The problem of missing residential mobility information in the German microcensus

Más información sobre el libro

The German Microcensus (MC) is a large-scale rotating panel survey conducted over three years, notable for its high case numbers and mandatory participation, making it valuable for longitudinal analyses. However, a significant challenge arises from the lack of information on residential movers due to the area sampling approach, which raises concerns about potential biases in longitudinal analyses, particularly regarding transitions between labor market states. This study addresses these issues by employing a methodology that reflects the nature of the missing data in the MC, utilizing the German Socio-Economic Panel (SOEP), which includes data on residential mobility. The focus is on transitions between labor market and marital states, revealing potential substantial biases in these transitions. To correct for the bias, two strategies are implemented: the inverse probability weighting (IPW) approach and selection models. The IPW approach is integrated within generalized linear models, utilizing generalized estimating equations for weight estimation. Selection models are applied to incompletely observed contingency tables, employing loglinear and logistic modeling techniques, with particular attention to identification issues. The findings suggest that both strategies are effective in reducing bias in the analyses.