This project estimates the effects of a new educational intervention for school drop outs in the city of Rotterdam: the socalled Neighborhood schools. The effects are identified by exploiting the restriction on the number of youths that can participate in these schools. The schools start with 100 young persons. At the moment the assignment rules for the experimental and control group are being developed and implemented in cooperation with the city of Rotterdam and the neighborhood schools. In addition, the structure of the data collection is developed.