This course is closed, registration is no longer possible.
Inferential Network Analysis introduces statistical methods for analysing networks. Networks, or graphs, are sets of nodes and their connecting ties. Networks are used to model a wide range of complex social and political phenomena, such as policy networks among political actors around the design of new regulation; lobbying ties of interest groups to policy makers or lobbying coalitions; recruitment of experts into international organizations; patronage relationships; multilevel or collaborative governance systems; international relations, including conflict, alliances, trade, and migration; financial deals between organizations; political debates among actors about policies; or the diffusion of policies among states. Systems like these usually evolve over time in complex ways, and as researchers we want to understand the formation of ties between nodes at the micro level in order to understand how the system evolves. We also want to understand the adoption of behavior as a consequence of being embedded in a network.
The central questions are:
- How can we model any data where the observations are not independent and identically distributed (i.i.d.)?
- How can we explain and model connections between nodes (or characteristics of nodes) using covariate data and theories about the endogeneity in the data?
- And how can we simulate such processes forward in time to predict future states of the network or the characteristics of the nodes?
The course “Inferential Network Analysis” introduces a range of statistical models for explaining and predicting the formation of ties in networks using characteristics of the nodes, their ties, and the network. You will consider the exponential random graph model (ERGM) as the workhorse model of statistical network analysis in the first half of the course, including specification, estimation, and implementation in R.
You will then discuss extensions to temporal and valued relations before you consider alternative modeling choices, including the family of latent space models, the quadratic assignment procedure, the stochastic actor-oriented model, and the relational event model. Finally, you will also learn about the use of network autocorrelation models to explain the state or behavior of a node by considering the node’s network embeddedness. All models will be discussed theoretically, in application, and practically using R. Participants will be given daily assignments to solve after class to maximize the learning experience.
