Searcharxiv⌕ Search

arXiv subjects

Milada Dvorakova

Publications and source records attributed to Milada Dvorakova.

2 recordsLinked to original sources

Accounting for Attrition in Respondent-Driven Sampling Cohort Studies

Respondent-driven sampling (RDS) is widely used to recruit populations lacking conventional sampling frames. Because RDS oversamples well-connected individuals, weights are commonly applied to improve representativeness. Although methods for cross-sectional RDS analyses are well developed, guidance for longitudinal studies is limited, despite additional selection bias from missed visits and dropout. We develop a framework that jointly accounts for unequal inclusion probabilities at recruitment and attrition during follow-up. The approach combines baseline RDS weights with inverse-probability-of-non-response weights estimated from observed follow-up histories, producing visit-specific weights for estimating prevalences, means, and regression parameters. We clarify when the RDS design should inform attendance models rather than being addressed only during outcome analysis. Using simulations of RDS recruitment, we assess prevalence and standardized prevalence differences under varying recruitment and attrition mechanisms, focusing on bias, variance, and confidence interval coverage. Weighting the response model by baseline RDS weights was important when an unmeasured determinant of degree also predicted response, even when it did not predict the outcome. We illustrate the framework using data from the Montreal site of Engage, a multisite RDS cohort of sexual minority men in three Canadian cities, from 2017-2023. Follow-up involved intermittent missed visits, attrition, and terminal dropout that was more common among participants with smaller networks. Our findings provide practical guidance for combining design-based and attrition adjustments to support valid inference in longitudinal RDS studies.

stat.AP↗

Estimating hidden population size from a single respondent-driven sampling survey

This work is concerned with the estimation of hard-to-reach population sizes using a single respondent-driven sampling (RDS) survey, a variant of chain-referral sampling that leverages social relationships to reach members of a hidden population. The popularity of RDS as a standard approach for surveying hidden populations brings theoretical and methodological challenges regarding the estimation of population sizes, mainly for public health purposes. This paper proposes a frequentist, model-based framework for estimating the size of a hidden population using a network-based approach. An optimization algorithm is proposed for obtaining the identification region of the target parameter when model assumptions are violated. We characterize the asymptotic behavior of our proposed methodology and assess its finite sample performance under departures from model assumptions.

stat.ME↗