The Core-Periphery Divide in Science

At the heart of this research is the concept of a core-periphery structure—a pattern seen in many networks, from social circles to transportation systems. In science, the “core” consists of highly cited, well-connected elite researchers who frequently cite each other’s work. The “periphery” includes the vast majority of scientists whose work receives less attention and whose connections are weaker. Using data from the Microsoft Academic Graph (covering over 200 years of publications), the study mapped citation networks in three fields: nanoscience, chemical physics, and human-computer interaction.

The findings are stark: over time, the gap between “core” and “peripheral” scientists has widened. Connections among elite “core” authors have grown stronger, while interactions among ordinary “peripheral” researchers remain sparse. It’s a scientific version of the “rich get richer” phenomenon: once a researcher gains visibility, their work attracts more citations, funding, and collaborations, creating a self-reinforcing cycle. This pattern mirrors the Matthew Effect, a term coined in sociology to describe how advantage accumulates disproportionately for those already ahead.

How Inequality Manifests in Citations

To measure this divide, the study used three distinct yet interconnected lenses:

• Connection Strength: Core authors cite each other 4–5 times more intensely than expected by chance. Peripheral authors, however, cite one another 10% less than random networks predict.
• The “Rich Club” Phenomenon: Like exclusive social clubs, top-cited authors form tightly knit circles. Their mutual citations are disproportionately high, even after accounting for their productivity.
• Assortativity: Over time, scientists are less likely to cite peers with similar levels of influence, suggesting a growing preference for targeting established names over lesser-known work.

These patterns hold across all three disciplines studied, indicating a systemic issue rather than a field-specific quirk. For example, in nanoscience—a rapidly growing field—the percentage of “core” authors has shrunk since the 1980s, even as their absolute numbers grew. This reflects a concentration of influence: a shrinking elite holds outsized sway over what research gets noticed.

Why This Matters: The Cost of Ignoring the Periphery

Scientific progress thrives on diverse ideas. When attention clusters around a few stars, several risks emerge:

Lost Innovations: Breakthroughs often come from unexpected places. Peripheral researchers—working in niche areas or challenging dominant theories—may struggle to gain traction.

Career Barriers: Early-career scientists and those from underrepresented regions or backgrounds face steeper climbs to visibility, deepening long-standing inequities in funding, collaboration, and career advancement.

Echo Chambers: Overly dense core networks can lead to intellectual inbreeding, where popular ideas are reinforced while novel ones are overlooked.

The study also tested what happens when “core” authors are removed from the citation network. The result? The remaining periphery fragments further, revealing weak connections among ordinary authors. This suggests that the core isn’t just dominant—it’s structurally essential to holding the network together. Without intentional efforts to bridge the divide, the periphery risks becoming increasingly isolated.

If we fail to intentionally bridge the growing divide in science, we risk silencing the periphery—those researchers whose ideas, though less visible, may hold the seeds of tomorrow’s breakthroughs. The cost is not only a lost opportunity for individuals, but a diminished scientific future for us all.

Pathways to a More Inclusive Future

The paper isn’t just a diagnosis—it’s a call to action. To foster a healthier scientific ecosystem, the authors propose three strategies:

• Amplifying Marginalized Voices: Journals, conferences, and funding agencies could prioritize work from early-career researchers or those outside traditional power hubs.
• Rewarding Collaboration: Incentivizing cross-institutional and interdisciplinary partnerships might dilute the rich-club effect.
• Transparent Metrics: Moving beyond simplistic measures like citation counts to assess impact could reduce bias toward established names.

Looking Ahead

This research opens a series of critical questions: How do collaboration networks reinforce citation inequalities? Can digital platforms (such as preprint servers or open peer review) democratize attention? And what role do emerging fields, like AI-driven research, play in either mitigating or exacerbating these divides?

While the trends are concerning, the study offers a glimmer of hope. Even among less visible scientists, sub-groups of scientists are beginning to cite and support one another more frequently—an encouraging sign that grassroots communities can foster visibility. By intentionally designing policies that nurture these connections, the scientific community can shift from a hierarchy of stars to a mosaic of diverse, interconnected voices.

If science is to remain vibrant and forward-looking, the solution lies not in flattening all hierarchies, but in intentionally building bridges. Funding mechanisms, institutional policies, and academic evaluation systems must evolve to recognize not just disciplinary excellence, but the full spectrum of intellectual and epistemological diversity. Supporting interdisciplinary collaboration, investing in early-career researchers, and valuing contributions from underrepresented voices can help transform isolated clusters into interconnected communities of innovation.

At its core, science is a collective endeavor. Ensuring that all contributors—not just the elite—can participate fully isn’t just about fairness; it’s about safeguarding the creativity, resilience, and adaptability that drive discovery.

The original article on which this essay is based is: Wang, Haoyang, Win-bin Huang, and Yi Bu. “The attention inequality of scientists: A core-periphery structure perspective,” Information Processing and Management (2025).