A Physically Embodied Robot Teacher (PERT) as a Facilitator for Peer Learning

Abstract

This research full paper describes our work on the development, and evaluation of a physically-embodied robot teacher (PERT) as a facilitator for Peer Instruction (PI) in a high school science classroom. Embodied pedagogical agents are particularly useful in teaching young learners because of the engagement, and enjoyment they engender. Teaching robots are among possibilities in the future classroom; they may also represent solutions to the looming, acute global teacher scarcity. However, their design creates unique computational challenges. Moreover, robot-human systems involving a single robot in interaction with multiple humans, are not common. PI has been reported in many studies as a learner-centered approach that supports conceptual learning, metacognition, cognitive load sharing, and higher order learning, through peer scaffolding. However, existing studies have reported only applications by human teachers in regular classrooms. This study highlights an approach for enabling a PERT as a PI facilitator. Based on conflicting reports of the significance of agent embodiment for learning, we present the report of two experimental studies from our assessment of the system, employing the mixed method approach. We compared participants' perception of the robot's social presence and teaching quality when the robot is physically present and when it is absent (voice-only). Our findings are based on data from 14 participants in each study. The Social Presence indeX (SPiX) and the teacher quality index (TQI) were used for quantitative data collection for social presence and teaching quality, respectively. Qualitative data were based on live observation and participants' open-ended feedback. Quantitative findings show no significant difference in participants' perceptions in both conditions but qualitative feedback provided more information. Our findings show: i) students are willing to accept non-human teachers, and ii) their expectations of them are similar to those for human teachers; iii) an improved affective system will promote increased presence; iv) participants noted the robot as `neutral', with important implications for culturally-responsive education; v) robot teachers could be potential technology-inspired solution to global teacher scarcity and vi) in addition to promoting deeper learning, PI can be a useful pedagogical approach for non-human teachers. This study is in line with anticipated ability of future technological advancements to support machines that are as effective as intelligent human tutors.