In Frontiers in psychology ; h5-index 92.0
Group discussion is a common and important form of learning. The effectiveness of group discussion could be facilitated by the adaptive support of virtual agent. Argumentative knowledge construction is beneficial to learners' acquisition of knowledge, but the effectiveness of argumentative scaffolding is not consistent in existing studies. In this study, a total of 64 college students (32 groups, two participants and one computer agent in each group) participated in the experiment and they were assigned to the experimental condition (16 groups) and the control condition (16 groups). In the control condition, the computer agent would give an idea from semantically different categories according to the automatic categorization of the current discussion. In the experimental condition, the computer agent provided argumentative scaffolding after giving diverse ideas to support participants' deep processing. The argumentative scaffolding included two prompt questions, "do you agree with me?" and "could you give the reasons to support your viewpoint?." The dependent variables were the interaction quality, network centrality, the breadth and depth of discussion, the self-reported of discussion effectiveness and the degree of change before and after the discussion. Findings revealed that compared with the control condition, the participants were more likely to discuss the keywords provided by the virtual agent and reported more comprehensive understanding of the discussion topic, but surveyed less ideas and interactions during the discussion under the argumentative condition. This study suggests that the argumentative scaffolding may have both positive and negative effect on the group discussion and it's necessary to make a choice.
Gao Hongli, Xu Sheng, Yang Lei, Hu Xiangen
2022
argumentation, argumentative knowledge construction, artificial intelligence in education, cognitive diversity, group discussion, intelligent tutoring system
