Model-Based Estimation with Link-Tracing Sampling Designs

 by Stephen Thompson
Pennsylvania State university

Ove Frank
Stockholm University

 Research Report 1998:1

 Department of Statistics, Stockholm University, S-106 91 Stockholm, Sweden


Samples from hidden and hard-to-access human populations are often obtained by procedures in which social links are followed from one respondent to another. Inference from the sampling to the larger population of interest can be affected by the link-tracing design and the type of data it produces. The population with its social network structure can be modeled as a stochastic graph with a joint distribution of node values representing characteristics of individuals and arc indicators representing social relationships between individuals. In this paper maximum likelihood estimators of population graph parameters and predictors of realized population graph quantities are described. Differences between these estimators and conventional data summaries such as sample means, subgraph means, and expansion predictors are noted.

 Key words: Snowball Samples, Adaptive Sampling, Graph Sampling, Ignorable Designs, Link-Tracing Designs, Network Sampling, Likelihood, Predictive Likelihood. 

Close this Window

Last update: 1998-02-05 / MC