Dynamic assessment and deep learning in elementary writing instruction: A bibliometric study

Abstract

This study conducts a bibliometric analysis to map global research trends on Dynamic Assessment (DA) and deep learning in elementary writing instruction from 2019 to 2025. Using 499 Scopus-indexed documents, the analysis employed VOSviewer to generate network, overlay, and density visualizations based on co-occurring terms. The results show two dominant thematic directions: (1) foundational writing pedagogy reflected in terms such as writing instruction, skill, essay, and child, and (2) rapid growth of technology-driven themes, including deep learning, machine learning, artificial intelligence, and ChatGPT. The cluster and temporal maps reveal that narrative writing now functions as a bridge between reflective learning approaches and AI-supported scaffolding tools. Compared with previous studies that examined DA, deep learning, or writing instruction separately, this research provides the first comprehensive bibliometric evidence of their convergence. The study’s novelty lies in identifying an emerging integrated research ecosystem where DA principles, deep learning strategies, and intelligent technologies collectively shape future directions in writing pedagogy. These findings suggest the need for adaptive, technology-enhanced assessment models that support meaningful and reflective writing development in elementary learners.