Whether error correction benefits L2 language writing has provoked much controversy and generated a multitude of research with conflicting results. As previous research has not yet reached a consensus on the said issue, more empirical studies are still needed, particularly in the EFL context where error correction is much more heavily used than in the ESL context. Engaging Taiwanese college students in an 18-week experimental instruction, this study aimed to investigate whether error correction could lead to improved quality and accuracy in students’ new pieces of writing, whether such an improvement would be affected by different types of errors, and how students of varied writing proficiency benefit from and respond to error correction. The findings show that (1) error correction is effective in promoting the overall quality of student essays, particularly for intermediate to high proficiency students; (2) error correction cannot improve the accuracy of student writing except for high proficiency students; (3) verb-related errors are more susceptible to error treatment than other types of errors; and (4) high proficiency students responded more favorably to teacher corrective feedback than low to intermediate achievers. Based on the results, this study suggests that individual oral feedback is needed to provide students with clear directions for essay improvement, especially for beginning writers or students who are placed at low to intermediate levels.