Compositional Analysis of Writing Instructional Practices Performed with Digital Tools: A Meta-Analysis

Abstract

Meta-analyses have consistently shown that ICT-supported writing interventions are effective, yet existing syntheses often provide limited insight into their instructional design. To address this gap, the present meta-analysis adopts a component-based instructional design perspective to examine which instructional features embedded in ICT-supported writing interventions are associated with improvements in students’ writing quality. We estimated the overall effect on writing quality (1) and tested whether design principles, their operationalization into specific instructional activities (2), feedback and scaffolding features (3), and the delivering agent (teacher, ICT, or both) (4) moderated effects. Following PRISMA guidelines, studies were identified through ERIC, PsycINFO, Web of Science, and Scopus, yielding 42 empirical studies, focusing on typically developing students. ICT-supported interventions showed a large overall positive effect on writing quality (g = 1.264, 95% CI [0.957, 1.570], t(39.93) = 8.33, p <.001). Contextual variables (educational level, language context, study design, duration, and intensity) did not significantly moderate effects. While instructional design principles did not emerge as significant moderators, their concrete operationalization through instructional activities did: written presentation of objectives and direct instruction were linked to stronger effects. Dynamic (adaptive) scaffolding was also linked to stronger effects, and example-based activities yielded larger effects when delivered by both ICT and teachers. Overall, the results suggest that analyzing ICT-supported writing interventions at a fine-grained, componential level provides clearer insight into how instructional design shapes learning outcomes and can inform the development of more effective tools.