From Evolution to (R)evolution: Modeling Thematic Innovation in the Italian Scientific Landscape

Innovation is widely recognized as a driver of scientific research and the evolution of knowledge production over time. Focusing on statistics as a STEM field, this study examines how innovation dynamics unfold in Italy within a domain characterized by methodological heterogeneity and cross-disciplinarity. The data cover the Italian population of researchers affiliated with the SECS-S/01 to SECS-S/05 sectors, as defined by the Italian Ministry of University and Research. N = 11, 789 abstracts published between 2012 and 2022 were retrieved. Given the short length of documents and annual publication data, we propose a methodological strategy that complements Structural Topic Modeling with two ex post measures: the Dynamic Structural Innovation Index, which captures the structural semantic distinctiveness of each topic with respect to the core of the field; and the Structural Topic Innovation Index, which identifies temporal topic-level innovation by combining semantic differentiation with yearly topic presence and growth.