The current study focuses on the issue of error correction in L2 writing, specifically whether coded vs. uncoded feedback would affect learners' ability to self-correct on a two-draft composition. In two classroom sessions, 19 intermediate/advanced learners of Spanish as an FL were given 30 minutes to write compositions which narrated the events depicted in two sets of drawings. In the following classroom sessions, subjects were given 20 minutes to correct their errors. On the first composition, errors were indicated by underlining and on the second, errors were indicated and coded according to a list of codes provided to subjects, representing syntactic, lexical, and mechanical errors. Paired-samples t-tests revealed significant differences between the first and second drafts of the compositions in both conditions, but also revealed significant differences between the second drafts of the coded and uncoded compositions, suggesting that feedback is important in improving learners' performance on a second draft, but also that the type of feedback does indeed influence learners' ability to self-correct. Written questionnaire data revealed learners' perceptions regarding ease of correction with and preference for coded feedback.