Integrating migrants into representative cross-sectional and longitudinal survey designs 1 |
|
| Chair | Professor Jürgen Schupp (SOEP/DIW Berlin ) |
| Coordinator 1 | Professor Lucinda Platt (LSE) |
| Coordinator 2 | Professor Narayan Sastry (ISR/Univ Michigan) |
Germany experienced an influx of estimated 890,000 refugees seeking protection in 2015 and an additional 210,000 by the end of September 2016. This poses major challenges to policymakers, administrators, and actors in civil society. For these reasons, the Institute for Employment Research (IAB), the research center of the Federal Office for Migration and Refugees (BAMF-FZ), and the German Socio-Economic Panel (SOEP) have forged a partnership to establish a household panel study on recent refugee migration to Germany: The IAB-BAMF-SOEP Refugee Survey. The longitudinal sample includes about 3,200 families. About 4,500 adult respondents initially participated in personal interviews of wave 1 between June and December 2016.
Researchers who plan to sample refugees and asylum seekers in Germany as well as many other countries are regularly facing problems of accessing appropriate sampling frames. For this project, we were able to use the Central Register of Foreigners held by the German Federal Office for Migration and Refugees (BAMF) as a sampling frame. The register is restricted to documented migration and contains information about foreign nationals including refugees and asylum seekers who are currently living in Germany.
This presentation focuses on the sampling design of the IAB-BAMF-SOEP Refugee Survey and provides insights on the practical implementation of a large-scale sample of refugees in Germany. First, we will discuss advantages and disadvantages of using the Central Register of Foreigners as a sampling frame. Second, we provide detailed information on the various stages in the sampling process, starting with the clustering of the register data into primary sampling units and ending with the stratified sampling of individuals within each sampled cluster. Finally, we will examine issues related to design effects, nonresponse adjustment, and post-stratification.
The panel study “Labour Market and Social Security” (PASS), established by the Institute for Employment Research, is a central data source for research on unemployment, poverty and the welfare state in Germany. Since 2006, we annually interview approximately 12.000 persons in more than 8.000 households. The study’s design particularly enables evaluating the situation of recipients of basic income support, called unemployment benefits II (UB II) in Germany. An additional sample of the residential population enables researchers to make comparisons between the two groups and answer questions beyond a pure welfare focus.
The current migration movements to Germany permanently change the structure of UB II recipients: about 400.000 recognised refugees and asylum seekers are entitled to welfare benefits in 2016. To ensure the continued provision of representative and well-grounded data on UB II in Germany, we implemented different strategies to integrate this special migrant population successfully into PASS.
In our annual refreshment samples of new entrants into the UB II system, we oversampled Syrian and Iraqi refugees in the most recent wave. Together they compromise the largest groups of refugees to Germany in the recent years and speak predominantly Arabic. We drew special refugee samples from the administrative data of the German Federal Employment Agency. With the help of native speaking researchers, we translated our questionnaire programme to Arabic and designed special modules for the migrant population. We closely coordinated the fieldwork with our survey institute to ensure high quality of the interviews with the migrant population in CATI and CAPI modes. Thus, we could realise participation rates and or more than one third and about 800 interviews in 2016.
Our special migrant population is on average since 1.5 years in Germany, with recognised asylum status and entitlement to UB II. Thus, we can provide insights for a group whose integration into labour market and society is of great interest to politics. The dual sampling structure of PASS allows comparisons between the refugee sample and immigrants who have been living here for longer already, the UB II recipients, or the general population.
We offer to present our approach to integrating the new migrant population into an existing panel study and want to share our experience. After a short general overview of PASS we would like to focus on the challenges that integration of a new migration population in a running panel study poses to survey management. We would like to present our different strategies, with particular reference to sampling design and fieldwork organization (e.g. interviewer training, tracking techniques and follow-up rules) in order to ensure a high response among the new migration sample of PASS. Furthermore, we would like to outline our questionnaire design and the research potentials, which PASS provides about migration, labour market and welfare benefit.
With the influx of refugees in Germany during the last two years the population for the household panel survey PASS has changed. PASS is an annual household panel survey for labour market, welfare state, and poverty research in Germany. The sample for PASS is drawn out of all recipients of social benefits from the Federal Employment Agency’s register. During the last 10 years of PASS migrants from Russia and Turkey participated in the panel. With the new influx of refugees Arabic people with cultural particularities need to be integrated, too. In 2016 – PASS wave 10 – we were faced with several challenges to meet the needs of this new population: e.g. preparing questionnaires in the Arabic language, providing Arabic speaking interviewers, contacting refugees in residential accommodation, finding the right ways to attain contact, conducting a standardised PASS interview, identifying persons living together in households.
With our paper we will present some first experiences with survey management and fieldwork strategies we made in said wave 10. PASS incorporates a full set of field strategies to meet the high quality requirements of data collection measurements (e.g. tracking measures, mixed mode, refusal conversion, incentives). In wave 10 a subsample of Syrians and Iraqis was drawn as refresher sample. With the mixed mode design in PASS we realised CATI and CAPI interviews. Switches from one method to the other were allowed as well as changing the interview language. We allowed different settings during interviews in the CAPI field such as support through third parties or partly self-administration of the questionnaire. The strategies applied allow for comparing the results and giving insight in the cultural particularities for measurement instruments.
During the next PASS waves we are going to optimise measures and strategies for data collection. Mainly the integration of refugees as panel members over several waves will be a great challenge for the future. First data from tracking procedures (letter between waves) will be available.
The increased influx of migrants into European countries presents a huge challenge both to official statistical bodies, survey researchers, and academics. The need for more comprehen-sive and reliable data has been acknowledged both by the European Union (see, e.g., the EU regulation (EC) 862/2007 on community statistics on migration and international protection or the Declaration of Zaragoza (2010) stressing the need to have common indicators on migrants) and international statistical bodies such as the United Nations, the OECD, and the World Bank (see, e.g., the Task Force on Improving Migration and Migrant Data Using Surveys and Other Data, also referred to as the ‘Suitland Working Group’). In response to these initiatives many European countries have made efforts to improve their national statistical infrastructure on migrants. However, a major issue in migration research is the question of how these new populations integrate into their destination countries and how the integration process can be managed by integration policies. Therefore, it is of great interest both for academic researchers and policymakers how different societal and political contexts facilitate or hamper this integration process, which on top of the aforementioned information needs calls for more data on migrants that is comparable across different national contexts. Based on an expert study in nine western European countries (Denmark, Germany, Italy, Netherlands, Romania, Poland, Spain, Sweden, and United Kingdom), this talk will discuss the possibilities and restrictions of a comparative sampling design in these countries.