Automated Writing Evaluation Feedback in English Learning and Teaching

Abstract

Writing is a fundamental skill in an individual’s life, serving as a medium to express opinions, record important information, and create imaginative worlds. For students, writing is more significant because it is a necessity to achieve academic success no matter what subjects they are studying (Graham, 2018). However, writing is a complex task because it requires students to master both basic cognitive processes such as handwriting or spelling, and complex cognitive processes such as idea generation, transformation of ideas, and writing modification (Graham & Perin, 2007). Moreover, they also need sufficient practice to gain proficiency (Burstein et al., 2020). These abilities do not develop naturally; thus, schools play a vital role in nurturing students with these abilities by offering adequate practice and instruction in writing (Graham, 2018, 2019). An important instructional practice in the classroom is providing writing feedback. By showing the gap between students’ current capabilities and expected outcomes, feedback can help students improve their writing quality (Biber et al., 2011; Graham et al., 2015).

However, providing real-time, detailed, and cross-subject feedback at scale poses a challenging task for educators (Burstein et al., 2020). In this case, Automated Writing Evaluation (AWE), defined as the capability of a computer technology that employs artificial intelligence, to evaluate and score written text (Shermis & Hammer, 2013), can serve as a viable choice, given its ability to offer immediate feedback to students (Lee, 2020).

This paper illustrates the affordances and constraints of AWE by reviewing findings from research studies, and discusses its application in instructional settings. Furthermore, it presents a comparison between instructors’ feedback and AWE feedback, before providing practical implications for teaching English for Academic Purposes (EAP).