Investigating Social Effects in Online Networks Through Triadic Analysis

The cognizance of users’ motives behind online interactions can offer sound bases to predict how a social network may evolve and support a host of applications. Social scientists commonly use triadic analysis, which considers three-way relationships, to investigate how social effects drive interaction patterns among users in physical, offline networks. Drawing upon this work, we seek to investigate certain congruencies of the social effects on two types of online networks. The first group consists of online networks designed to promote sharing and communication and includes Facebook, Twitter, YouTube, and Slashdot. The second group consists of citation networks formed by the virtue of acknowledging and distinguishing prior work and includes DBLP, HepPh, Web-Google, and U.S. Patents. This dissertation carries out a comprehensive analysis of the effect of social forces through the means of transitivity and social capital theories on these networks. Through transitivity among triads, we study the influence of offline social factors, namely, stature, relationship strength, and egocentricity on the interactions. We observe that: (i) social stature has a significant influence on the interactions in all networks except in the network of U.S. patents; (ii) the influence of relationship strength is dominant in all networks; and (iii) egocentricity shows a noticeable effect in all networks except U.S. patents. Next, through a novel angle of social capital, we explore the extent to which a particular network is characterized by brokerage or closure capital as well as how strong and weak ties contribute to the presence and absence of these two types of capital. Our analysis depicts that networks composed of strong ties comprise of both brokerage and closure triads. On the other hand, networks composed of weak ties have a higher presence of brokerage triads. In view of the distribution of the constituent triads within the brokerage and closure classes, each network reveals different proportions strongly pointing to its purpose and objectives that drive users’ interactions. Finally, based on the brokerage and closure capital theory, we propose an approach to predict the formation of a new social relationship. We find that brokerage and closure capital are significant predictors of tie formation