Improving Students' Argumentation Skills Using Dynamic Machine-Learning-Based Modeling

• 成果类型: Article
• 署名作者: Wambsganss, Thiemo; Janson, Andreas; Soellner, Matthias; Koedinge, Ken; Leimeister, Jan Marco
• 署名单位: University of St Gallen; Universitat Kassel; Carnegie Mellon University; Universitat Kassel
• 刊物名称: INFORMATION SYSTEMS RESEARCH
• ISSN/ISSBN: 1047-7047
• DOI: 10.1287/isre.2021.0615
• 发表日期: 2025
• 关键词: SOCIAL COGNITIVE THEORY; information-technology; design science; systems; education; thinking; argue; acceptance; analytics; guidance

摘要

Argumentation is an omnipresent rudiment of daily communication and thinking. The ability to form convincing arguments is not only fundamental to persuading an audience of novel ideas but also plays a major role in strategic decision making, negotiation, and constructive, civil discourse. However, humans often struggle to develop argumentation skills, owing to a lack of individual and instant feedback in their learning process, because providing feedback on the individual argumentation skills of learners is timeconsuming and not scalable if conducted manually by educators. Grounding our research in social cognitive theory, we investigate whether dynamic technology -mediated argumentation modeling improves students' argumentation skills in the short and long term. To do so, we built a dynamic machine -learning (ML)-based modeling system. The system provides learners with dynamic writing feedback opportunities based on logical argumentation errors irrespective of instructor, time, and location. We conducted three empirical studies to test whether dynamic modeling improves persuasive writing performance more so than the benchmarks of scripted argumentation modeling (H1) and adaptive support (H2). Moreover, we assess whether, compared with adaptive support, dynamic argumentation modeling leads to better persuasive writing performance on both complex and simple tasks (H3). Finally, we investigate whether dynamic modeling on repeated argumentation tasks (over three months) leads to better learning in comparison with static modeling and no modeling (H4). Our results show that dynamic behavioral modeling significantly improves learners' objective argumentation skills across domains, outperforming established methods like scripted modeling, adaptive support, and static modeling. The results further indicate that, compared with adaptive support, the effect of the dynamic modeling approach holds across complex (large effect) and simple tasks (medium effect) and supports learners with lower and higher expertise alike. This work provides important empirical findings related to the effects of dynamic modeling and social cognitive theory that inform the design of writing and skill support systems for education. This paper demonstrates that social cognitive theory and dynamic modeling based on ML generalize outside of math and science domains to argumentative writing.