Co-authorship networks: a review of the literature
| DOI | https://doi.org/10.1108/AJIM-09-2014-0116 |
| Date | 19 January 2015 |
| Published date | 19 January 2015 |
| Pages | 55-73 |
| Author | Sameer Kumar |
| Subject Matter | Library & information science,Information behaviour & retrieval |
Co-authorship networks:
a review of the literature
Sameer Kumar
Asia-Europe Institute, University of Malaya, Kuala Lumpur, Malaysia
Abstract
Purpose –The purpose of this paper is to attempt to provide a review of the growing literature on
co-authorship networks and the research gaps that may be investigated for future studies in this field.
Design/methodology/approach –The existing literature on co-authorship networks was identified,
evaluated and interpreted. Narrative review style was followed.
Findings –Co-authorship, a proxy of research collaboration, is a key mechanism that links different
sets of talent to produce a research output. Co-authorship could also be seen from the perspective
of social networks. An in-depth analysis of such knowledge networks provides an opportunity to
investigate its structure. Patterns of these relationships could reveal, for example, the mechanism that
shapes our scientific community. The study provides a review of the expanding literature on
co-authorship networks.
Originality/value –This is one of the first comprehensive reviews of network-based studies on
co-authorship. The field is fast evolving, opening new gaps for potential research. The study identifies
some of these gaps.
Keywords Social networks, Co-authorship networks, Literature review,
Research collaboration networks, Research collaborations
Paper type Literature review
Introduction
Research collaboration is a key mechanism that links distributed knowledge and
competencies into novel ideas and research avenues (Heinze and Kuhlmann, 2008).
In simpler words, research collaboration connects different sets of talent to produce a
research output. Co-authorship in research articles is considered a reliable proxy
of research collaborations (Barnett et al., 1988; Melin and Persson, 1996). Using
co-authorship to measure research collaboration has been a subject of significant
interest since the 1960s (Beaver, 2001; de Solla Price and Beaver, 1966; Glänzel
and Schubert, 2005; Melin and Persson, 1996). From bringing different talents
together to giving scientific credibility, research collaborations may accrue
several benefits for researchers (Beaver, 2001; Aghakhani et al., 2013). However,
understanding research collaborations fromthelensofsocialnetworksisafairly
young research area.
A social network is a representation of entities (two or more) connected among one
another due to some kind of relationship. The idea of social networks was made
popular by the famous “Small world”experiment of Stanley Milgram in the 1960s
(Milgram, 1967) and “Six degrees of Kevin Bacon”parlour game in the 1990s (Adamic,
1999). Recent studies on large-scale networks by two prominent physicists, B arabási
(Barabási and Albert, 1999) and Watts (Watts and Strogatz, 1998), have provided new
insights into the topologies of networks. The Watts-Strogatz model suggests a single
parameter model, which interpolates between an ordered finite dimensional lattice and
a random graph (Albert and Barabási, 2002) (see Figure 1).
Aslib Journal of Information
Management
Vol. 67 No. 1, 2015
pp. 55-73
©Emerald Group Publis hing Limited
2050-3806
DOI 10.1108/AJIM-09-2014-0116
Received 9 September 2014
Revised 3 November 2014
Accepted 13 November 2014
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/2050-3806.htm
The study is supported by University of Malaya, project number: RP020D-14AFR.
55
Co-authorship
networks
Albert-László Barabási found that the self-organizing networks also have a scale-free
property (Barabási and Bonabeau, 2003). Preferential attachment and growth are two
prominent features of scale-free networks. Preferential attachment means that a node
shows preference for node which it is better connected in comparison with its
neighbouring nodes. Growth means that real networks also demonstrate the featu re of
adding more nodes and links over time. The scale-free networks model overturned the
long standing random networks model of Erdős and Rényi (Erdős and Rényi, 1959,
1960), which postulated non-existence of hubs in a network. Network transitivity and
Community Structure (Girvan and Newman, 2002) are two other properties found in
many networks.
In a social network, the stress is on the relationships between the actors. However,
the attributes of the nodes are not ignored; rather, they are seen in the light of the
relationships that the actors have among themselves. Moreno (1953), the founder of
Psycho-Drama, was one of the first researchers to work in the area of social networks.
During his time and many years after that, the field of social networks was known
as Sociometry. Ever since Moreno, several researchers, e.g., Balevas, Kochen,
Levi-Strauss, Linton Freeman (Linton, 1977) and Howard Aldrich (Aldrich and
Zimmer, 1986), from diverse disciplines, like psychology, anthropology, sociology and
business, have contributed immensely to our idea of social networks (Borgatti
et al., 2009).
A developed set of mathematical algorithms, known as social network analysis
(SNA) (Wasserman and Faust, 1994), are applied for the analysis and visualization of
networks. SNA is a sociological approach to discover the topological properties of a
network. Centrality (degree, betweenness, closeness and pagerank, among others) at
the local level and clustering coefficient, degree distribution, geodesic distance and
communities formation at the global level are often investigated through SNA
(Newman, 2003). Whereas degree of a node is the number of direct connections it has,
closeness and betweenness centralities are path-based, indicating the relative position
of a node in the network (Newman, 2003). SNA has been used in various settings to
examine different phenomena, from organization behaviour (Borgatti and Foster, 2003)
to the spread of obesity (Christakis and Fowler, 2007). The technique has been used to
study the exchange of resources among actors (Haythornthwaite, 1996).
Regular Small-world Random
p = 0 p = 1
Increasing randomness
Source: Watts and Strogatz (1998)
Figure 1.
Watts-strogatz model
56
AJIM
67,1
To continue reading
Request your trial