Online Program

Network Sampling Coverage in RDS: How Much of the Network Do We See?

James Moody, Duke University 
M. Giovanna Merli, Duke University 
Sharon Weir, University of North Carolina 
Jing Li, China National Center for STD Control 
*Jeff Smith, Duke University 

Keywords: hard to reach populations; hiv/aids; sampling; social networks; RDS

This paper explores the network coverage of an RDS on commercial sex workers (CSW). Using unique information on the social networks of CSW obtained from two overlapping studies of CSWs, RDS and a venue based sampling approach (PLACE), we use new network estimation tools (Smith, 2011), based on exponential random graph models from local networks, to construct the likely network from which our observed RDS has been drawn. We then use this to estimate population coverage rates by simulating chains consistent with the observed RDS chains. We discuss the results in light of (a) potential estimation improvements implicit in network information, (b) strategies for improving the coverage rates, and (c) multiple sources of potential variability in coverage.