Promoting accuracy of collocation use in L2 writing: the role of data-driven learning in indirect corrective feedback

Abstract

This study investigates the value of integrating data-driven learning (DDL) into indirect corrective feedback (CF) as a revising tool, with a focus on promoting accuracy in using collocations in the second language (L2). A cohort of 118 participants was divided into three groups, namely the indirect CF group supported by DDL, the indirect CF group supported by a dictionary, and the no-CF control group. The experiment ran for 16 wk, and during this time, the DDL and dictionary groups wrote 15 argumentative essays, received indirect feedback on collocation errors, and revised their errors with the help of a corpus or dictionary. The control group carried out the same 15 essay tasks without CF. All groups undertook a pretest, posttest, and delayed posttest. The results revealed that DDL coupled with indirect feedback was effective in promoting accuracy in collocation use over a sustained period of time. Additionally, the study found that corpus queries were superior to dictionaries in resolving collocation errors and reducing error rates. The findings suggest that using DDL in combination with indirect feedback is an effective tool for resolving L2 collocation errors.