Sampling

Sampling is generally used to gather data on a population. In order to statistically be able to infer something about the entire population from the elements in the sample, the sample elements must be drawn with known positive probability. Good estimates are based on the underlying selection procedures.

GESIS statisticians primarily research the topics of sampling and estimators and thus offer years of experience and unique expertise. This is clearly seen in the development and providing of telephone sampling of landlines and mobile networks. The Gabler-Häder method has meanwhile become the standard for landline sampling. The combination of landline and mobile phone sampling is a must and challenge for telephone surveys to be able to best cover the entire population. Current research is focused on developing adequate selection frames and weighting techniques.

Planning and designing of samples has ever increasing praxis relevance for international comparative studies and this is an area where GESIS has years of demonstrable project and research experience as part of large-scale projects such as ESS, SHARE und PIAAC. The goal of this research is to develop statistical comparison criteria for the quality of such surveys utilizing design effects, whereby the inclusion of models for non-response in the estimates continues to gain significance.