Narrative review: Trends of local research on Data-Driven Learning (DDL)

This narrative review aims to delineate current research trends in topical areas, design characteristics, and key findings from empirical studies that investigated the efficacy of Data-Driven Learning (DDL) in local educational contexts over the last decade. To this end, we analyzed 36 academic journal papers, as well as graduate theses and dissertations, using PRISMA guidelines. Major findings indicate a predominance of research on the hybrid area, integrating cognitive and affective aspects. Middle-school and university students with varying English proficiency levels were most represented. Mixed-methods design, involving 11-50 participants over 1-8 sessions, was most common, focusing on vocabulary and grammar learning. Pre- and post-tests were most favored and an indirect approach to DDL dominated using a general corpus. In addition, DDL generally benefited English learners across all proficiency levels, with more learning gains observed within groups compared to between groups. Learners recognized DDL's language learning benefits, showing preferences for group activities. Common issues included challenges in inferring meaning and linguistic patterns, and overly complex example sentences. The review findings provide insights for future research designs and the pedagogical implementation of DDL in local settings.