Global Trends in Flipped Learning Research: A Bibliometric and Thematic Analysis

In the last decade of accelerated digital transformation, the Flipped Learning Model (FLM), as a technology-supported and student-centered pedagogical model, has become increasingly visible in instructional designs, but the research accumulation in this field is fragmented. This study aims to map the scope, temporal development and conceptual structure of academic production in the field of FLM from a holistic perspective. To this end, as a result of the search conducted in the Web of Science and Scopus databases in the subject area of "flipped learning" OR "flipped classroom", a total of 8,683 records were obtained, 4,188 of which were WoS and 4,495 of which were Scopus. By removing 2,896 duplicate records and including only English-language research articles, a final dataset of 5,787 articles distributed across 1,639 journals published between 2012 and 2025 was created. Data analysis was conducted in RStudio using the Bibliometrix/Biblioshiny package, applying annual production and citation trends, author-institution-country-level collaboration networks, keyword co-occurrences, trend topic analysis, factor analysis and thematic mapping techniques, and findings were supported by descriptive content analysis. The results show that the field has moved from a phase of rapid expansion to a phase of maturity, with significant growth particularly after 2013 and a peak publication volume in 2019-2022. An interdisciplinary structure with health sciences, engineering, mathematics, language education and K-12 contexts was observed, with the US and China identified as the most productive countries, while institutions based in East Asia and Southern Europe formed strong but relatively limited networks of international collaboration. Conceptual analyses indicate a transition from early work focused on implementation and course design to themes of engagement, motivation and self-regulation, and more recently to artificial intelligence, online/blended learning and learning analytics. The findings suggest that the research on FLM is dynamic yet geographically uneven and that there is a need for more international, equity-oriented and AI-integrated designs in the future.