"The paper discusses techniques for sampling the 'migrant background' population in Germany, which comprises all first-generation immigrants, all non-citizens born in Germany, and all children with at least one parent fulfilling one of these criteria. Random walk sampling and random digit dialing techniques are feasible for sampling this population as a whole, but inefficient for subgroups. Telephone directories provide biased representations of the population, and the large proportion of non-pubs disqualifies their use. The Central Register of Foreigners excludes naturalized immigrants and introduces a socio-economic bias toward the less successful. Snowballing overrepresents persons with larger ethnic networks. The center sampling technique may encounter particular problems in Germany due to settlement patterns and legal issues affecting certain immigrants. Local authority Population Registers provide the best representation of the population. Foreign citizenship fails to identify the target population as it largely underestimates numbers and distorts the social structure. Place of birth is a suitable criterion to identify the Aussiedler population (ethnic German immigrants from eastern Europe and the former Soviet Union). In most cases, however, foreign names best serve the purpose of unbiased sampling. Therefore, name-based sampling in the Population Registers is the method of choice. However, the decentralized administration of Population Registers makes this a costly endeavor and although there is a certain legal sampling interface, there are still legal obstacles to optimal implementation of this sampling procedure." (author's abstract)
展开▼
