Peer assessment is useful for the improvement of English writing. However, individual differences exist between students and might affect how they assess peers' works. Hence, individual differences need to be considered in peer assessment. Among various individual differences, cognitive styles have essential impacts on student learning. Thus, this research investigated assessment differences between teachers and students in the context of peer response from a cognitive style perspective. The results suggest that there were no assessment differences in the first draft but assessment differences did exist in the revised drafts. This might be because students might feel difficult to identify improvement due to their insufficient experience. On the other hand, Holists usually paid more attention to the content at a superficial level while Serialists had difficulties in the connection between topics and the content. In brief, the findings from this study demonstrated that each student took a unique way to do assessment. Furthermore, such findings can guide instructors to deliver effective peer assessment in the future.