Performing causal analyses in the field of Economics of Education has been constrained by the lack of large panel datasets. With some exceptions, mostly in the UK and the US, most available datasets are cross-sectional survey data that make it difficult to assess the educational performance over time.Administrative data on classes and schools are better suited for some causal analysis than survey data because they make it possible to observe all pupils of all schools in a given geographic area.In this project, newly available German administrative school data are to be prepared for research purposes. These data contain information on all pupils of a set of regions. It is to be assessed which analyses the new data will allow to realize.A further goal is to use the new datasets to start performing a first analysis.