The Role of Direct and Indirect Written Corrective Feedback in Improving Iranian EFL Students’ Writing Skill

AbstractThe call for longitudinal evidence on the efficacy of written corrective feedback (WCF) for ESL and EFL writers has been made repeatedly since Truscott (1996) claimed that it is ineffective, harmful, and therefore; should be abandoned. Though controversy continues as to whether error feedback helps L2 student writers to improve the accuracy and overall quality of their writing (Ferris, 1999; Truscott, 1996; Truscott, 1999), most studies on error correction in L2 writing classes have provided evidence that students who receive error feedback from teachers improve in accuracy over time. This study was an attempt to investigate the role of direct and indirect written corrective feedback in improving EFL students’ writing skill. The participants were sixty pre-intermediate students in “Iranians” institute in Ardabil. Twenty students were in direct-feedback group, twenty students were in indirect-feedback group and twenty students were in no-feedback group. The variance of the number of errors of three classes in session one (pretest, appendix 1) was approximately equal, so their ability in this structure was the same. We gave the students TOEFL tests about definite and indefinite articles in five sessions. There were twenty-two tests and forty blanks in each session. The tests were selected from three books titled: TOEFL Grammar Flash, Rahnama TOEFL, and Longman Preparation Course for the TOEFL Test. Treatments were made on experimental groups through giving feedback and correcting their errors. This was fulfilled during five sessions (one session per week.). Two direct-feedback and indirect-feedback groups received the treatment in terms of giving feedback about their errors while the no-feedback group did not receive any kind of feedback. In correcting the papers of direct-feedback group, we gave them direct feedback on their errors. But in correcting the papers of indirect-feedback group we just underlined the errors. And in correcting no-feedback group we corrected their errors but we did not give them their papers. One-way ANOVA was used to analyze the data for three groups. Group differences were considered significant when ρ < .05. The analysis of data is presented in this paper. Because the time interval of tests was the same and the type of testing (fill in the blanks test) also was the same and variances were equal so we analyzed just two phases of pretest (session1) and posttest (session5)