How effective is feedback for L1, L2, and FL learners’ writing? A meta-analysis

Background

Feedback is important for writing improvement, but research suggests that not all learners benefit equally from feedback.

Aims

This meta-analysis examined if different feedback treatments given by different feedback agents are differentially effective for first (L1), second (L2), and foreign language (FL) learners at secondary schools and universities.

Method

Our meta-analysis included studies with an experimental or quasi-experimental design. We computed effect sizes for surface- and deep-level writing outcomes.

Results

Across 200 comparisons, surface-level feedback significantly improved surface-level outcomes (g = 0.58), with higher effects for FL (g = 0.69) than for L2 learners (g = 0.34). However, such feedback may have detrimental effects on FL deep-level outcomes (g = −0.23). In contrast, deep-level feedback demonstrated positive effects on deep-level outcomes (g = 0.80), with larger effects for L1 (g = 1.26) than for FL learners (g = 0.37; non-significant). Combined surface-and deep-level feedback positively affected both deep-level (g = 0.54) and surface-level outcomes (g = 0.36) for all learners.

Instructor feedback demonstrated overall moderate benefits for both outcomes showing greater impact on FL learners (g = 0.72) than on L2 learners (g = 0.35) for surface-level outcomes. Peer feedback showed medium to large effects for deep-level outcomes (FL learners: g = 0.76; L1 learners: g = 1.46). In contrast, algorithm-based feedback showed a non-significant medium effect on FL learners' surface-level outcomes (g = 0.53), and self-feedback showed a non-significant medium effect on FL learners’ deep-level outcomes (g = 0.55). Lastly, feedback proved effective for both university and secondary school students, thereby underscoring its crucial role in fostering writing.