Research performance and degree centrality in co-authorship networks. The moderating role of homophily
| Date | 21 November 2016 |
| DOI | https://doi.org/10.1108/AJIM-07-2016-0103 |
| Pages | 756-771 |
| Published date | 21 November 2016 |
| Author | Kamal Badar,Terrill L. Frantz,Munazza Jabeen |
| Subject Matter | Library & information science,Information behaviour & retrieval,Information & knowledge management,Information management & governance,Information management |
Research performance and
degree centrality in
co-authorship networks
The moderating role of homophily
Kamal Badar
University of Balochistan, Quetta, Pakistan
Terrill L. Frantz
Peking University HSBC Business School, Shenzhen, China, and
Munazza Jabeen
University of Balochistan, Quetta, Pakistan
Abstract
Purpose –The purpose of this paper is to explore the nature of the relationship between a scholar’s
research performance (using weighted journal-impact factor average) and their degree centrality; the
impact of author-homophily (in terms of gender, institutional sector, academic age, academic ranks,
province and city) on this relationship is investigated as well.
Design/methodology/approach –Using scientific publishing data and journal-impact factors from
Thomson Reuters’ISI Web of Science (SCI) and Journal Citation Reports, respectively, the domestic
co-authorship network of chemistry researchers in Pakistan during 2002-2009 was constructed then
modeled via ordinary least squares regression.
Findings –Results show that the personal characteristics of a researcher do not necessarily
lead to high degree centrality, i.e. attributes may not be causal to co-author relationships.
Instead, high degree centrality is more so a function of the forerunning research performance
of the researcher: those whom publish more in terms of impact factor, attract more co-authors
(high degree centr ality). Moreover , the relationship between research performance and degree
centrality is posit ively moderated by ag e and province homoph ily and negatively mo derated by
city homophily.
Research limitations/implications –Data are sourced wholly from the Pakistani chemistry
research community; results many not be generalizable to other sub-populations or the wider research
community.
Practical implications –The findings provide insights to performance-seeking authors: knowing
that their research performance enhances their centrality, which in-turn may lead to increased research
performance and various other desirable professional outcomes. In addition, researchers can look
toward establishing similar (homophilous) or dissimilar (heterophilous) ties knowing that the
relationship between research performance and centrality will likely be stronger when similarity or
dissimilarity exists.
Social implications –This study supports the idea that high research performance attracts more
potential co-authors, which in-turn may lead to ever greater research performance, which suggests that
the research community will be fragmented between high- and low-performing researchers. Also
researcher will have similar or dissimilar ties in terms of various characteristics which in turn
moderate the research performance centrality relationship.
Originality/value –This paper counteracts the empirical belief that researchers are attractive as
potential co-authors according to their personal and professional characteristics. It is actually their
research performance and homophily or heterophily of their ties which matters.
Keywords Pakistan, Research performance, Homophily, Network centrality, Co-author network,
Degree centrality
Paper type Research paper
Aslib Journal of Information
Management
Vol. 68 No. 6, 2016
pp. 756-771
©Emerald Group Publishing Limited
2050-3806
DOI 10.1108/AJIM-07-2016-0103
Received 21 July 2016
Revised 1 October 2016
Accepted 4 October 2016
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/2050-3806.htm
756
AJIM
68,6
Introduction
Research performance is the quintessential metric for research faculty. Whether one is
motivated by career advancement or personal fulfillment, publishing research is a
rigorous condition for membership in the scientific community and employment at
research-oriented universities. Moreover, while some faculty members consider
research their defining characteristic (Crittenden, 1997), some maintain that teaching-
oriented faculty also benefit from conducting research (Webster, 1986), as research an d
teaching activities are considered complementary (Marsh and Hattie, 2002).
While some researchers choose to single-author, depending on the field,
the vast majority of academic research is published jointly with co-authors
(Abramo et al., 2009). Research conducted collaboratively with others can be full of
tension, strife and inefficiencies (Hackett, 2005), but from the perspective of an
individual’s published-research yield, teaming up with others is indispensable (Lee
and Bozeman, 2005). Lately, multiple-author research has become a necessity as
cutting-edge research becomes increasingly complex and interdisciplinary
(Cummings and Kiesler, 2005).
Collectively, scholars co-authoring research publications form an expressive social
network, namely, a co-authorship network (Katz and Martin, 1997; Kumar, 2015a;
Newman, 2004a), though there is debate whether such relationships are an indicator
for meaningful scientific collaboration (Katz and Martin, 1997). Along with the rise
of computing technology and increasing volumes of scientific publishing output,
co-authorship networks have become extensively studied, specifically by applying the
lens and tools of social network analysis (Newman, 2004b).
The construct of centrality is an indication of the advantageous position
of an actor in a social network according to the actor’s relationships and the
structure of the network (Badar et al., 2015). An actor in a co-authorship network
can benefit from the knowledge flowing within the network to derive enhanced
research performance outcomes; a highly central actor is structurally positioned to
benefit more so than others less central. Previous studies have investigated and
identified the positive influence of network centrality on research performance
outcomes in co-authorship networks in various contexts (Badar et al., 2013, 2014;
Bordons et al., 2015; Eaton et al., 1999; Fischbach et al., 2011; Gonzalez-Brambila et al.,
2013; Gonzalez-Brambila, 2014; Lee et al., 2012; Li et al., 2013; Liao, 2011). In addition,
a few studies have also reported diminishing returns of network centrality for
authors in the co-authorship network (Badar et al., 2015; McFadyen and Cannella,
2004; Rotolo and Petruzzelli, 2013). While the aforementioned studies have rigorously
tested the network centrality leading to performance hypotheses, the direction of
causality or the presence of reverse causality has largely been ignored. It can equally
be argued that high research performance causes authors to be central (Perry-Smith
and Shalley, 2003). Moreover, the notion of homophily or similarity in co-authorship
ties has largely been ignored. We believe that homophily might moderate the
relationship between research performance and network centrality (Perry-Smith and
Shalley, 2003).
This paper extends previous research of co-authorship networks by focusing on a
co-authorship network of faculty members from Pakistan publishing in chemistry and
its sub-fields. We propose and test whether high research performance in an earlier
time period (aggregate impact factor) causes authors to be central in a later time period
(degree centrality), and whether homophily in ties –in terms of gender (Sax et al., 2002;
Xie and Shauman, 1998), institutional sector (Lawler, and Yoon, 1998), academic age
757
Moderating
role of
homophily
To continue reading
Request your trial