Design-based estimators for snowball sampling
2010 (English)Conference paper (Refereed)
Snowball sampling, where existing study subjects recruit further subjects from amongtheir acquaintances, is a popular approach when sampling from hidden populations.Since people with many in-links are more likely to be selected, there will be a selectionbias in the samples obtained. In order to eliminate this bias, the sample data must beweighted. However, the exact selection probabilities are unknown for snowball samplesand need to be approximated in an appropriate way. This paper proposes differentways of approximating the selection probabilities and develops weighting techniquesusing the inverse of the selection probabilities. Some numerical examples for smallgraphs and simulations on larger networks are provided to compare the efficiencyof the weighting techniques. The simulation results indicate that the suggested re-weighted estimators should be preferred to traditional estimators with equal sampleweights for the initial snowball sampling waves.
Place, publisher, year, edition, pages
2010. 23-27 p.
Research subject Statistics
IdentifiersURN: urn:nbn:se:su:diva-88948OAI: oai:DiVA.org:su-88948DiVA: diva2:614939
Workshop on Survey Sampling Theory and Methodology August 23-27, 2010 Vilnius, Lithuania