How to Analyze It
The basics of Social Network Analysis (SNA) Reporting
The general goal of social network analysis is to analyze and show the structure of inter-relationships among sets of entities in a social context. Although the results from social network analysis can be usefully discussed in an entirely qualitative and descriptive way, the field has a strong quantitative and theoretical underpinning with connections to network theory, graph theory and probability theory among others. A basic social network analysis that can be implemented by anyone would include two basic components:
- A summary of network structure based on commonly used summary statistics
- A visual representation of network structure.
Visual representations are important to understand network data and convey the result of the network. Visualization often facilitates qualitative interpretation of network data. There are lots of choices for social network analysis software depending on the level of expertise of the user, ranging from very simple to very complex. One thing to be aware of is that software for social network analysis that allows users to select "off the shelf" analysis will probably impose constraints on the set of metrics and visualizations that is available. That said, there is a core of basic network summary statistics which have robust interpretations, and a similar core set of approaches to network visualizations that will be useful in most situations. These core approaches are available in most, if not all, of the software for social network analysis . For those who are new to social network analysis , we strongly advise consulting an expert in social network analysis if you want to go beyond the most basic approaches.
The Fundamental Concept: Every Network is Also an Array
The homology between networks underlies all of the analyses which occur under the heading of social network analysis , but what does that homology mean? Here is a toy example that illustrates the concept. We have a social network that consists of three friends, Mary, Mungo and Midge. We can represent the network as a triangle in which the corners are drawn as nodes (also referred to as vertices, sometimes) and connected by lines (often referred as edges, connections or links).
We can also represent the same network as a two dimensional array (or matrix) in which each person is represented by one row and one column, with the order of the rows and columns being the same. Now, we specify that the entries in the array correspond to the relationship "person in row x is a friend of person in column y". We enter a 1 in the corresponding cell of the array if the relationship is true and a 0 if it is not true. We assume that the relationship is meaningless for people in relation to themselves, so the main diagonal of the matrix (from top left to bottom right) is all zeros. The two views of our toy example look something like Figure 1.
Levels of network analysis
1. Dyad level
– Cases are pairs of actors (nodes)
– Variables have a value for every pair of actors. Variables are properties of relationships between pairs of actors. From the example above, the variable "friendship" has a value of 1 or 0.
2. Node level (note: nodes can be collective actors)
– Cases are actors
– Variables have a value for each actor. Variables are properties of the node’s position in the network. Metrics are often used to determine the roles of nodes in a network. Of these the most prominent are degree, closeness and betweenness.
Degree is the sum of the links attached to a node, both in- and out-going.
Closeness is the reciprocal of the sum of all the geodesic (shortest) distances from a given node to all others.
A higher betweenness value for a node means that it is on higher number of shortest-paths between nodes, which is an indication of the node’s importance to the network.
3. Group / Whole Network level
– Cases are entire networks
– Variables have a value for each group/network. Variables are properties of the network structure