Introduction
Evidence-based policymaking is the cornerstone of modern global governance, relying on the effective translation of scientific research into actionable guidelines. This critical task is largely managed by Intergovernmental Organizations (IGOs) such as the World Health Organization (WHO) and the Intergovernmental Panel on Climate Change (IPCC). Despite the importance of this process, the micro-mechanisms governing how science enters the policy sphere have remained largely opaque. A recent study investigating over 230,000 papers cited by IGOs reveals that this knowledge flow is dominated by a tiny, highly interconnected elite of "Policy-Influential Scientists" (PI-Sci), driven by a pronounced Matthew effect where past influence dictates future visibility.
Mapping the Science-to-Policy Pipeline
To understand how research moves from the laboratory to the legislative floor, researchers linked policy document citations from the Overton database with academic metadata from Scopus. The analysis shows that while direct citations are rare—only 0.71% of papers published since 2015 are directly cited in IGO documents—the broader scientific landscape is closely connected to policy through citation networks.
As illustrated in Figure 1, approximately 93.3% of all papers are connected to a policy citation within just three to four steps in the citation network. This suggests that while few papers are "policy-facing," the vast majority of scientific knowledge resides in the immediate intellectual neighborhood of global policy, regardless of the specific scientific domain.
The Dominance of Policy-Influential Scientists
The study identified a small group of researchers, termed "Policy-Influential Scientists" (PI-Sci), who act as the primary conduits for science-to-policy translation. Across 23 distinct research domains, PI-Sci make up fewer than 5% of all policy-cited authors but account for 30% of the total policy-cited literature.
This concentration is particularly extreme in mature, highly organized fields. In Climate Modeling, for example, a mere 0.62% of scientists (258 individuals) generate 30% of the policy-cited work. Notably, over 40% of the top 100 climate PI-Sci have served as official authors for the IPCC, highlighting the overlap between formal advisory roles and bibliometric influence. These elite authors often form "rich-clubs"—dense, internationally spanning co-authorship networks that coordinate the flow of information to IGOs.
The Matthew Effect and the Role of Status
One of the most striking findings is the presence of a "Matthew effect" in policy citations: a scientist's past policy influence is the single strongest predictor of their future influence. Once a researcher establishes a foothold in the policy sphere, their impact compounds rapidly, creating a feedback loop of influence.
Figure 2 demonstrates the predictive power of various factors. While simple metrics like publication volume (N) or journal prestige (J) provide a noisy estimate of policy impact, accounting for an author's past policy citations (A) and their use of policy-cited references (R) significantly improves the model's accuracy, bringing the ratio of actual to estimated citations close to 1.0. This trend holds true across different global regions, suggesting that IGOs rely heavily on established reputations and existing advisory networks.
Further evidence for the importance of status comes from a semantic matching experiment. When comparing papers with nearly identical content, those authored by a PI-Sci were nearly twice as likely (50.1% vs 28.1%) to be cited in policy documents. This implies that the identity and network of the author are often as important as the intrinsic relevance of the research itself.
Field Maturity and Knowledge Cascades
The structure of influence is not uniform across all sciences. In mature fields like climate science, influence is hyper-concentrated around central hubs. In contrast, emerging or fragmented fields like Data Science and AI exhibit a more dispersed influence structure, lacking a clear collaborative core.
Additionally, the research identified a "synchronized policy cascade." Major IGOs like the UN and WHO act as source nodes, integrating new research from PI-Sci first. This knowledge then trickles down synchronously to smaller, more specialized policy bodies, reinforcing the central role of elite scientists in shaping the global policy agenda.
Conclusion
The reliance on a narrow "policy elite" raises important questions about the diversity of evidence used in global governance. While this concentration may facilitate efficient knowledge transfer through trusted channels, it may also obscure valuable research from the Global South or non-English speaking regions. Understanding these bibliometric pathways is essential for ensuring that global policy remains grounded in the most comprehensive and diverse scientific evidence available.
Summary updated