Rather than looking at variables like grammatical feature type, proficiency, and LI as separate, we must start to view them as interrelated and codependent. In this way, educators may finally learn how to maximize the effect of corrective feedback (CF). A meta-analysis of 13 studies was conducted to examine four styles of CF (implicit prompt, implicit reformulation, explicit prompt, and explicit reformulation) along with the following variables: grammatical feature type, L1 similarity (with Mandarin Chinese), and proficiency level. Results suggest that effect size is highly dependent upon the grammatical feature targeted. Explicit feedback appears to have a scaffolding effect for more semantically or syntactically complex grammatical features. Implicit reformulations (recasts) also provide scaffolding, yielding larger effect sizes than implicit prompts when used with Li dissimilar features that are more semantically or syntactically complex.
