Hybrid Feedback: The Efficacy of Combining Automated and Teacher Feedback for Second Language Academic Writing Development

While researchers and practitioners agree that students need more writing practice and feedback, overburdened teachers often do not have sufficient time to read, mark, and give feedback on students' multiple drafts. Automatic Writing Evaluation (AWE) systems have emerged as a possible solution to give immediate feedback to writers. However, AWE systems lack individualized feedback and feedback on content and can diminish the social and communicative dimensions of writing. Thus, some researchers have advocated that AWE should be construed primarily as a complement to, rather than a replacement of, teacher feedback.

Currently, there is a lack of research on the effectiveness of hybrid feedback, or the combination of teacher and AWE feedback, in the academic writing classroom for supporting the development of second language writing. The current study has started to address this gap by examining if hybrid feedback resulted in differences in approaches to writing, language, content, and organization of writing between a class that received hybrid feedback and a class that received only teacher feedback. A mixed methods design first collected quantitative data and then augmented the quantitative results with in-depth qualitative data. First, pre, post and delayed post-treatment writing tasks were administered to both groups to compare writing in terms of scores and various fine-grained writing indices. A questionnaire on changes in cognitive processes was conducted for both groups, and questionnaire data on the perception of AWE was collected from the experimental group. Second, a focus group interview was conducted as a follow-up to the quantitative stage from the experimental group. A mixed MANOVA comparing changes between and within groups was used to analyze the questionnaire data and changes in writing, and thematic analysis was used to interpret the qualitative data.

The findings suggest that although AWE feedback has limitations, including insensitivity to context, learner needs, meaning, and inability to provide dialogic feedback, combining it with teacher feedback may address some of its limitations, help motivate students to revise and write more often, facilitate autonomous learning, and reduce teachers' workload.