Tuesday 16th July
Wednesday 17th July
Thursday 18th July
Friday 19th July
Download the conference book
Surveying immigrants in the absence of a sample frame 1 |
|
| Convenor | Dr Yana Leontiyeva (Institute of Sociology, Academy of Sciences of the Czech Republic) |
| Coordinator 1 | Dr Agata Gorny (Centre of Migration Research, University of Warsaw) |
| Coordinator 2 | Ms Joanna Napierala (Centre of Migration Research, University of Warsaw) |
The aim of the proposed session is to attract contributions that address the methodological challenges posed by surveying respondents with immigrant-backgrounds, especially when an appropriate sample frame is not available. Conceptualising the sampling immigrants as a hard-to-reach population is relevant for (a) large national surveys that often underrepresent immigrants and (b) specific immigrant oriented surveys that often fail to produce data suitable for sophisticated statistical analysis.
The session welcomes contributions that address the following topics:
1) Use of innovative sampling methods like respondent-driven sampling, time-space sampling, quota sampling, random route sampling with focused enumeration, onomastic method in migration research.
2) Evaluations of survey data quality for immigrants in representative national and international surveys.
3) Presentation of particular immigrant surveys with a special focus on sampling techniques.
There are many strategies that can be used for sampling hard-to-reach populations. For example, snowball techniques, onomastic procedures or facility-based surveillance. But the results of such surveys are often biased due to the methodology.
This presentation shows a new approach how samples can be drawn by using gravity analytical methods. In our example we used probabilistic models developed by Huff (1964) and Nakanishi & Cooper (1974) for estimating the probability of belonging to the national minorities in the border region between Germany and Denmark.
Therefore the distance to the Danish-German border, the election results of the party for the minority and the densities of national typical surnames and organisations are taken into account. A projection on the desired sample size calculates the needed number of respondents for each location.
Literature:
Huff, D. L., 1964: Defining and Estimating a Trade Area. Journal of Marketing 28(3): 34-38.
Nakanishi, M. und L. G. Cooper, 1974: Parameter Estimation for a Multiplicative Competitive Interaction Model - Least Squares Approach. Journal of Marketing Research 11(3): 303-311.
Centre sampling (CS) is a technique used to realise probabilistic surveys even in the situation where the population list is completely missing or partially unknown. This method was originally devised to gather detailed information on the presence of immigrants in a particular area, and it is subject to further developments both on the applied and the theoretical level, with particular reference to the robustness of the results. This latter aspect is related to the comparison with other, potentially alternative, sampling techniques, which might rival CS performance, provided their actual applicability. In fact, a fundamental advantage of the CS is the ability of devising a theoretical scheme that is consistent with a set of constraints and limitations in terms of access to data, which in general tend to thwart the development of the survey. On the one hand, there is indeed still much work to be done with respect to the methodology. However the empirical evidence deriving from the numerous applications of CS seems to provide reassuring results, both in terms of usefulness and of effectiveness of the method. Since the 1990s, almost forty surveys on the foreign migrants' universe have been directly designed and developed using the CS scheme. The total number of statistical units included in those samples is more than 180,000.
Ref.: Baio G., Blangiardo G. C. e Blangiardo M. (2011), "Centre sampling thecnique in foreign migration surveys: a methodological note", in Journal of Official Statistics, vol 27, 3, pp. 1-16
With the introduction of the new statistical category "persons with migration background" in 2005, the understanding of immigrant population in Germany has changed. Migrant surveys increasingly aim at capturing much more heterogeneous subgroups: foreign citizens, naturalized persons and German citizens with at least one parent that is born abroad. This appears to be a difficult task due to a lack of official registration data for building a proper sampling frame.
The paper seeks to address this challenge and presents an innovative sampling frame - the so called Center Sampling technique - and its application for surveying persons of Bulgarian origin in Hamburg. The technique consists in interviewing persons at selected meeting points such as commercial, religious and educational centers. Main assumption is that each person visits at least one of the selected centers, so that a sample is drawn randomly at the selected centers. The bias of the sample is addressed ex post by weighting that is based on a calculation of the likelihood of each respondent to be included in the survey.
Both methodological challenges and chances of the sampling technique are discussed. While the method allows for including hard-to-reach subgroups like naturalized or unregistered persons, it poses a question about the probability of inclusion of persons who avoid visiting migrant-oriented centers. Finally, it is asked to which extent a survey based on this sample frame is an apt technique for drawing a sample in a representative way for migrant groups without an access to official population registers.
Sampling rare populations without sampling frames is challenging. Nearly all techniques like screening, snowball, RDS, quota and onomastic sampling have serious drawbacks, most often unknown selection probabilities or serious undercoverage. Many sampling techniques rely on spatial clustering of selection units, therefore the use of this principle for the selection of RDD blocks was examined in two nation-wide German RDD samples. The target population were Russian and Turkish migrants. Both groups are rare populations in Germany. A new RDD sampling procedure was implemented in a CATI call scheduling system with varying block sizes. A first survey (n=640) was done in January 2011, a second survey (also with n=640) with a different implementation started in January 2013. The presentation will explain the implementation in detail. By comparing both surveys, sampling efficiency, response-rates, summaries of para-data and the effects on dependent variables will be reported. The results show that the new procedure can be used in practice for RDD.