Effectiveness and moderating factors of computer-mediated feedback in L2 writing: A meta-analysis

Computer-mediated feedback (CMF) for writing has garnered growing attention among practitioners and researchers. However, comprehensive meta-analyses in English as a foreign language (EFL) and English as second language (ESL) contexts remain scarce. This study synthesized 19 valid effect sizes from 26 experimental studies conducted between 2021 and 2024 to evaluate the effectiveness of CMF in L2 learners’ writing performance and determine the moderating factors influencing its variability. The results revealed CMF large overall effect (g = 1.602) with significant variability across studies. A moderator analysis identified meaningful differences based on feedback source, feedback timing, learner proficiency, and task characteristics. Automated and immediate feedback demonstrated the strongest impact (g = 0.937 and g = 0.875) highlighting their importance in facilitating real-time corrections. Intermediate learners benefited the most (g = 0.789) while advanced and beginner learners showed comparatively smaller effects. Greater improvements were observed in academic writing tasks other than argumentative writing tasks. Additionally, integrating educational software alongside CMF implementation at the university level appeared to enhance writing proficiency. These findings suggest that CMF significantly contributes to L2 writing development, particularly in structured academic environments.