Sessions‎ > ‎

Intermediate SNA

Instructor: Steve Borgatti

This is a 4-day (Mon-Thu) workshop designed for people interested in gaining both a deeper and broader look at network theory and methods. A key objective is making connections between different network concepts, such as seeing the underlying unity behind families of measures, including those at different levels of analysis, such as measures of cohesion and measures of centrality. It is assumed that participants are familiar with UCINET and have some knowledge about network concepts.

Website
The website for this module, which contains slides, resources, and a detailed schedule, can be found here:

Topics include:
  • Characterizing whole networks both in terms of cohesion and shape, and examining change over time
  • Relations among relations, including change over time
  • Advanced centrality measures, including induced (vitality) measures, betweenness variants, closeness variants
  • Combining attribute data with network data
  • When to use ERGM and QAP
  • Descriptive models of actor strategies as revealed by longitudinal data
  • Using the UCINET command line interface (CLI), for example:
->net = loadexcel(kracknet.xls)  //load excel file and create ucinet file
->draw net //visualize network
->dsp corr(indeg(net),outdeg(net)) //correlate indegree with outdegree
->cent = centrality(net)  //compute centrality measures
->holes = holes(net) //compute structural holes measures
->attr = loadexcel(krackattr.xls) //import attributes from Excel
->data = joincols(cent,holes,attr) //combine centrality, structural holes, and attribs into one file
->saveasexcel data mydata.xls //export to excel
  • Learning the best ways to import data structured in widely varying ways
  • Deriving the right variable for the research question, including things like measuring change
  • Integrating network variables and results into a larger study -- i.e., how do you use this stuff to generate knowledge write publishable papers. 
Schedule
  • Monday: importing data, working with 2-mode data, constructing dyadic proximities, relating relations, network change
  • Tuesday: stochastic modeling via QAP and ERGM
  • Wednesday: node-level analysis, including egonet composition, structural holes, and centrality
  • Thursday: clustering, cohesion and network shape.
  • For a detailed schedule, see the Intermediate website.
Software
  • We will be using both MPNET and UCINET. Please visit our software page.
Recommended Textbook
Additional Readings
  • Bonacich, P. 1987. "Power and Centrality: A Family of Measures." American Journal of Sociology. 92:1170-1182. [^pdf]
  • Borgatti, S.P. 2005. Centrality and network flow. Social Networks. 27(1): 55-71. [pdf]
  • Borgatti, S.P. and Everett, M.G. 2006. A graph-theoretic perspective on centrality. [pdf].