Design and Mobility Effects in the Micro Census and Impacts on the Social Statistics

The Working Group

Coordinated by Johann Bacher, Johannes-Kepler-University Linz (JKU)

For more information on the JKU's contributions to PUMA see here.

Participants from the Statistics Austria:

Josef Kytir (Head of the Directorate "Population"), Daniela Gumprecht, Alexander Kowarik, Angelika Meraner, Cornelia Moser, Matthias Till

Description of the Working Group

In the Micro Census each selected household stays within the survey sample for five waves. This fact leads to several challenges for analysing personal characteritics such as employment status, school attendance etc.:

• There is a cluster effect: Each household constitutes a cluster. The persons within a household (especially when surveyed multiple times) are more similar to each other than persons from different households. Thereby, a decrease in accuracy occurs.
• The Micro Census - as an address sample - is not conceptualised as an individual panel. However, if household members are surveyed throughout five waves, certain mobility effects can be traced: A person within a household might not be met for a repeated survey, e.g. due to moving. If there is a correlation between the failure reason (e.g. moving) and the analysed feature (such as employment status), usually systematic biases appear, which can be called Mobility and Wave Effects. 

The Working Group researches both of these effects and their consequences for central characteristics of the Social Statistics (employment status, job-seeking behaviour, education, NEET, ESL etc.). In addition, it will be investigated if and to what extent the distorting effects of "outgoing mobility" might balance the "incoming mobility" amongst the Micro Census households.

The participants will also evaluate, whether the longitudinal integration  (micro-data linkage) of adminstrative data (esp. HV and AMS) might allow to using the Micro Census as a "real" individual panel.

Aims and Objectives

Objective 1: Contributing to the problem awareness for correct calculation of standard errors in the micro census

Objective 2: Providing methods for the correct calculation of standard errors in the micro census

Objective 3: Using the panel structure of the micro census

Objective 4: Analysing the mobility effects in the micro census 

Work plan

• Documentation of the proceeding of Bacher et al. in previous publications
• Development of an R-module for the calculation of standard errors in the micro census with Bootstrap
• Testing of the R-module by Working Group members and Feedback
• Workshop on the R-module

Objective 2: Estimation of the Impacts of Design and Mobility Effects

Associated Partner