Social Networks: Statistical Approaches
A 3-Day Remote Seminar Taught by
John Skvoretz, Ph.D.
To see a sample of the course materials, click here.
The best preparation for this course is to take Introduction to Social
Network Analysis, taught by Stephen Borgatti, March 11-13.
The study of social networks focuses on relationships among the units of some population, and on how the structure of these ties affects outcomes experienced by both the units and the population. Often the units are persons or individuals, but they may be families, households, corporations, or nation states.
Social network analysis is a set of methods for describing, quantifying and analyzing the properties of social networks. This seminar is survey of statistical methods for analyzing social network data. It will focus on testing hypotheses about:
- network structure (e.g. reciprocity, transitivity, centralization),
- the formation/dissolution of ties based on attributes (e.g. homophily),
- local structural effects.
Starting April 8, we are offering this seminar as a 3-day synchronous*, remote workshop for the first time. Each day will consist of a 4-hour live lecture held via the free video-conferencing software Zoom. You are encouraged to join the lecture live, but will have the opportunity to view the recorded session later in the day if you are unable to attend at the scheduled time.
Each lecture session will conclude with a hands-on exercise reviewing the content covered, to be completed on your own. An additional session will be held Thursday and Friday afternoons as an “office hour”, where you can review the exercise results with the instructor and ask any questions.
*We understand that scheduling is difficult during this unpredictable time. If you prefer, you may take all or part of the course asynchronously. The video recordings will be made available within 24 hours of each session and will be accessible for two weeks after the seminar, meaning that you will get all of the class content and discussions even if you cannot participate synchronously.
MORE DETAILS ABOUT THE COURSE CONTENT
The course begins with models for the local structure of dyads and triads, and next moves to models based on the assumption of dyadic independence. We will then consider models that permit structured forms of dependence between dyads.
Topics include statistical models for local structure (dyads and triads) and graph-level indices, biased net models for complete networks and for aggregated tie count data, exponential random graph models, and stochastic actor-oriented models. Each day will be divided into a presentation of methods and a lab using those methods. This workshop assumes that you have already taken a first course in network analysis, or have acquired equivalent knowledge through self study.
All software uses the R environment as implemented in RStudio or directly in the standard console interface. Analysis relies on functions and packages developed for that environment including custom functions developed by the instructor and the packages: network, sna, statnet, and RSiena.
If you’d like to take this course but are concerned that you don’t know enough R, there are excellent on-line resources for learning the basics. Here are our recommendations.
Who should Register?
This workshop does assume that you have already taken a first course in network analysis (or have acquired equivalent knowledge through self study). You should be familiar with such network concepts as density, reciprocity, centrality and centralization, and clustering coefficients.
You should also be familiar with statistical concepts at the first-year undergraduate level, including such topics as basic probability distributions, hypothesis testing, t-tests, chi-square tests, OLS regression, and logistic regression.
1. Analysis of hypotheses about reciprocity, multiplexity, exchange, transitivity, density, degree, centralization, clustering coefficient and other graph‐level indices
a. Testing for effects like reciprocity, multiplexity, and exchange in dyads
b. Testing for effects like transitivity and closure in triads
c. Evaluating hypotheses about graph-level indices like density, mean
degree, centralization scores, clustering coefficients against
2. Biased net models for aggregate tie counts and complete networks
a. Attraction and repulsion mechanisms for homophily
b. Unidimensional and multidimensional analysis of intergroup ties
c. Complete network models: forms of dyadic dependence in models for
cross-sectional data and in models for longitudinal data
3. Exponential random graph models (ergm), their specification and estimation with statenet
a. Global and local form of an ergm
b. Model specification and estimation in statnet
c. Families of ergms from Bernoulli to social circuit models
d. Ergms for two-mode network data, for valued network data, and for
longitudinal network data
4. Stochastic actor‐oriented models (saoms) and RSiena
a. Modeling the rate function, the evaluation function, and the
b. Model specification and estimation in RSiena
c. Common effects in saoms for ties and behavior
“This course was really helpful because I had access to real R scripts that I can use and replicate for my own research. John understands every detail of complex network models and explains them in a very clear way.”
Seungho (Andy) Back, University of Toronto
“I thought this was a fantastic course. This course helped to clarify and expand upon what I had previously learned. Learning how to test if certain network structures are significantly different from chance was useful and getting a detailed explanation of how to interpret ERGM terms and ERGM model fit statistics was much better through this course than what I’ve previously tried to teach myself with textbooks.”
Megan Evans, Pennsylvania State University
“The material is very relevant to audiences outside of its origins in sociology and related fields. It demonstrates the presence of mathematical rigor in an aspect of studying relationships that allows researchers to go beyond descriptive statistics and pretty pictures.”
Steve Vejcik, TransUnion
“Very helpful for understanding advanced SNA. Nice instructor.”
Huiwen Xu, University of Rochester