Centralization of Research Influence and Declining Novelty: Text-Based Approaches to Global Scientific Hegemony

• In Event: 40858 - The Politics of Science Today

Abstract

We investigate how increasing centralization of research influence among elite countries affects the novelty and disruption of scientific knowledge production. While science is often perceived as a global enterprise, disparities in access, attention, and influence shape the trajectory of research in ways that may stifle innovation. We argue that as researchers worldwide increasingly follow dominant scientific trends set by elite nations, intellectual diversity declines, limiting the potential for serendipitous discoveries. Using a dataset of 27 million academic publications from OpenAlex (1995–2020), covering 122 disciplines, we employ text-based methods—including phrase extraction, topic modeling, and similarity measures—to map international networks of research influence. Unlike traditional citation-based metrics, our approach directly analyzes the content of research to track how scientific ideas spread and evolve across national and disciplinary boundaries. Our findings reveal a decline in research novelty, as measured by textual indicators of disruption and intellectual "churn." Countries in the global periphery increasingly align their research agendas with those of dominant nations, focusing on similar topics while narrowing the scope of inquiry. This suggests that centralization of influence is not merely about recognition but about shaping what gets studied globally. While prior work attributes declining scientific disruption to the sheer scale of modern research, we highlight the role of global research hegemony in limiting conceptual innovation. By integrating computational text analysis with sociological perspectives on scientific inequality, we provide new insights into the mechanisms through which dominant nations shape global knowledge production. Our work underscores the need to rethink how research agendas are set and the broader consequences of a homogenized global scientific landscape.

Authors

• Charles Jonathan Gomez, University of Arizona

• Yeaeun Kwon, University of Arizona

• Jeffrey Shen, University of Arizona

• Sarah Stueve, University of Arizona

• Harshvardhan Singh, University of Arizona

• Jessica Dawson, Army Cyber Institute