Assessment Through Simulation of Respondent-Driven Sampling Estimates When Used in Telephone Survey Sampling

• Activity Number: 252
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Jenny Yang*+ and Donglin Zeng and Robert Agans
• Companies: Carolina Survey Research Laboratory and The University of North Carolina at Chapel Hill and Carolina Survey Research Laboratory
• Keywords: respondent driven sampling ; simulation ; network ; cell phone ; telephone ; RDS

Abstract

Respondent-driven-sampling (RDS) has recently been utilized as a valid method in survey methodology for reaching hard-to-reach populations. However, the assessment of its efficacy in expanding sample size as well as its estimation methods has largely focused on the in-person implementation which includes an initial convenience seed. This implementation varies greatly from its sibling implementation in telephone survey sampling, which is most heavily reflected in the assumptions used for deriving estimates. Through the use of simulation to generate numerous different types of correlated networks, we compare RDS in a telephone setting in comparison to the in-person setting and assert that RDS is a robust sampling method. We also explore practical matters in regards to the implementation of RDS for telephone surveying such as the number of waves to reach equilibrium and obtaining an accurate parameter estimate.