An Intelligent Learning Environment for Case-Based Argumentation
A number of intelligent tutoring systems (ITSs) have been developed that engage students in argument exchanges, but few of them have focused on teaching genuinely complex argumentation skills. The current research focuses on the hypothesis that such skills can be taught by an ITS that uses an AI model of case-based argumentation to generate examples dynamically. To test this hypothesis, we developed CATO, an intelligent learning environment that supports reasoning tasks that involve past cases, including an induction task and a written argumentation task. CATO generates argumentation examples at the students’ request while reducing some of the distracting complexity in the students’ task. In a controlled experiment, which took place in the context of a first-year legal writing course, CATO’s component-wise, example-based instructional approach was compared to small-group instruction led by an experienced instructor. Both approaches were about equally effective in teaching basic argumentation skills. However, on a far transfer test, for which students wrote advanced legal memos, it was revealed that the component-wise example-based approach was less effective in preparing students to apply basic skills in complex contexts. Thus, while examples can be effective in argumentation instruction, it is important to use them in an “integrated” instructional approach, where complexity is present from the start.