Statistical Tests of the Assumptions behind Respondent Driven Sampling in a Survey of Female Sex Workers in Shanghai

William Whipple Neely, University of Wisconsin at Madison
Giovanna Merli, Duke University
Feng Tian, University of Wisconsin at Madison

We carry out an empirical assessment of the assumptions underlying Respondent Driven Sampling (RDS) in the context of a survey of female sex workers (FSWs) in Shanghai. By combining knowledge of the social structure of the target population with probability models implicit in the RDS methodology, we construct explicit hypothesis tests for the appropriateness of the assumptions behind the RDS methodology. This work is a component of a larger project titled “Sexual Behavior, Sexual Networks and STDs in China,” which focuses on the social and behavioral determinants of HIV and other STDs in China, given a background of rapidly changing social norms and behaviors, a rising incidence of STDs and concerns that HIV will soon spread from high-risk groups to the general population.