Visit GESIS homepage
MISSY: Homepage aufrufen
Metadata for Official Statistics
FDZ GML: Homepage aufrufen

Microcensus Scientific Use Files from 2012 onward as a Rotational Panel

From the Scientific Use File of the Microcensus 2012 onwards, it is possible to independently merge multiple survey years into panel datasets using longitudinally consistent identification numbers (IDs). The newly created longitudinal datasets provide advanced methodological and substantive analysis possibilities. Since sampling districts remain in the sample for a maximum of four years and a quarter of the sample is replaced annually, the observation period can extend up to four years.

Due to the complete renewal of the sample in 2016, cross-sectional data from the survey period 2012 can only be linked up to and including the 2015 survey. Additionally, due to the new regulation of the Microcensus from 2020 onwards, linkages from the survey period 2016 are only possible up to and including the 2019 survey. The cross-sectional data of the Scientific Use File 2020 can be merged with the cross-sectional data of subsequent years in the future.

The paper (available in German) describes the combination of the Microcensus Scientific-Use-Files 2016-2019 into a person panel data set. The procedure is described in detail, the number of sampling districts, households and persons and the reconstruction of the rotation groups are presented. The presentation of the procedure concludes with consistency checks. Afterwards, a small sample analysis is shown to illustrate the analysis potential.

  • Backes, N., & Pforr, K. (2023). Verknüpfung von Mikrozensus-Querschnitterhebungen 2016-2019 zu Paneldaten. (GESIS Papers, 2023/11). Köln: GESIS - Leibniz-Institut für Sozialwissenschaften. [.pdf]
Stata-Code
Combination 2016, 2017, 2018, 2019 mz_panel_16171819.do

The paper (available in German) describes how researchers can independently create panel files and what has to be taken into account. In addition, there are programs for SAS, SPSS and Stata for linking the cross-sectional data.

  • Herter-Eschweiler, R.; Schimpl-Neimanns, B., 2018: Möglichkeiten der Verknüpfung von Mikrozensus-Querschnitterhebungen ab 2012 zu Panels [.pdf]
SAS-, SPSS- and Stata-Program

Please note:

The syntax files should be downloaded and used with statistics programs. The browser view provides only an overview. Characters with an Umlaut and line breaks are sometimes displayed incorrectly.

Version 1 (longitudinal orientation):
For analysing gross changes and survival analysis
  • 2012, 2013, 2014 and 2015
    		•

    panel_12131415_v1.sas
    panel_12131415_v1.sps
    panel_12131415_v1.do
  • 2012, 2013 and 2014
    		•  panel_121314_v1.sas
    panel_121314_v1.sps
    panel_121314_v1.do
  • 2012 and 2013
    		•  panel_1213_v1.sas
    panel_1213_v1.sps
    panel_1213_v1.do
    Version 2 (cross-sectional orientation):
    For the analysis of net changes or cumulated data from consecutive years
    panel_1213_v2.sas
    panel_1213_v2.sps
    panel_1213_v2.do
    Nonresponse adjustment and design weights in the German Microcensus panel 2012 – 2015

    Since the Scientific Use File of 2012, the annual data can be merged into a rotation panel. In accordance with the principle of area sampling, residential movers are not followed. The core methodological problem is whether the resulting dropout leads to biased results. If this dropout is to be corrected by means of design weights, some information is currently missing from the data; among other things, the differentiation of the reasons for dropout – spatial mobility and death. In this report, the missing information on deaths is imputed using a model-based procedure. It is then shown by way of examples how longitudinal weights can be created. Furthermore, effects of missing information on deaths are examined. If the longitudinal populations contain deaths incorrectly, the longitudinal weights are higher than with the correct procedure and lead to compensation of deceased persons. Against this background, better data provision is proposed.