Back to the homepage of Tom Snijders

## SOCIAL NETWORK ANALYSIS

General information about social network analysis can be found at the web site of INSNA, the International Network of Social Network Analysis.

Note the following links:

You can here get some information about the following statistical methods for social network analysis, and download the associated computer programs:

1. program SIENA, Simulation Investigation for Empirical Network Analysis, for the statistical analysis of the evolution of entire social networks, based on data consisting of two or more repeated measurements of the network; this program is based on Snijders (2001);.
2. program BLOCKS, for stochastic block modeling for undirected, directed and valued graphs.
3. program ZO, for analysis of graphs with given degrees, or if you wish, 0-1 matrices with given marginals;
4. program SNOWBALL, for estimation of the size of a (hidden) population from a one-wave snowball sample.

The program BLOCKS is a program for stochastic block modeling, based on Snijders and Nowicki (1997) and Nowicki and Snijders (JASA, 2001). The method is based on Gibbs sampling, which is one of the many methods of Markov chain Monte Carlo. Therefore it is rather time-consuming.

This program can be used for undirected as well as directed graphs, but also for undirected or directed valued graphs (where you could think of 3 to 6 values).

The program is written in Delphi, for use under Windows. The current version is 1.8 (June 2007).
BLOCKS can be executed from the StOCNET environment, but also as a stand-alone program.

This method was applied, e.g., in

ZO (for Zero-One) is a collection of programs for the analysis of 0-1 matrices with given marginal sums.
Special attention is given to matrices with a structurally zero diagonal, which represent graphs (if symmetric) or directed graphs (if not symmetric).
These programs can be used for the statistical analysis of graphs and directed graphs according to the distributions known in the social network literature as the U|{Xi+} , the U|{Xi+},{X+i} , and the U|{Xi+},{X+i},M distributions.

The programs in ZO implement the methods published in the articles

For related articles, also see
• An application:
Bonacich, Ph., Oliver, A., and Snijders, T.A.B., Controlling for size in centrality scores.
Social Networks, 20 (1998), 135-141.
• Another simulation method:
A.R. Rao, R. Jana, and S. Bandyopadhyay (1996). A Markov chain Monte Carlo method for generating random (0,1) matrices with given marginals.
Sankhya ser. A 58, 225-242.
• A review of some simulation methods:
John M. Roberts, jr. (2000). Simple methods for simulating sociomatrices with given marginal totals.
Social Networks 22, 273-283.

The program SNOWBALL computes estimates of the size of a population (this method is intended to be used for so-called hidden populations) according to the methods proposed in
O. Frank & T.A.B. Snijders (1994). Estimating the size of hidden populations using snowball sampling
Journal of Official Statistics 10, 53-67.

The program is written in Turbo Pascal, and was compiled using Delphi.

You can download a zipped collection of files, which can be unzipped using PKUNZIP or WINZIP.

The included text file README.TXT, gives the basic information for executing this program.