EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Respondent-Driven Sampling Model Evaluation for Sampling without Replacement in Estimating Hidden Populations

Anjikwi Yakubu, Danjuma Jibasen, Ikeme John, Emmanuel Torsen

International Journal of Development Mathematics (IJDM) · 28.03.2026 · 3 (1) · 193-209

DOI: 10.62054/ijdm/0301.16 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Development Mathematics (IJDM), online erschienen am 28.03.2026. Verfasst von Anjikwi Yakubu und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study evaluates respondent-driven sampling (RDS) models that use sampling without replacement to estimate characteristics of hidden populations. Traditional RDS estimators, such as Salganik and Heckathorn (SH-RDS) and Volz and Heckathorn (VH-RDS), assume sampling with replacement and require many recruitment waves to reach statistical equilibrium, which is rarely achieved in practice. Most real-world RDS studies are conducted without replacement and use fewer waves, leading to biases like overrepresentation of highly connected individuals. Recent estimators, such as Gile’s Successive Sampling (G-SS), address some limitations but still face challenges, including instability with large samples, broad confidence intervals, and inadequate handling of non-random recruitment and seed selection. To address these issues, a new estimator is proposed that incorporates sampling without replacement and strategic multiple-seed selection. Simulations and real-world data analysis (using the Project 90 dataset) demonstrate that estimator performance varies by sample size. For small samples SH-RDS and VH-RDS are most accurate for gender estimation. For larger samples (, the proposed estimator is most efficient, with minimal variance. G-SS shows moderate, reliable performance, while the Naïve estimator becomes less reliable as the sample size increases. Analysis also reveals that the proposed estimator performs well for groups with higher connectivity, though variance remains high for the “Unemployed” group. Overall, the proposed estimator was recommended for large samples and complex networks, especially among hard-to-reach populations.

Zitieren

APA

Yakubu, A., Jibasen, D., John, I., & Torsen, E. (2026). Respondent-Driven Sampling Model Evaluation for Sampling without Replacement in Estimating Hidden Populations. International Journal of Development Mathematics (IJDM), 3(1), 193-209. https://doi.org/10.62054/ijdm/0301.16

Vancouver

Yakubu A, Jibasen D, John I, Torsen E. Respondent-Driven Sampling Model Evaluation for Sampling without Replacement in Estimating Hidden Populations. International Journal of Development Mathematics (IJDM). 2026;3(1):193-209. doi:10.62054/ijdm/0301.16

BibTeX
@article{Yakubu2026,
  author = {Yakubu, Anjikwi and Jibasen, Danjuma and John, Ikeme and Torsen, Emmanuel},
  title = {Respondent-Driven Sampling Model Evaluation for Sampling without Replacement in Estimating Hidden Populations},
  journal = {International Journal of Development Mathematics (IJDM)},
  year = {2026},
  volume = {3},
  number = {1},
  pages = {193--209},
  doi = {10.62054/ijdm/0301.16}
}
Kostenlos starten →