Answers are grounded in this paper and cite their sources.
The “collaboration premium,” higher citations for international research teams, is a foundational finding in scientometrics. However, existing studies conflate geographic diversity (where authors work) with diversity in authors' backgrounds, and leave unresolved whether the premium reflects superior novelty or wider audience reach . We address these gaps by analyzing 139,560 publications (2021–2025) across six policy-priority research domains (AI, Quantum, Space, Management, Sustainability, Wellbeing). Using a validated LLM-based method to infer the regional origin associated with author names, we separate Geographic Diversity (Single/Domestic/International Institution) from Regional-Origin Diversity (Single/Multiple Origins) and examine their associations with citation visibility at three levels: individual papers, national portfolios, and topic clusters. Three findings emerge. First, teams that are both internationally collaborative and diverse in regional origin are 1.6–2.8 times more likely than single-institution, single-origin teams to produce a Top 10% cited paper; the premium persists when teams of the same size are compared, and in five of six domains it is then statistically indistinguishable from the sum of the two single-dimension premiums. Second, at the country level, both dimensions correlate positively with citations, significantly so in five of six domains. Third, the associations are level-dependent: within topic clusters, regional-origin diversity is positively associated with citations in every domain, whereas between clusters the association is weaker and negative in sign in four domains, although not distinguishable from zero—a pattern consistent with a Simpson's paradox. These descriptive findings are consistent with an audience-based interpretation of the collaboration premium and suggest that policies aiming to enhance research visibility may benefit from attending both to international partnerships and to the diversity of the domestic research workforce.
No takes yet. Share an insight, caveat, or question.
Mejia et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: