Network Dependencies in Social Space, Geographical Space, and Temporal Space

This page has material for the workshop at NetGlow 2022 that will give you hands-on experience in fitting a stochastic actor-oriented model to longitudinal networks using RSiena in R. The workshop is divided into two parts.

Part I

The first part is based around a simple example of longitudinal network and aims to illustrate how to calibrate (estimate) the SAOM so that the simulation model reproduces the second network starting in the first. No prior experience of R is required and the workshop will walk you through pre-prepared code. Basic knowledge about SNA and a laptop with R and the packages RSiena, sna, and network (instructions for installation are provided on the RSiena webpage).

Download R script from here and a reformatting script from here

In the workshop you will have the R script open in R (or RStudio) and we will go through the code, line by line. At one point (line 35 in the script) you will need to use the function reshapeRSienaDeps() to plot simulated networks. To do that open and execute the reformatting script.

The hands-on exercise will be accompanied by these introductory slides.

For a very basic to working with networks in R see Minimal intro to SNA in R part 1, part 2, and part 3.

Background reading

Snijders, T.A.B., van de Bunt, G.G., and Steglich, C.E.G. (2010). Introduction to actor-based models for network dynamics. Social Networks, 32, 44-60.

Tom A.B. Snijders (2005). Models for Longitudinal Network Data.
Chapter 11 in P. Carrington, J. Scott, & S. Wasserman (Eds.), Models and methods in social network analysis. New York: Cambridge University Press, pp. 215-247.

Part II

The second part of the workshop provides an overview to extensions of SAOM and other advanced issues. In Part I we will have built some basic experience of using the standard selection and influence models, including running standard diagnostics. We will cover some recent extensions, such as analysis of multiplex ties, the joint evolution of one-mode and two-mode networks, influence models for continuous outcomes, and multilevel analysis using multigroup and sienaBayes. A focus will be put on data formats and some key interpretative issues in working with the output. We will draw on some example code from the RSiena webpage and a laptop with R and the packages RSiena, sna, and network is recommended for following the workshop.

The slides are available here.

R scripts, other didactic material, and literature on stochastic actor-oriented models can be found on the RSiena webpage. Comprehensive documentation is provided in the RSiena Manual.