Mapping institutional change research in higher education: A BERTopic-based topic modelling study

Research on institutional change in higher education has grown quickly, but it remains fragmented across disciplines and theoretical traditions. This study maps the thematic structure of this research in educational sciences and traces how it has changed over 34 years. I analysed the titles and abstracts of 2,623 publications indexed in the Web of Science Core Collection (Education & Educational Research, 1992–2025) with BERTopic, an embedding-based topic modelling tool. Placeholder masking kept the names of countries, nationalities and languages from shaping the topics, and key model settings are reported for reproducibility. The model produced 24 interpretable topics covering 1,824 publications. The general Structural Reforms topic fell from 65.6% of topic-assigned publications in 1992–2000 to 12.5% in 2023–2025, while more specific themes such as sustainability, transformational leadership and diversity, equity and inclusion gained weight. In 2023–2025, Digital Transformation became the largest topic (14.8%), and Artificial Intelligence Integration rose from almost zero to 8.1%. Technology-related research developed in successive waves: e-learning, emergency remote teaching during the COVID-19 pandemic, digital transformation and generative artificial intelligence. Each wave reframed digitalisation with new concepts instead of simply continuing the previous one. These results suggest that institutional change in higher education is studied through several parallel pathways rather than a single dominant narrative. For policymakers and institutional leaders, the findings point to the need to build the capacity for change as a lasting institutional capability rather than responding to each technological, sociological and political disruption separately.