Because phonological, semantic, and morphosyntactic characteristics of grammatical features can have a significant impact on form-focused instruction, utilization of different grammatical features to test new language teaching techniques may conflate determinations of efficacy or inefficacy. The purpose of this study was to holistically examine different types of instruction, comparing them with grammatical features to evaluate effectiveness. Forty-six experimental studies of form-focused instruction were selected for study. Comparison of effect sizes suggests that the efficacy of form-focused instruction differs considerably based upon the type of grammatical feature targeted. Input-based instruction (e.g.,input enhancement or explicit rule presentation) appears more useful for features like the plural -s, past -ed, and third person singular -s, which are phonologically insalient, yet morphologically regular. Output-based instruction (e.g., corrective feedback or recasts), in contrast, appears more effective with grammatical features such as questions, phrasal verbs, conditionals, and articles, which are syntactically or semantically complex. Overall, the results suggest that differences in grammar be considered before curricula or pedagogical interventions are designed.