Answers are grounded in this paper and cite their sources.
Respondent-driven sampling (RDS), traditionally used for sampling rare populations, presents challenges when applied to larger and more heterogeneous subgroups within the general population. Unlike traditional probability sampling, RDS is network-based, with sample size and characteristics commonly unknown until fieldwork concludes. Estimating sample convergence and efficiency becomes challenging when surveying large subpopulations with only a small fraction sampled, especially in studies with multiple analytical focuses and different recruitment patterns. This study, based on a survey of domestic workers in three Chinese cities (2019–2020), addresses these challenges through improved convergence tracking, estimator evaluation, and efficiency assessment. Monitoring sample distribution changes by participation time rather than recruitment wave is important, especially when small subgroups exist. Viewing homophily as intraclass correlation provides a new approach that captures complex recruitment patterns and enables variance estimation across more software tools. Design effects and coefficients of variation are useful metrics for evaluating sample efficiency and precision.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: