Abstract

Previous research of KEM was able to show, that, when no person registers are available, existing sampling methods are problematic. Housing unit listing as well as random route samples are confronted with errors on the level of households selection and interviews. These sampling methods are influential used in German wide, Europe wide and internationall studies.
This study intends to develop and test survey procedures, which provide researches with reliable methods in comparable sampling situations. Sampling approaches for population surveys which do not require a register are investigated. The aim is to achive samples which are easy to implement and enable unbiased estimates. In particular building- and street section samples are investigated.
The sampling approaches are simulated in a realistic setting with real data. A measure of size (MoS), which determines the household density of a region, is developed, in order to reduce costs.