Font Size: 

Errors inevitably affect data. Statistical Institutes deal with them with rules designing data relationships and imputation for missing information. In this paper, we report the use of classification trees to improve this process when using longitudinal administrative data in the estimation process.