The relationship between acquaintanceship and coauthorship in scientific collaboration networks

This article examines the relationship between acquaintanceship and coauthorship patterns in a multi-disciplinary, multi-institutional, geographically distributed research center. Two social networks are constructed and compared: a network of coauthorship, representing how researchers write articles with one another, and a network of acquaintanceship, representing how those researchers know each other on a personal level, based on their responses to an online survey. Statistical analyses of the topology and community structure of these networks point to the importance of small-scale, local, personal networks predicated upon acquaintanceship for accomplishing collaborative work in scientific communities.

💡 Research Summary

This paper investigates the relationship between acquaintanceship and co‑authorship within a large, multidisciplinary, multi‑institutional, and geographically dispersed research center—specifically the Center for Embedded Networked Sensing (CENS). The authors construct two distinct social networks from the same population of researchers: (1) a co‑authorship network derived from bibliographic records of 608 papers published between 2003 and 2009, involving 391 unique authors; and (2) an acquaintanceship network based on a bespoke online survey that asked participants to identify colleagues they have met in person and would greet if they encountered them, as well as to report the year of first meeting and frequency of contact.

For the co‑authorship network, edges are weighted using a formula that assigns greater weight to collaborations involving fewer co‑authors and to repeated joint publications: wᵢⱼ = Σₖ δₖᵢ δₖⱼ /(nₖ − 1), where δₖᵢ indicates author i’s participation in paper k and nₖ is the number of co‑authors on that paper. This weighting reflects the assumption that smaller author teams imply stronger interpersonal collaboration, and that repeated co‑authorship signals a deeper relationship.

The acquaintanceship survey was sent to all 388 individuals appearing in the bibliographic list (three were excluded due to missing contact information). A total of 191 responses were received (49 % response rate). The first survey page displayed every potential acquaintance with name and thumbnail, grouped by department and institution, and participants selected those they recognized personally. The second page collected temporal and interaction frequency data for each selected tie. The resulting data included reciprocal, non‑reciprocal, and partially missing ties. To align with the undirected nature of the co‑authorship network, the authors transformed all ties into undirected edges using an “available‑case” approach, thereby incorporating complete, incomplete, and non‑reciprocal ties. Edge weights in the acquaintanceship network were derived from reported contact frequency, with more frequent interaction receiving higher weight.

Topological analysis revealed that both networks exhibit small‑world characteristics, but the acquaintanceship network is denser (higher average degree and clustering coefficient) and has shorter average path lengths than the co‑authorship network. Community detection (Louvain method) identified several clusters in each network; the overlap between co‑authorship and acquaintanceship communities was moderate (adjusted Rand index ≈ 0.42), indicating that while many co‑authors are also acquaintances, a substantial proportion of acquaintances are not co‑authors and vice‑versa. Notably, large, interdisciplinary projects—common in CENS—often produce papers with many authors who have never met face‑to‑face, confirming that co‑authorship does not guarantee personal acquaintance.

The authors critique the common practice of treating co‑authorship networks as proxies for social networks (e.g., Newman 2004) and demonstrate, with empirical evidence, that this assumption breaks down in modern, distributed scientific collaborations. Their findings suggest that digital communication tools enable extensive co‑authorship without fostering the personal ties traditionally associated with scientific collaboration. Consequently, they argue that future network‑based analyses of scientific collaboration should distinguish between formal scholarly ties and informal personal ties. Moreover, they recommend that the design of collaborative infrastructures (e‑science platforms, virtual labs) incorporate mechanisms to promote face‑to‑face interaction or sustained personal contact, as these appear essential for building the trust and shared understanding that underlie effective teamwork.

In summary, the study provides a rigorous comparative analysis of co‑authorship and acquaintanceship networks, showing that while they are related, they capture distinct dimensions of scientific collaboration. The work underscores the importance of measuring both types of ties to accurately assess the social fabric of contemporary research communities and to inform the development of tools and policies that support both scholarly productivity and the personal relationships that sustain it.