The Impact of Computer-Based Feedback Strategies on Students' Learning: A Network Meta-Analysis

Résumé En

Feedback to learners is considered as a crucial component of successful learning. Elaborated feedback (EF) and Answer-Until-Correct feedback (AUC) have been identified among the most effective for learning in a computer-based environment. Based on 46 studies with samples from 11 countries, the current study used frequentist network meta-analysis to investigate the effectiveness of different EF and AUC strategies on students’ learning outcomes among high school, university, and adult students according to their prior knowledge level. We used 117 effects size to compare the following strategies: EF with explanation (EFE), EF with hints, guidance or cues (EFH), EF with worked examples (EFW), AUC with knowledge of correct response (AUCK), AUC with hints or guidance (AUCH), and AUC with explanations (AUCE). Results showed (1) all types investigated feedback produced positive significant effects on students’ learning outcomes except EFH, (2) AUCE seemed to be the most effective feedback strategy for enhancing students’ learning, (3) EFE is the only types of feedback that produced a positive significant effect on low prior knowledge students’ performance, while (4) all types investigated feedback produced positive significant effects on high prior knowledge students’ performance except EFH, and (5) the effects of feedback vary according to the investigated moderators: learning order level, learning domain, and test format. These findings suggest more studies on feedback strategy with explanation (EFE and AUCE) among students needed to be conducted.