Effects of Robots' Revoicing on Preparation for Future Learning

Abstract

In the method of preparation for future learning, learners often engage in constructive interaction with expressing, listening to and integrating their own multiple voices. In order to identify a specific discourse structure underlying successful collaboration, we used a remotely controllable robot as a member of a small discussion group of college students who solved a challenging physics problem. For the robot to act as a listener who solicits voices from students in the group, we manipulated its ways of "revoicing": it performed minimum revoicing of students' keywords without evaluative comments in one condition and guiding revoicing towards scientific models in the other condition. Comparing these two conditions in addition to a human-only condition, we found that the robot's minimum revoicing fostered students' agency and reflection on their mental models, which prepared them to learn from a lecture and solve a transfer problem. The role of listener for PFL was discussed.