INTERACTIONAL COMPUTER-ASSISTED FEEDBACK: ENHANCING LEARNERS' COHESIVENESS AND AUTONOMOUS CORRECTION IN ENGLISH AS A FOREIGN LANGUAGE WRITING CLASSES
English as a Foreign Language (EFL) learning setting is influenced by many crucial factors such as comprehensible input, age, explicit/implicit teaching strategy, small class size, and learning style. An interactional teaching-learning methodology is another essential factor assisting both teachers (Ts) and learners (Ls) to reach the intended learning outcomes, especially in the EFL writing classes. However, such classroom interaction is hindered by the large-sized and mixed-abilities EFL classes. Accordingly, an open-source Learning Management System (LMS) is a viable solution that paves the way for Computer-assisted Corrective Feedback (CCF) which, in return, caters for online Ts-Ls interaction between and learner autonomy. The purpose of this quasi-experimental study was to find the effect of explicit, interactional CCF on the Ls' autonomous correction of errors related to the target features of lexical and grammatical cohesive markers. During the one-semester study and via the open-source LMS Canvas, forty Egyptian freshman and sophomore Ls received online, explicit, and interactional CCF for the first and final drafts of their required academic writings to find its effect on their autonomous correction of the target features. The results of the pre-, immediate, and delayed post-tests assessing the Ls' cohesive EFL writing before, during, and after the treatment were analyzed using a paired sample t-test. The mean score in the pretest (m = 2.6, s = 0.83781) was less than its counterpart in both the immediate posttest (m = 4.4, s = 1.27877) and the delayed posttest (m = 5.5, s = 0.93233). Furthermore, the paired sample t-test revealed how this type of CCF assisted the Ls to cognitively process and correct any persisting errors in their EFL writing t (39) =17.664, p = 0.000. Furthermore, the Ls' attitude towards the usefulness of this type of feedback was measured using a one to five Likert scale questionnaire. The Ls' responses demonstrated how 50% found it extremely useful, 35% very useful, 5% somewhat useful, and 10% not useful at all. Findings of both the paired sample t-test and the questionnaire indicate how explicit, interactional CCF fosters an interactional learning environment and, accordingly, facilitates Ls' autonomous error correction of the target features. However, more research is needed to develop open-source LMSs that cater for time-saving CCF and the tailored objectives of EFL writing classes.