![]() |
![]() |
Jichuan Wang, PhD1, Robert Carlson, PhD1, Russel S. Falck2, and Harvey A. Siegal2. (1) Community Health, Wright State University, 3640 Colonel Glenn Hwy., Dayton, OH 45435, 937 775-2084, jichuan.wang@wright.edu, (2) Community Health Department, Wright State University, 3640 Colonel Glenn Hwy, Dayton, OH 45435
A new form of chain-referral sampling, Respondent Driven Sampling (RDS), has been recently developed by Heckathorn to sample hidden populations. RDS is designed to eliminate the sources of bias caused by the non-random selection of an initial sample, as well as to reduce biases due to volunteerism and masking, in the regular chain-referral sampling. RDS has been successfully applied to studying hidden populations, such as drug users with HIV infection and injection drug users. In this study, RDS was applied to sample young adults who had recently used MDMA. The sampling began with 16 MDMA users, who served as “seeds” to initiate recruitment chains. A total of 402 young adults, age 18-30, were recruited. Dual incentives and recruitment coupons were employed in the sampling process. Analysis of the sample showed that sample compositions converged/stabilized within a limited number of recruitment waves, and the ultimate sample compositions were independent of the characteristics of the initial sample. In this study, compositions of the targeted population were estimated and compared with the sample compositions. Bias due to volunteerism and masking was assessed, and the social structures in which respondents were embedded were analyzed, using the “by-product" information provided by the RDS. In addition, advantages, assumptions, and limitations of the RDS were discussed.
Learning Objectives:
Keywords: Statistics, Survey
Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.