Abstract The current study is an attempt to investigate the effect of written corrective feedback on improving grammatical accuracy of EFL learners at different proficiency levels. It also aims to explore the effect of corrective feedback on learners’ output. Two classes of low proficient students (N:10) and two classes of high proficient ones (N:9) were selected based on convenience sampling. The participants were administered a pre-test, two treatments and a post-test, were given to all the participants in four separated sessions. Since the study does not involve random selection of the participants, a quasi -experimental design was employed. As for the pre-test, the participants were asked to describe a picture in their own words and then complete a multiple-choice test and a cloze-test. The second and fourth sessions involved two treatments which had the same formats as in the pre-test, that is, picture description task, multiple-choice tests and cloze task. After each treatment session, they were given written corrective feedback which was explained to them in details. The post-test which was the same as the pre-test, was administered in the last session. In the current study, there were five research questions whose data analysis procedure was based on quantitative statistics, and data management and analysis were performed using SPSS (version 26), (2019). It should be mentioned that the entire analysis procedure of this study was limited to nonparametric statistics since the number of the participants resulted in rejection of parametric assumptions such as normality of distribution. Findings indicated that the high-proficient learners did not seem to experience a significant change with the corrective feedbacks. On the other hand, the low-proficient learners’ scores turned to be significant in both variables (gain scores from pre-test to post-test and improvement in the accuracy rate). Besides the verb form, the participants’ writings were analyzed based on other criteria of linguistic features including capitalization, spelling, word formation, subject/verb agreement, verb tense, wrong word, word order, plural/singular, missing words. Through such detailed checklist, the scores of the learners were investigated with more details and potential areas of changes were detected with more accuracy. It was also found that the linguistic features, which were significantly improved in low-proficient learners, were capitalization, verb tense, and missing words
