MPIDR Technical Report

Fertility assessment with the Own-Children Method: a validation with data from the German Mikrozensus

Abstract

The German Mikrozensus is a data set that is frequently used to study fertility behavior. With its large sample size, the dataset also offers the opportunity to investigate childbearing among population subgroups, such as particular migrant groups. Because birth histories are not surveyed in the Mikrozensus, scholars have frequently employed the Own-Children Method (OCM) to generate fertility schedules from this data. The OCM draws on information on the number and ages of the co-residential children in order to estimate the fertility schedule of women. It is clear that this method leads to an underestimation of fertility, because children do not necessarily live with their biological mother. It is also clear that the bias is stronger for older women, because children may have left parental home. In the past, it was not possible to assess the bias. With the Mikrozensus 2008, we are able to validate the OCM by relying on respondent’s own report on the total number of children. We furthermore examine whether the OCM leads to worse estimates for migrants and non-migrants. The results show that the OCM provides reasonable fertility estimates for women in western Germany up to age 40, irrespective of migrant background. For eastern Germany, there is a stronger bias at later ages than in western Germany. This relates to the early age at childbirth in this region and the more diverse family structures.