Online or Face-to-Face Self-Regulation? Technology-Supported or Non-Technology Self-Regulation? A Meta-Analysis Study on the Effect of Self-Regulation on Achievement in English Language Learning

Abstract

Self-regulation refers to the learner's initiative and goal-setting and involves cognition and metacognition, with an interconnection between them. Self-regulated learning environments have been provided through online, face-to-face, and blended learning. The term has been widely studied in educational psychology and language learning. Thus, this paper investigates its effect on English language learning achievement. The meta-analysis method was used to combine the quantitative data extracted from primary studies. Primary studies published up to 2024 were reviewed across the selected databases using predetermined keywords. Inclusion and exclusion criteria yielded 71 effect sizes from 57 primary studies. The heterogeneity test (Q df = 71, 781.913; I 2 = 91.048) indicated a random-effects model. The overall effect size for the REM model was g = 1.189, indicating a large effect. Publication bias was tested with Rosenthal's classic fail-safe N and Orwin's fail-safe N, and no publication bias was detected. Moderator analysis was conducted to identify potential sources of heterogeneity. Thus, research place, publication type, education level, language skills, sample size, treatment duration, the context in which self-regulation was applied (online, face-to-face, or blended), and self-regulation type (technology-supported or non-technology-supported) were identified as possible moderators. All moderators, except language skills, were found to be important, with significant differences in effect sizes. Conducting online self-regulation and technology-supported self-regulation proved more effective, with effect sizes of g = 1.494 (huge) and g = 1.403 (very large), respectively.