The effect of differential recruitment, non-response and non-recruitment on estimators for respondent-driven sampling

摘要: Respondent-driven sampling is a widely-used network technique, designed to sample from hard-to-reach populations. Estimation the resulting samples an area of active research, with software available compute at least four estimators population proportion. Each estimator claimed address deficiencies in previous estimators, however those claims are often unsubstantiated. In this study we provide simulation-based comparison five existing focussing on conditions which recent address. We find no consistently out-performs all others, and highlight each be preferred.