Statistics – Methodology
Scientific paper
2012-01-09
Statistics
Methodology
20 pages, 1 table, 5 figures, 1 supporting information file
Scientific paper
Respondent-driven sampling (RDS) is a commonly used substitute for random sampling when studying hidden populations, such as injective drug users or men who have sex with men, for which no sampling frame is known. The method works like a snowball sample but can, given that some assumptions are met, generate unbiased population estimates. One key assumption, not likely to be met, is that the acquaintance network in which the recruitment process takes place is undirected, meaning that all recruiters should have the potential to be recruited by the person they recruit. Here we investigate the potential bias of directedness by simulating RDS on real and artificial network structures. We show that directedness is likely to generate bias that cannot be compensated for unless the sampled individuals know how many potentially may have recruited them (i.e. their indegree), which is unlikely in most situations. We propose three indegree-based estimators for RDS on directed networks, and show that they can be used in the situation when only outdegrees are observed, either together with prior information or in a sensitivity analysis taking uncertainty of indegree properties of the network into account.
Britton Tom
Liljeros Fredrik
Lü Xiancai
Malmros Jens
No associations
LandOfFree
Respondent-driven Sampling on Directed Networks does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Respondent-driven Sampling on Directed Networks, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Respondent-driven Sampling on Directed Networks will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-461403