Sampling of Hard-to-Reach Populations
Christian Hoops, Ipsos Public Affairs and Prof. Dr. Kai-Uwe Schnapp, Universität of Hamburg, in Proceedings of the European Survey Research Association Conference, 07/2013
A presentation by Christian Hoops to the European Survey Research Association Conference held in Ljublijana shows different ways to conduct survey studies with hard-to-reach populations. First he describes strategies that can be used to sample populations with low incidence. On the one hand classical sampling methods like random digit dialing or random route are too expensive, on the other hand alternative sampling methods can only generate samples of unknown composition. For example, snowball techniques, onomastic procedures or facility-based surveillance. So the results of such surveys are often biased due to the methodology.
With the aim to get a comparison of these methods Hoops defines the sample space which consists of all people with target characteristics who could be included in the sample. So the established survey methods can be described by the cardinality of the sample space and by the variance of the inclusion probabilities. At the end he suggests to analyse each sample problem individually to find the best survey method.
More about the conference program you will find here.