Back to Search
Start Over
What connections lead to good scientific performance?
- Source :
- Scientometrics. 118:587-604
- Publication Year :
- 2019
- Publisher :
- Springer Science and Business Media LLC, 2019.
-
Abstract
- This paper concentrates on the connections in the collaboration network and aims to explore what kinds of connections improve the joint output of the two nodes in connection, using the collaboration data of top institutions in the field of Information Science and Library Science for the period 2007–2016. More intensive connections are found between top institutions, and most institutions are connected into the largest component. The effect of international connection on performance is compared between US and non-US institutions. The homophily of centrality, tie strength and h-index measured as assortativity coefficient is described to show how institutions of similar properties tend to connect with each other in the graph. Furtherly, a negative binomial regression model is employed to investigate the relationship between the homogenous connections and the citation counts received by the connections. Characteristics of connections that contribute to good performance are then obtained.
- Subjects :
- Computer science
Assortativity
05 social sciences
Connection (vector bundle)
Negative binomial distribution
General Social Sciences
Library and Information Sciences
050905 science studies
Data science
Homophily
Information science
Field (computer science)
Computer Science Applications
Graph (abstract data type)
0509 other social sciences
050904 information & library sciences
Centrality
Subjects
Details
- ISSN :
- 15882861 and 01389130
- Volume :
- 118
- Database :
- OpenAIRE
- Journal :
- Scientometrics
- Accession number :
- edsair.doi...........b1b0bafb196c97c04ee0692cf2a366e3
- Full Text :
- https://doi.org/10.1007/s11192-018-02997-7