Degrees of Uncertainty with Uncertain Degrees in Respondent-Driven Sampling

• Activity Number: 66
• Type: Topic Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Joseph Blitzstein*+
• Companies: Harvard University
• Keywords: social network analysis ; networks ; respondent-driven ; network sampling

Abstract

Respondent-driven sampling (RDS) is a contact-tracing sampling design (introduced by Heckathorn) which is widely and increasingly being used to study hard-to-reach populations such as injection drug users and individuals at high risk of HIV infection. We compare model-based vs. design-based estimators in this setting, and study the bias-variance tradeoffs which result depending on what assumptions one is willing to make about the population of interest and about the network structure.