The effectiveness of corpus in written revision tasks for English language learners

This study aims at exploring the effectiveness of EFL learners' use of corpus data to review and self-correct lexico-grammatical errors in their writings. The participants (n=34) were divided into an experimental (n=17) and a control (n=17) group. Both groups were intact groups. Adopting a mixed-methods approach, the learners' performance were analysed quantitatively, and their attitudes were explored qualitatively. The participants wrote five compositions followed by a revision session per composition. The categorization of errors is based on surface structure errors taxonomy. To explore the effectiveness of the correction methods, the percentages of accurately corrected mistakes were examined. The differences between different error categories was examined similarly. Finally, a post writing was compared to the error ratio of the first composition to reveal whether there is a long-term effect of the experimental treatment. The results indicated a positive short-and long-term effect for both groups, however the experimental group outperformed the control group. The interviews showed that the learners had a positive attitude towards corpus consultation after their short experience with it.