Which Progress for Poverty Studies can we expect from new large Data Sources?

Large data sources would allow us to test the impact of neighborhood characteristics, such as poverty rates , on the attitudes and behavior ot residents. The article explores the feasibility of existing large date sets for such a purpose. Unfortunately, none of the three sets reviewed, the Microcensus, the ALLBUS and the SOUP, allows for such multi-level analyses, because data cannot be regionalized due to data protection or insufficient sample size. To overcome these problems in a limited sense, it is suggested to pursue a “puzzle strategy” to combine data from different existing data sets.