Comparing university organizational units and scientific co-authorship communities

TL;DR

This study analyzes university co-authorship networks to understand collaboration patterns within organizational units, revealing how informal communities relate to formal structures using network analysis and community detection methods.

Contribution

It introduces a comparative analysis of co-authorship communities and formal organizational units within a university using complex network measures and bipartite community detection.

Findings

Communities often cross formal organizational boundaries.

Centrality of individuals varies significantly across units.

Network features align with typical complex network properties.

Abstract

A co-authorship network of scientists at a university is an archetypical example of a complex evolving network. Collaborative R&D networks are self-organized products of partner choice between scientists. Modern science is, due to the immanent imperative of newness, strongly interdisciplinary. Crossovers between the different scientific disciplines and organizational units are observable on a daily basis. Since collaborative research has become the dominant and most promising way to produce high-quality output, collaboration structures are also a target for research and management design. We study co-authorship covered by the Citation Index within University College Dublin, a large Irish university. We focus especially on collaborations between organizational units, such as schools or colleges. We compare the centrality of the brokerage individuals within their organizational units. The network of co-authorship is analyzed with standard network measures, with typical features for complex networks observed. We use the bipartite network of authors and papers for the identification of communities of collaborators and clusters of papers. The observed communities are compared to the formal organizational units of the university.