The German Microcensus (MC) is a large scale rotating panel survey over three years. Because of its high case numbers and its mandatory participation the MC is a valuable survey for longitudinal analyses within a 3 years frame. A serious obstacle is the missing information about residential movers which is due to the area sampling approach used for the MC. This raises the questions whether longitudinal analyses, like transitions between labour market states, are biased and how different methods perform that promise to reduce such a bias.
These questions are examined by a methodology that carefully reflects the nature of the missing data in the MC. This is achieved by the use of similar survey, the German Socio-Economic Panel (SOEP), which covers residential mobility. The emphasis is on the transitions between labour market states and marital states. The results indicate that there may occur a substantial bias in the analysed transitions.
For the correction of the bias two different strategies are used: inverse probability weighting (IPW) approach and selection models. The IPW approach is proposed in the context of generalized linear models. For the estimation of the weights the generalized estimating equations technique is used. The selection models are proposed for the incomplete observed contingency tables. In formulating selection models the loglinear and logistic modelling techniques are used. Special emphasis is given to the identification issues of the selection models. The finding is that both strategies can be quite effective in the reduction of the bias.
SchlagworteBetriebswirtschaftslehre GEE Generalized estimating equations Inverse probability weighting IPW Labour market analysis Missing data Panel survey Residential mobility Selection models Statistik
Ihr Werk im Verlag Dr. Kovač
Möchten Sie Ihre wissenschaftliche Arbeit publizieren? Erfahren Sie mehr über unsere günstigen Konditionen und unseren Service für Autorinnen und Autoren.