Network Analysis

Definition of Networks

Networks are social relational systems characterized by a set of actors and their social ties. They consist of a finite set of identifiable nodes; the relationships among these nodes (actors) are represented as ties (edges). Ties are dichotomous: present or absent, or unordered (undirected) or ordered (directed ties). The nodes represent a single entity that potentially may take part in the relationships under study. A network analysis takes the characteristics of nodes and the characteristics of the relations connecting the nodes into account. Attributes—additional information about the actors—are characteristics, such as the behavior, attitudes, or other properties of actors. Networks are represented as a (social) graph, which is defined by a set of nodes together with the set of pairwise relationships among them. Networks present an opportunity and constraint structure: on the one hand offering access to resources flowing through the ties and, on the other hand, restricting choices and controlling behavior.

Definition and Goals of Network Analysis

Network analysis is an interdisciplinary field of research with a history in sociology and anthropology. It provides the formal mechanisms for representation, measurement, and modeling of relational structure and is based on the assumption that actions and decisions of actors are dependent on the context and the actions of other actors. The social structure is an area of (inter)action in which emergent patterns of behavior can be observed, and structural variables help analyze the resulting interactions. Generally, a network consists of actors (nodes) that are connected with each other through ties (edges). The actors can represent people, organizations, countries, or other entities. The relations and ties are conduits for the flow of resources in the form of knowledge, finances, collaboration, and so on. The units of analysis are, therefore, the interactions between the actors.

In a network analytical approach different types of data can be distinguished:

• attributes, which are descriptors of the individual actors in the network, such as age, sex, profession, or political affiliation;
• relational data that are derived from the interactions of at least two actors, such as Country A imports Product x from Country B; and
• structural characteristics of the overall network that can be derived from the relational characteristics of all studied actors, such as density or centralization.

The main goal of network analysis is to understand the emergence of the network structure and its consequences by description, visualization, and statistical modeling. In more formal terms, network analysts analyze which independent variables have led to the observed structure and how the social structure influences other emerging social processes. In this form of “structural analysis,” the relationships among actors become the first priority, and individuals' properties are secondary in the study of the flow of structural regularities that might influence actors' choices and their resulting behavior. The result is an approximation to the structure of a more complex system for purposes of studying a particular property (such as the diffusion of a disease in a community). Ultimately, complex situations can be represented using multiple relationships (multiplexity), such as group membership, friendship, hate relationships within the group, hierarchy and reporting structure among the group, and the strengths and frequencies of interactions. There are different ways of looking at the social structure: The researcher can either look at the outcomes of the existing network structure, at the emergent behavioral patterns, or the characteristics of the overall structure.

...