A study of blog networks to determine online social network properties from the tie strength perspective

Date29 April 2014
Pages381-398
DOIhttps://doi.org/10.1108/OIR-01-2013-0022
Published date29 April 2014
AuthorTerry Hui-Ye Chiu,Chien-Chou Chen,Yuh-Jzer Joung,Shymin Chen
Subject MatterLibrary & information science,Information behaviour & retrieval
A study of blog networks
to determine online social
network properties from the
tie strength perspective
Terry Hui-Ye Chiu, Chien-Chou Chen,
Yuh-Jzer Joung and Shymin Chen
Department of Information Management, National Taiwan University,
Taipei, Taiwan
Abstract
Purpose – Most studies on tie strength have focused on its definition, calculation and applications,
but have not paid much attention to how tie strength can help analyse online social networks.
Because ties play different roles in a network depending on their strength, the purpose of this paper is
to explore the relationship between tie strength and network behaviours.
Design/methodology/approach – The authors proposea simple metric for tie strength measurement
and then apply it to an on line social netwo rk extracted from a blog netw ork. These networks a re
massive in size and have technology for efficient data collection, thereby presenting the possibility
of measuring tie st rength objectively. From the results s everal key social net work properties
are studied to see how tie strength may be used a s a metric to explain cer tain characteri stics in
social networks.
Findings – The online networks exhibit all the structural properties of an actual social network, not
only in following the power law but also with regard to the distribution of tie strength. The authors
noted a strong association between tie strength and reciprocity, and tie strength and transitivity in
online social networks.
Originality/value – This paper highlights the importance of analysing online social networks from
a tie strength perspective. The results have important implications for the development of efficient
search mechanisms and appropriate group leaders in virtual communities.
Keywords Online social networks, Tie strength, Strong ties, Weak ties, Clustering coefficients
Paper type Research p aper
Introduction
A social network is a social structure comprising nodes (such as a set of actors
or organisations) and dyadic ties (such as relations or interactions) that connect
these nodes. Studies on the structural properties of social networks have revealed
some common features (e.g. degree distribution and clustering coefficients), and these
features were studied to understand the behaviour of social networks. For instance
analysing degree distribution and clustering coefficients can identify regions where
innovations occurred and thus find ways to accelerate development of innovations
(Fleming et al., 2011). Thus understanding social network properties and network
behaviour may help people to design such networks more effectively and efficiently.
Moreover identifying additional essential network properties would increase that
effectiveness and efficiency.
Studies of social network properties often assume that all relationships in
a network are equal. However, in real life, people maintain many relationships with
varying tie strength. For example close friends have much stronger ties than those of
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1468-4527.htm
Received 26 January 2013
Second revision approved
9 August 2013
Online Information Review
Vol.38 No. 3, 2014
pp. 381-398
rEmeraldGroup PublishingLimited
1468-4527
DOI 10.1108/OIR-01-2013-0022
381
Online social
network
properties
acquaintances; therefore Granovetter (1973) has designated the former as “strong ties”
and the latter “weak ties”. In general tie strength refers to the intensity and tightness of
a tie between two nodes, measured by, for example, intensity of friendship or frequency
of interaction. Sometimes, however, weak ties are instrumental in finding better jobs
and in diffusion of ideas and information (Brown and Reigen, 1987; Burt, 2004;
Granovetter, 1974; Weenig, 1993; Zhao et al., 2012), and some weak ties are more
influential on decision making (Steffes and Burgee, 2009). Strong ties can reinforce
support and beliefs between groups or organisations an d thus enhance trust and
information exchange (Kraatz, 1998; Holzinger et al., 2009). McGee et al. (2011) have
shown how close friends (peers with strong ties) in the Twitter network tend to be
geographically close compared to peers with weak ties.
Ties of different strengths play different roles in networks, and these different tie
strengths may affect the social structure differently. Granovetter (1973) showed that
networks with a high clustering coefficient tend to fragment into local groups that have
many strong ties and a small set of weak ties whose primary purp ose is to act as a
bridge to remote groups.
The clustering coefficient is a structural property of social networks, and the
Granovetterexample indicates that it can be used to predict certain network behaviours.
Tie strength is not considered in calculating the clustering coefficient; however, it is
notable thatthe resulting predictions cluster nodes according to their tie strengths. Thus
tie strength could be the underlying cause of such clustering; in other words these
differenttie strengths may have caused such network behaviours. Therefore it is logical
to directly explore the relationship between tie strength and network behaviours.
Previous studies have attempted to link tie strength to certain structural properties
summarised by Newman (2003). Researchers found that tie strength, network density,
network centrality (Kate et al., 2010), and homophily (Shin et al., 2011) have a positive
influence on users’ involvement in the network. Kivran-Swaine et al. (2011) found that
reciprocity (density) in a dyadic relationship may indicate a strong tie in the Twitter
network. Bapna et al. (2011) revealed that tie strength is correlated with trust and
reciprocity in the Facebook network. Onnela et al. (2007) reported that the local
structure is associated with tie strength in the mobile network and removing weak ties
can disintegrate the network. Thus it is clear there is a relationship betwe en certain
structural properties and tie strength, and our present pap erexamine s how tie strength
can be used to explain certain network behaviours, as measured by some structural
properties, in online social networks. The online social network is selected due to its
scale and ability to retain long periods of interactions. This allows researchers to
analyse the network more accurately and objectively.
The paper is organised as follows. First we provide background information and
describe general characteristics of social networks. Then we explain our methodology
and rationale for collecting data and using an online social network as the proxy for an
actual network. Next we describe how tie strength is measured and used to analyse the
experimental results. The subsequent section examines the structural proper ties of our
network and demonstrates how tie strength is associated with those properties. Finally
we discuss some implications and present our conc lusions.
Related work
Tie strength
In his seminal paper Granovetter (1973, p. 1316) used “tie strength” to characterise
interpersonal relationships in a social network as follows: “the strength of a tie is
382
OIR
38,3

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT