Robot-assisted language learning involves the use of robots to aid students in learning and expressing language. In this educational approach, robots engage in interactive social behaviors. A recent systematic review and meta-analysis by Wang and Cheung investigated different types of robots’ social behaviors in language learning and their impacts. The review synthesized 59 empirical studies to explore how different social behaviors exhibited by robots have contributed to language learning and in what contexts.
Results showed that robots’ social behaviors have predominantly been observed in contexts involving fewer than 80 K-12 students, for learning English vocabulary within a single session. In the process, robots interacted with students socially via non-verbal and verbal behaviors. Non-verbal behaviors consisted of physical appearance, gesture, posture, face and eye behaviors, and environment and space. Verbal behaviors included vocal behaviors, verbal immediacy, verbal feedback, and verbal adaptivity. These social behaviors conveyed social signals that indicated robots’ roles, instructional support, and a friendly real-life learning atmosphere.
Regarding the effects of robots’ social supportive behaviors versus neutral behaviors, this study found a positive overall effect size (g = +0.27). Specifically, when robots played the role of tutors rather than peers, robots’ social supportive behaviors produced better language learning. In addition, anthropomorphic robots showed a slightly higher effect than cartoon-like robots. Mixed behaviors had the largest effect, followed by non-verbal behaviors and verbal behaviors. This study provides insights into future research and practice in the field of robot-assisted language learning.
