AMQuestioner: Training Critical Thinking with Question-Driven Interactive Argument Maps in Online Discussion

Critical thinking, which requires logical analyses on the problems and keeping open-minded to others' viewpoints, is a crucial skill when participating in online discussions. While existing works have explored visualizing the components of an argument in a map, i.e., argument map, to support critical thinking tasks, few of them have incorporated educational elements that aim at training critical thinking in online discussion. In this paper, based on a formative study (N = 57), we develop AMQuestioner, a critical thinking training tool that allows question-driven interactions with argument maps automatically extracted from a post thread. In AMQuestioner, users can explore others' claims with a chatbot via suggested questions and conduct critical thinking exercises by answering generated questions related to any claim in the map. A mixed-design study (N=24) reveals that, compared to a baseline tool without question-driven features, participants after training with AMQuestioner demonstrated significantly more improvements in independently writing arguments that are detailed, specific, and relevant to the topic. Participants with AMQuestioner also exhibited a stronger inclination toward open-mindedness to others' arguments during the three-days training process. We discuss design implications for future critical thinking training tools.

CCS Concepts: • Human-centered computing → Human computer interaction (HCI); Collaborative and social computing.