Educational chatbot research: text mining and bibliometrics

Chatbots have received increasing attention in education due to their ability to converse in natural language and provide personalised support. The present work provided a comprehensive overview of this area, focusing on (1) publication trends, publication sources, countries/regions, and institutions; (2) scientific collaborations; and (3) research themes and their evolution. To achieve this, we combined bibliometric indicators, social network analysis, and topic model methodologies to analyse 536 papers published between 2000 and 2023 in the Web of Science. First, the findings reveal an increasing and continuous focus on educational chatbot research, with the International Conference on Artificial Intelligence in Education and the United States as the leading source and country, respectively. Second, countries/regions and institutions engaging in research collaborations were more productive. Third, topic analysis highlights the popular applications of chatbots in language, healthcare, and mathematics education for promoting engagement; supporting metacognitive, collaborative, and affective learning; and reducing learning stress. Finally, we highlight the implications and suggest directions for future research on educational chatbots based on the findings. This work is expected to facilitate researchers’ and practitioners’ understanding of educational chatbot research and raise their awareness of the various research frontiers and possible future directions for said research.