Abstract

"Process-generated and administrative datasets have become increasingly important for labour market research over the past ten years. Major advantages of this data are large sample sizes, absence of retrospective gaps and unit non-responses. Nevertheless, the quality and validity of the information remain unclear. This paper contributes to this subject, focusing on the variation of research results due to alternative data cleansing procedures. In particular, the paper uses the general set up for data cleaning proposed by Wunsch/Lechner (2008) in evaluating the outcome of training programmes in Germany. First results are limited to the sensitivity of the construction of the sample populations used for the counterfactual analysis. The results emphasize that sample construction seems to be robust to the scenario used for the data cleansing." (Author's abstract, IAB-Doku) ((en))