Many social research questions aim at understanding relationships
between different phenomena, and increasingly complex multi-variate statistical models are often employed to address these questions. The suitable choice of statistical model must be grounded in the nature of the research aim itself, with simple descriptive analyses forming the starting point for constructing robust statistical models. In this short article, we will demonstrate the importance of data description. Serving as an example study and cautionary tale, we will critically discuss a recently published research finding related to arson attacks on refugees in Germany. Ignoring descriptive figures, this study comes to wrong conclusions due to solely focussing on misspecified, complex statistical models. We therefore highlight how mismatches between descriptive figures and statistical models ask for caution and conclude by reminding the scientific community of the key role of descriptive data analysis at both the outset and conclusion of an analysis.
