Advanced Network Analysis
Description
This is a 4-day (Mon-Thu) workshop taught by Steve Borgatti. The workshop is intended for participants who already have some experience with network analysis, but would like to deepen and broaden their knowledge. Topics include: analysis of network change, advanced centrality methods (including KeyPlayer methods), advanced approaches to 2-mode data, analyzing negative ties, working with multiple relations, and testing unusual hypotheses. In addition, participants will learn more about UCINET and NETDRAW efficiently (including powerful undocumented features). Finally, we will analyze some substantial datasets that are more interesting than the toy datasets we usually use for teaching.
Analyzing Network Change. Siena models are great for analyzing longitudinal data, but sometimes we need to (a) find simpler and more stable models, (b) summarize trends over time, (c) create variables that assess network change at different levels of analysis (e.g., node level change statistics), and (d) model patterns in network change as a function of other variables (e.g., some nodes make new friends through old friends, others acquire new ties independently. We can correlate these tendencies with personality characteristics).
Advanced Centrality. How to use more complex centrality measures such as Bonacich's Beta Centrality, including how to select beta. New classes of centrality measures, such as ones calculated by measuring a network statistic, removing a node, and then recalculating the network statistic to see how much that node was contributing to the network as a whole.
2-Mode data. We discuss advances in both constructing better 1-mode projections of 2-mode data, and analyzing 2-mode data directly.
Software Tips and Tricks. For example, how to correlate all members of a set of independent variables with all members of a set of dependent variables, but without correlating all the indeps with each other. How to compute shortest paths using valued data.
Slides
Schedule
Monday
Graph theory and matrix algebra refresher
Tuesday
Advanced centrality
Working with attributes, including incorporating them into centrality calculations
Wednesday
Randomization tests, including LRQAP (logistic regression QAP) Part I
Analyzing network change Part I
Thursday
Analyzing network change Part II
Advanced data entry formats
2-mode data
All-request ...
Readings