Evaluating an Alternative Frame for Address-Based Sampling in Germany: The Address Database From Deutsche Post Direkt

Abstract

In Germany, the population registers with addresses of individuals can be used for address-based sampling. However, unlike countries with a centralized register, municipalities in Germany administer their registers themselves. This not only makes sampling for a nation­wide survey more costly and cumbersome but may also result in gaps in the gross sample, as selected municipalities may refuse to allow their registers to be used for sampling pur­poses. If substitute municipalities are not available, other sampling methods are required. The present study tested the feasibility of using the address database from Deutsche Post Direkt (ADB-DPD) as an alternative frame for address-based sampling in Germany. We simultaneously conducted two almost identical surveys in the German city of Mannheim with gross samples of equal size (N = 3,000). One sample was drawn from the city’s popu­lation register, the other from the commercial ADB-DPD. Our findings suggest that the ADB-DPD performs well both in terms of survey response and up-to-dateness. Due to relatively low costs and the fast provision of addresses, the ADB-DPD could be particu­larly attractive for survey projects with limited budgets and tight schedules. However, these benefits come at considerable cost. First, the use of the ADB-DPD is limited to self-admin­istered surveys. More importantly, in the net sample of the DPD survey, women and young persons were considerably underrepresented. This indicates coverage issues about which DPD provided no further information. Based on our analyses, we offer practical insights into the feasibility of using the ADB-DPD for sampling purposes and suggest avenues for future research.