Given the increasing bias in random digit dial sampling, could respondent-driven sampling be a practical alternative?

Increasing cellular phone service and nonresponse are causing random digit dial (RDD) users to search for alternative ways of sampling geographically large populations. This study evaluated the feasibility and utility in using a modified version of respondent-driven sampling (RDS) as an alternative method.

Using RDS, 469 residents of Albany County, New York were enrolled into a telephone-based health survey. Participants answered health and RDS feasibility questions. Results were compared to a previously collected RDD sample and census data.

Participation was high (81.4%) and participants referred at least one peer 65.9% of the time. The RDS method produced a more ethnically diverse sample, otherwise respondent demographics were similar to the RDD sample. The most common reason for participating (51.9%) was because a peer told them about the study; 44.9% would not have participated in an RDD study. Persons not willing to participate in a RDD study reported being less healthy and less likely to participate in healthy activities (e.g., have a physical exam in the past 24 months).

Although more research is needed, RDS methods may be developed into a viable alternative for collecting health data from large general populations.