The present study compared the efficacy of computer-mediated peer corrective feedback (CF) among two groups of high-proficiency (n=11) and low-proficiency (n=19) postsecondary Japanese EFL students. They completed a 10-week process writing composition, during which the students were given explicit instruction in metalinguistic peer CF. They then received CF from their peers and instructors. The researchers conducted a quantitative analysis of compositions, and the students' perceptions of the process were elicited using qualitative surveys. The results corroborate previous studies which found low-proficiency learners lack the aptitude to engage in meaningful peer CF. However, the qualitative portion of the study found that both proficiency groups had largely positive views of the peer CF process and revealed some interesting distinctions between the groups.