Formative Assessment Practices as Predictors of Self-Regulated Learning among University EFL Students in Vietnam

Abstract

Formative assessment can support self-regulated learning, yet limited empirical evidence has established how specific formative assessment practices predict different dimensions of self-regulated learning in Vietnamese higher education. This study examined university EFL students’ levels of self-regulated learning and perceptions of formative assessment practices, investigated the relationship between the two constructs, and determined the predictive contributions of teacher feedback, peer feedback, and self-assessment and reflection to overall self-regulated learning and its dimensions. Using a quantitative cross-sectional predictive design, data were collected through a 50-item self-report questionnaire completed by 268 undergraduate students at a public university in Vietnam. Descriptive statistics, Pearson correlation, and multiple linear regression analyses were conducted. Students reported relatively high levels of self-regulated learning and generally positive perceptions of formative assessment. Formative assessment and self-regulated learning were strongly and positively correlated (r = .848, p < .001). The three formative assessment practices jointly explained 72.9% of the variance in overall self-regulated learning. Self-assessment and reflection was the strongest predictor, followed by teacher feedback, while peer feedback made a smaller but statistically significant contribution. The dimension-specific analyses further revealed that different formative assessment practices predicted different aspects of students’ regulation of learning. These findings contribute context-specific evidence that learner-involving assessment is particularly important for developing self-regulated learning in university EFL education. EFL instructors should therefore integrate structured self-assessment, reflective activities, timely teacher feedback, and scaffolded peer feedback. Future research should employ mixed methods and multi-institutional samples to strengthen explanatory depth and generalizability.