摘要：With sensitive subjects there is the risc of a high nonresponse- or
untruthful answers rate with direct questioning on this matter. In this case randomized
response techniques are an alternative for an unbiased estimation of
the parameters under study. The price that has to be paid for this is an increase
of the variance of the estimators. In this paper for two of such estimators
for proportions the variances are derived for simple random sampling without
replacement. Furthermore by a simulation study the sampling distribution of
various estimators, that are based on such questioning designs are compared
and graphically presented. The results give information about the quality of
altogether 24 estimators, their differences and limitations.
关键词：Datenqualit¡§at, Stichprobenerhebung, Randomisierte Antworttechniken,
Nonresponse, Ziehen ohne Zur¡§ucklegen, Effizienz.
