Chatbot-based training on logical fallacy in EFL argumentative writing

Abstract

ABSTRACT EFL learners generally have the problem of logical fallacies in EFL argumentative writings. Logical fallacies are errors in reasoning that can undermine EFL argumentative writing quality. Explicit training on logical fallacies may help learners deal with the problem and enhance their self-efficacy and proficiency in EFL argumentative writing, which educational chatbots may further enhance. However, few studies have been conducted in this direction. To fill the gap, we developed a chatbot for out-of-class, self-regulated training on logical fallacies in EFL argumentative writing. Fifteen Chinese EFL undergraduate and graduate students engaged in the training for five weeks. Semi-structured interviews and transcript analysis were conducted to investigate the participants’ perceived impacts of the training on self-efficacy and proficiency in EFL argumentative writing. To triangulate the interview results, we conducted pre–post questionnaires on the participants’ writing self-efficacy and analysed their pre–post argumentative writings based on the Illinois Critical Thinking Essay Scoring Rubric. The results showed that the training on logical fallacies might improve EFL argumentative writing proficiency, although it might reduce writing self-efficacy. Based on the results, we called for more implementation and investigation of chatbot-based training on logical fallacies, especially the long-term training integrated with practice in writing tasks.