This article reports a quasi-experimental study of how error labeling in remediation exercises affects students' writing performance. Students in five sections of a course in written business communication composed the control group, whereas students in two sections composed the treatment group. On the first letter each group wrote early in the semester--before the treatment group began error labeling in remediation exercises--no significant differences were found in the writing performance of both groups. However, on their last letter written late in the semester, students in the treatment group significantly reduced the number of sentence-level errors they made, compared to students in the control group. The improved writing performance of the treatment group is attributed to error labeling in conjunction with error correction in the remediation exercises. (Contains 1 figure and 2 tables.)