Changes In Questionnaire
Within the framework of the additional programmes the following variables in the 2013 Microcensus are new:
- Private and company pension schemes (EF392 - EF397), supplementary employment data: Type of shift work performed (EF155 - EF159), health data (duration of illness or injury, type of accident; type of treatment (EF467 - EF475), body height and weight (EF476 - EF477), officially recognised disability status and degree of disability (EF465 - EF466), risk of illness, smoking (EF478 - EF482), nationality of parents, if they have or have had a permanent residence in Germany since 1960, year of move and, if naturalised, former nationality: EF349 - EF382). In the corresponding categorizations (mig_*, EF2001 ff.) there are further categories based on the additional program.
- The comparison of the ad hoc module "Occupational accidents and work-related health problems" (EF1081 - EF1096) with the 2007 correspondent module (EF1081 - EF1096) did not reveal any significant conceptual changes, but comparability is limited due to partially changed questions and answer categories (Liersch 2014:562).
- In order to improve the measurement of ILO labour status, in particular, the fieldwork and a number of questions have undergone changes. Although no significant changes have taken place since 2011, some deviations between the results of the Microcensus and other labour market statistics are still to be found in 2013. This applies, for example, to small and marginal jobs. In particular, changes due to methodological factors must be taken into account when comparing with previous surveys of the microcensus (see Körner/Puch 2009; Körner/Marder-Puch 2015).
- Questions on private life insurance, including the total sum insured, were surveyed in the 2013 supplementary programme with two questions (see EF392 - EF393); in 2009 only one question was asked (see EF393). This also applies to questions on capital-forming benefits (2013: EF395, EF397;2009: EF395), supplementary company pension schemes (2013: EF394, EF396;2009: EF394) and private life insurance (2013: EF392, EF393; 2009: EF393).
- For cross-sectional evaluations for the survey year, records are not needed that are coded with -7 "Selection area without surveyed household" and -8 "No vacant flat, private households and collective accommodations which failed to respond in EF30 and other variables.
- For households from the Jahresüberhang [i.e., households who should have been interviewed in the previous year but who were not interviewed until the following year] who were interviewed using the questionnaire from the previous year, no data are available for the current four-year additional programme and for the questions asked for the first time in 2012. In the original data, this is covered only in part by the variable “Jahresüberhang from the previous year” (EF5b). In the Scientific Use File (SUF), the code used in such cases is the value -6 “
- When constructing the extrapolation factor EF956 used for the variables of the 2013 ad hoc module, the survey method (EF7), the labour status type (EF29) differentiated according to sex (EF46) and age group (EF44) at federation (Bund) level, the labour status (EF29) differentiated according to sex (EF46) at federal state (Land) level (EF1), and the two nationality groups German/foreign nationals in the regions East/West were deployed to estimate the response probabilities (1st stage). The target population comprises persons in employment and persons in former employment
- The extrapolation factor EF957 is available for analyses of disability data (EF465, EF466). It can be used to compensate for item non-response and systematic differences between the disability data and official statistics on people with severe disabilities. In the case of persons with an officially certified disability (EF465=1-3), the Microcensus data at federal level were adjusted to the results of the official statistics on people with severe disabilities.
- The regional details federal state (Bundesland) and a rough classification of the community size classes (Gemeindegrößenklassen) – the latter is not shown for the small federal states Bremen and Saarland – are included in the Scientific Use File (SUF). With the help of a special code of the variable community size class, it is possible to distinguish between West and East Berlin (EF563 = 9 Berlin-East).
- The other variables in the SUF are also coarsened if necessary, so that each value in the univariate distributions comprises at least 5,000 persons from the target population.
- The values of the variable nationality are aggregated in such a way that each group of nationalities in the target population comprises at least 50,000 inhabitants. In cases where variables are coarsened, the most strongly populated category is shown in the SUF
The SUF is a de-facto anonymised 70% sample. Until 2011, the sampling units were households or apartments where all persons in a selected household or apartment were included in the sub-sample. From 2012 onwards, the sampling districts within a rotational quarter will be used as sampling units for the subsample. This, together with longitudinal consistent identifiers, makes it possible to generate panel data sets with the Scientific Use Files.
- For details, see “Zur Abgrenzung der Bevölkerungs- und Erwerbskonzepte im Mikrozensus 2013" [The demarcation of the population and labour force concepts in the 2013 Microcensus]. [.pdf] (in German)
Year Specific Documentation
Statistisches Bundesamt / GESIS (2017): Wichtige Informationen zur Nutzung des Mikrozensus Scientific Use Files 2013. [.pdf].
This document contains the information on this website as well as further details on the microcensus SUF 2013.
Design Weight: Target
Dwelling, Household, Persons
Design Weight: Method
The Microcensus is a single-stage, stratified cluster sample with a uniform sampling fraction for all strata. As a rule, the sampling fraction is 1%; it applies to all sampling units (dwellings, households, persons).
Because the selection probability of the Scientific Use File (SUF) basically remains constant at 70%, design weights can be created on the basis of the inverse of the sampling fraction of the Microcensus (1%) and the Scientific Use File (70%): w = 1 / (0.01 * 0.7).
Final Weighting: Method
For the final weight a two-stage procedure is employed: In a first step the net sample of the successfully interviewed households is adjusted to the gross sample of all households to be interviewed (compensation). In a second step this gross sample is benchmarked to the update of the current population figures (adjustment). (The extrapolation is based on the results of the current population update, as updated by the Census in 2011.)
The final weights of the microcensus result from a combination of this two-stage procedure of compensation and adjustment.
In order to achieve an improved adjustment to the published results, the extrapolation factors EF951 and EF952 of the SUF have been adjusted to the gender, age groups (under 15 years, 15 to under 45 years, 45 to under 65 years, 65 years or older) and the employment status (employed, inactive) within each federal state. In SUF 2013, unlike 2012, this post-stratification was constructed, taking the household constellation into account. The extrapolation factors are already multiplied by the reciprocal value of the subsample's selection probability and refer to 1,000 persons of the population.