230164 UE Sampling design, practical allocation procedures and Weighting (2009S)

This lecture deals with practical problems in the design and implementation of a survey: How to draw a representative sample? Why can a sample estimate depart from known distributions ? How can better coherence be achieved? Which role plays weighting for the precision of statistical tests. Students are invited to present their own research problems which may be addressed in particular during the course (for organizational reasons advance notice is recommended).

The introduction deals in particular with probability sampling. In practice several sample designs may be distinguished according to expected bias, precision and cost of data collection. In some cases data need to be weighted in order to achieve unbiased estimates.