The positive effects of dynamic written corrective feedback (DWCF) on linguistic accuracy are well-documented (Evans et al., 2010). However, studies on DWCF without exception have adopted a pretest--posttest research design; therefore, they were unable to explore the dynamics of development (Larsen-Freeman, 2006). In addition, all previous DWCF studies exclusively provided indirect feedback to students. Consequently, our knowledge is limited as to whether a modified version (providing direct feedback) of DWCF would be effective. In the present study, 24 university undergraduate students composed a total of 288 essays and received modified DWCF (direct feedback) on a weekly basis in two advanced writing courses over 12 weeks. The essays were analyzed by applying both pretest-posttest and time-series analyses. Statistically significant differences were found in the linguistic accuracy indices (errors per words, errors-free clause ratio) in the student's data between the pretest and posttest. The time-series analysis showed the dynamics of the development of linguistic accuracy. This study showed that a modified version of DWCF is also effective and provides deeper insight into the dynamic processes of linguistic accuracy development.