Factor-Based Argumentation Evaluation

Abstract

In the argumentation exercises of civil procedure law, ADR law [ADR Law 01, Kawamura 02], mock trial, and mock mediation [Kobayashi 98, Tanaka 05b], we have tried to be objective and clear about the superiority of the argument, but we have been conscious of the educational problem that the evaluations are inevitably based on the subjective judgment of the teachers. In more detail, this means that (1) the evaluation items are unclear, and (2) the method of assigning evaluation points is subjective. To cope with the problem of unclear evaluation items, we came to know the existence of the Intercollegiate Negotiation Competition (hereinafter referred to as “the competition”). The Intercollegiate Negotiation Competition [INC 05] is an intercollegiate negotiation and arbitration competition established in 2002 to increase public interest in negotiation and debate and to provide students with incentives to learn negotiation and debate [INC 05]. More than 100 people, including Japanese and foreign legal professionals, people from a wide variety of companies, and university faculty members, cooperate as judges in this negotiation competition every year. In the negotiation competition, the judges use a judgment chart to compare and evaluate multiple arguments. In the negotiation competition, the judges use a judgment chart to compare and evaluate multiple arguments. The judges use a set of evaluation items for argumentation skills, and the arguments are scored based on the total of the judges’ scores. The competition is evaluated comprehensively by combining the two arguments of Round A (arbitration = competitive) and Round B (negotiation = cooperative). However, the evaluation of each argument is left up to the scorers, and it is unavoidable that different scorers will give different scores to the same argument. If we can extract information from argumentation records that can be used to evaluate argumentation skills, we can not only support teachers in evaluating argumentations but also develop argumentation agents that can conduct advanced argumentations using these indicators.