The SOEP data comprises a sizeable number of immigrants to Germany and their descendants. Several user-friendly variables identify these groups (GERMBORN, CORIGIN, IMMIYEAR, MIGBACK) and thus give information on the migration background of all persons who have ever been a part of a SOEP household (i.e., the population from PPFAD, PPATH or PPATHL). In detail, GERMBORN and CORIGIN give information on the country of birth, with the exception of persons who immigrated to Germany before 1950 who are considered to have been born in Germany (the Federal Republic of Germany was founded in 1949). IMMIYEAR specifies the last year of immigration to the Federal Republic of Germany for all persons considered not born in Germany, and MIGBACK is useful to identify immigrant descendants by combining information on respondents and their (grand-)parents. In addition, GERMBORNINFO, CORIGININFO, IMMIYEARINFO and MIGINFO indicate the quality of information given in GERMBORN, CORIGIN, IMMIYEAR and MIGBACK, respectively.

All SOEP samples include immigrants to Germany and their descendants. The shares vary, however, across samples depending on the target population covered. Naturally, samples covering the entire residential population in Germany (Sample A, E, F, G, H, I, J, K, L1, L2, L3 and N) or specific groups such as persons from the former GDR (Sample C) contain a smaller number of immigrants and their descendants than the samples of foreigners and migrants (Sample B, D, M1, M2, M3, M4 and M5) or the sample of households in urban areas (Sample O).

Information for GERMBORN, CORIGIN, IMMIYEAR and MIGBACK and the respective INFO variables is collected primarily from the individual questionnaires (dataset PL) or the variations of the “biography / life history” questionnaires (integrated biographical data files and life-course information in dataset (BIOL) and from the additional youth questionnaire for 16-17-year-olds, in use since 2000 (dataset JUGENDL). In addition, information from the electronic household protocol for M1 (2013) and the retrospective survey (2012) of early childhood in the context of war (dataset BCBFK) was used (both datasets are not included in the standard data distribution).

GERMBORN specifies whether a person was born in Germany or in another country. Persons who immigrated to Germany before 1950 are considered as being born in Germany (the Federal Republic of Germany was founded in 1949; see also IMMIYEAR). To code GERMBORN, all relevant information (see Table 1: Information used for GERMBORN, CORIGIN, IMMIYEAR and MIGBACK) available on persons who have ever been a part of a SOEP household (i.e., the population from PPFAD, PPATH or PPATHL) was combined. The vast majority of persons who have ever been part of a SOEP household gave consistent information on their country of birth and GERMBORN was coded accordingly to the respondents’ answers. For part of the population, no direct information on the person’s country of birth was available. For both, persons for whom “(2) inconsistent information” or “(3) no information” (GERMBORNINFO) was available, additional indicators were used to code the GERMBORN values. In this process, information on a respondent’s citizenship and their parents’ migration biography were used. We coded the values on GERMBORN in the following order (with descending priority):

1. First, mothers’ immigration history and their place of residence at the time of the respondents’ birth were taken into account to determine the respondents’ probable country of birth. For instance, when a respondent was born after or in the year of their mother’s immigration to Germany, the respondent is considered to have been born in Germany. For the coding of a few cases, more detailed information on respondents’ month of birth and mother’s immigration month was available and used. When a mother’s immigration year was missing, the father’s immigration history was used to code a respondent’s country of birth.
2. In the next step, GERMBORN was coded for the remaining “(2) inconsistent information” cases. Respondents’ information on their country of birth, their citizenship, and parental information was taken into account to identify a respondents’ country of birth. The mode was calculated for inconsistent information on respondents’ and parental country of birth. In case of varying modes, higher values were given a preference when coding, to be more sensible to foreign countries of birth. For instance, a respondent who reported being born in Germany more often than being born abroad (country of birth), who had German citizenship (citizenship), and whose parents reported more often to be born in Germany than being born abroad (parental information) was considered to have been born in Germany.
3. In a last step, GERMBORN was coded for the remaining “(3) no information” cases. Respondents’ citizenship and parental information was used to approximate their most likely country of birth. By definition, information on their country of birth was missing. The mode of parents’ country of birth and citizenship was used for the coding of GERMBORN, too. For instance, respondents with German citizenship whose parents reported more often to be born in Germany than being born abroad were coded as being born in Germany.

Table 1: Information used for GERMBORN, CORIGIN, IMMIYEAR and MIGBACK

Source: v35