Research demonstrating the beneficial effects of corrective feedback (CF) for second language (L2) learning (e,g., Li, 2010) has almost invariably resulted from studies in which CF was provided immediately. Yet teachers are often encouraged to delay CF to avoid interrupting learners (Harmer, 2001). This study investigates how differences in the timing of CF on oral production affect L2 learning and learners’ reactions to CF. Theoretically, Immediate CF may facilitate L2 development by allowing learners to immediately compare their errors to accurate models (i.e., recasting, e.g., Doughty, 2001). The effectiveness of Immediate CF has also been linked to skill acquisition theory because some CF (i.e., prompting) is hypothesized to help learners proceduralize their L2 knowledge (Ranta & Lyster, 2007). This thesis introduces additional theoretical explanations to explain the effectiveness of both Immediate and Delayed CF. For example, reactivation and reconsolidation theory (Nader & Einarsson, 2010) holds that long-term mental representations are susceptible to change when they are recalled. Thus, both Immediate and Delayed CF may help learners alter their incorrect mental representations of language features if that CF reminds learners of those incorrect representations and provides them with accurate models. In a laboratory-based study, 90 intermediate-level adult ESL learners were randomly assigned to Immediate, Delayed, and No CF conditions. Learners took three pre-tests to measure their knowledge of the English passive construction: an aural grammaticality judgment test (AGJT), an oral production test (OPT), and a written error correction test (ECT). Next, they received some brief instruction on the passive. Learners then completed three communicative tasks in which the CF conditions were provided. These tasks were followed by immediate and delayed post-tests. Learners’ reactions to CF were elicited with a questionnaire. Mixed-design one-way ANOVAs revealed statistically significant improvement for all conditions over time on all measures, but no statistically significant differences between conditions. The questionnaires revealed that learners prefer Immediate CF, but that Immediate CF may constrain CF noticeability and learners’ independence, while Delayed CF may cause anxiety or embarrassment. In summary, altering the timing of CF did not differentially affect L2 development, but it did elicit different reactions from learners.