Evaluating Argumentation Effectiveness Using DeBERTa Token Classification Approach

• In Event: Situational Judgment Tests

Abstract

This study implements DeBERTa model for evaluating the quality of argument elements in Situational Judgment Tests. Utilizing transfer learning techniques, the model demonstrated high accuracy and agreement in assessing the quality key argumentative elements like "Position" and "Claim." The approach has significant implications for enhancing the evaluation processes of SJTs.

Authors

• Lingchen Kong, University of Florida

• Alexander MacIntosh, Acuity Insights

• Jinnie Shin