Network Analysis II: Advanced Topics
This is a course on inferential network analysis. The conventional categorization of data analytic methods into descriptive and inferential statistics can be fruitfully applied to network analysis. Descriptive methods of network analysis are important for illuminating structural features of a given network, but they cannot be used to build and/or test theories about the generation of networks. Inferential methods of network analysis can be used to test hypotheses about the generation and evolution of a network, derive measures of uncertainty for network indices, and find probabilistic models that accurately describe the overall features of a network.
The first week will look at regression alternatives and variations in the network context (Quadratic Assignment Proceedure, Network Auto-Correlation Model) which are both designed to account for network dependencies without formulating complex network statistics. We move from these models to Exponential Random Graph models (ERGMS) which can be parametrized to represent complex dependence processes and the effects of exogenous covariates. In the third and fourth weeks we will cover statistical models for longitudinal networks. These will include a longitudinal extension of the ERGM -- the Temporal ERGM, and the actor-oriented model of network dynamics (i.e., SIENA). We will present each model mathematically, discuss published social science applications of them, and utilize the models on example datasets.
All computing will be conducted in the R statistical software. Prerequisites include the Network Analysis I: Introduction at ICPSR or an equivalent introduction to network analysis, as well as familiarity with statistical modeling.
Fees: Consult the fee structure.
Location: ICPSR -- Ann Arbor, MI
Date(s): July 24 - August 18
Time: 10:00 AM - 12:00 PM