: a meta-analysis of student learning outcomes with implications for writing development

Abstract

The increasing integration of artificial intelligence (AI) into education has introduced new paradigms for teaching, learning, and assessment, especially in the development of students’ writing performance. This meta-analysis examines the effectiveness of blended, technology-enhanced, and AI-supported instructional interventions on student learning outcomes. Eleven studies representing 477 participants across primary, secondary, and higher-education settings were synthesized. The included studies represented heterogeneous educational contexts and assessed outcomes such as academic achievement, literacy, vocabulary and language learning, digital literacy, and motivation. Accordingly, the pooled effect represents an overall effect on student learning outcomes rather than a writing-proficiency-specific effect. Studies were identified through ERIC, Scopus, Web of Science, and Google Scholar and represented 477 participants across elementary, secondary, and higher-education settings and diverse disciplinary contexts. The results revealed a large and statistically significant pooled effect size [Hedges’ g = 0.89, 95% CI (0.76, 1.10)], showing that students who participated in technology-enhanced instructional interventions outperformed those receiving traditional instruction. Exploratory moderator analyses showed differences according to educational level, intervention duration, and technology type, with primary and secondary students, medium-length interventions (4–8 weeks), and interactive tools showing the strongest effects. The findings show the potential of thoughtfully integrated digital and AI-supported tools to enhance teaching and learning.