[On the problem of missing data: How to identify and reduce the impact of missing data on findings of data analysis]

The impact of missing data on the analysis of empirical data is a frequently unrecognized problem. Missing data may not only result in a decrease in the actual sample size but potentially biasing effects on statistical findings have to be considered as well. Two important points are made in this article: Firstly, it is shown why the identification of potential causes of missing data should be an inherent part of any data analysis; secondly, the handling of missing data should be based on appropriate assumptions in order to avoid biased results and problems concerning the interpretation of empirical findings.