Web Survey Bibliography
In Official Statistics, household and personal surveys gain increasingly in importance. A survey of great significance is for example EU-SILC, a study about poverty, social exclusion and living conditions in Europe. The sampling frame in Germany for this survey is the Official Statistics‟ Access Panel - called the ”Dauerstichprobe” (DSP) - that is conducted by the German Federal Statistical Office (DESTATIS). One problem in sampling from the AP is the voluntary participation and the resulting self-selection in the recruitment process. Thus, certain groups in the population (such as entrepreneurs, pensioner, etc.) are less willing to participate.
As a result of possibly low response propensities, one would expect biased and inefficient point and variance estimates. For that reason, it is important to take these propensities into consideration and investigate the influence of the response rate (i.e. the rate itself as well as the response pattern) on the given estimates.
The present paper will therefore address point and variance estimation of means and total values (e.g. of income) in different access panel participation scenarios using different estimation strategies. Based on the latest estimation results from the partner‟s response propensity estimates (cf. Rendtel/Amarov, FU Berlin) and on earlier work of the DACSEIS project (cf. http://www.dacseis.de), several scenarios will be implemented and examined within a Monte-Carlo simulation study.
Conference homepage (abstract)