Factors that influence triadic closure in social networks
A recent study has sought to establish a predictive model of the factors that influence the formation of closed triads in two popular microblogging sites: Twitter and Weibo. The property of strong triadic closure holds that in a situation where two nodes (A and B, for example) are linked by strong ties to a third node (C), nodes A and B will be predisposed to having a tie form between them. This working principle is visible in the growth and development of social networks, of which closed triads comprise some of the most basic elements. Although the universality of strong triadic closure is evident, the factors that influence its occurrence in specific situations vary depending on the context and nature of the social network in question.
To address this issue, the study defined these influencing factors (in the context of Twitter and Weibo) into several categories, chief among them network topography, demography, and social role. Network topography considered the structure of the network, demography concerned the gender and geographic location of individuals in the network, while social role was examined in terms of popularity of individuals. After analyzing network elements in Weibo and Twitter, the authors of the study determined that in terms of network topography, if the two unlinked users in a potential triad were linked via two-way relationships to their shared acquaintance, they were more likely to form closed triads than a cluster of three nodes correspondingly linked with just one-way relationships. Concerning demography, the authors concluded that the geographic location of users did not have a significant impact on the formation of closed triads based on examination of trios of users from the same city/province, as the very nature of online social networks ensures that geographic location is no longer a limiting factor for the establishment of strong ties. Regarding the second component of demography, however, the authors discovered that gender was a fairly influential factor, as women were significantly more likely to close a triad than men in equivalent circumstances. Lastly in terms of the factor played by social role, the authors made two distinct conclusions on the impact of user popularity. They found that although the popularity of a user did not have a significant impact on triadic closure among ordinary users mutually acquainted with the same popular user, they discovered that popular users were more likely to form closed triads with other popular users.
In regards to the strong triadic closure property, the various factors used to define the study’s predictive model could be interpreted as indicators of the strength of ties between nodes. In class, we did not specifically define what constitutes a strong and a weak tie, and the definition would definitely vary greatly depending on the situational context. This study sheds some light on what factors might influence triadic closure and might also be used to define the strength of ties between the nodes of a triad in terms of network topography and social role in particular.
Reference: http://keg.cs.tsinghua.edu.cn/jietang/publications/WWW14-Huang-et-al-triadic-closure-prediction.pdf