This research is taken from a classroom-based study on how the indirect approach in written corrective feedback (CF) differs from the direct approach in facilitating the learning of grammatical features in L2 writing. Eleven university students from a writing class were divided into three groups based on TOEIC scores. After writing narrative assignments, they were given three different types of feedback: direct, indirect, and metalinguistic. The participants were asked to revise their writings after receiving indirect CF and metalinguistic CF. Immediately after revising, they were also asked to answer an open-ended questionnaire on the feedback type. Their revisions and corrections were analyzed according to use of indefinite articles, and different feedback types showed slightly different reactions and attitudes. The highlevel students were not influenced much by the feedback type, while the mid- and low-level students favored the indirect approach for their thinking processes, and benefited from metalinguistic CF in the revision sessions. The indirect approach appeared to be more effective, if implemented appropriately, increasing working memory for self-regulated L2 writing processes.