Social Network Analysis

Social network analysis (SNA) is the examination of social networks and their behavior through the framework of network theory. In network theory, networks typically consist of nodes within the network connected by edges or links. In social networks, the nodes represent individuals, groups, organizations, or larger social systems. The links connecting the nodes signify relationships such as friendship, kinship, organizational position, sexual relationship, and so forth. Regardless of the level of social [Page 850]system or type of relationship, SNA seeks to describe the structure and pattern of these relationships and to understand both their root causes and end consequences. The study of these relationships has been used throughout anthropology, biology, communication studies, social psychology, public relations, and sociolinguistics to examine network dynamics and locate influential actors within each network. The study of social networks helps public relations professionals influence perception and spread strategic messages through a network by identifying and leveraging these influential actors.

Relationships in the network can be directed (“seeks advice from”), undirected (“shares information with”), positive (love, friendship, alliance, partnership), or negative (hatred, anger, rivalry). Network analysis can involve the qualitative study of these connections or a quantitative study of the number of relationships, regardless of how good or bad they are.

Clusters are used to identify and classify groups in a social network. A cluster is a tightly knit, highly bonded group both belonging to and distinct from the larger network. Identifying clusters has become one of the most important applications of SNA. As Amazon has demonstrated with its algorithms that show what other customers who bought certain items also bought, sales can be improved by identifying these linked groups in the larger network. To identify clusters, analysts often look at relationship density. A node's or actor's density is the number of links that connect that node or actor to a group of interest divided by the maximum possible number of links that could exist from that node or actor. The closer the density is to 1, the more connected that node or actor is to the rest of the network.

Analysts try to find the bridgers and the hubs in the network to best analyze network behavior. Bridgers are individuals who have connections in multiple clusters, and thereby bridge distinct subgroups in a network. This connection makes them important relay points between clusters. Bridgers are often overlooked because their significance is not obvious from their density. A better metric is betweenness centrality, a measure that calculates the number of shortest paths to other nodes that pass through that node. Analysts also use network constraint. An individual's network constraint measures how many links a node has with other nodes that are already connected to each other. High betweenness centrality and low network constraint indicate bridgers.

Hubs are individuals within a social network who have the most influence, or are sought after by other network members. They are best measured using in-degree centrality, or counting the number of directed links (“seeks advice from”) as opposed to undirected links (“shares information with”) from other nodes to the hub node. More advanced metrics count not only the number of directed links but also how influential the nodes seeking the hub node are in the network.

...