We provide a comparative perspective on the contributions of the special issue with regard to their applied methods and findings. In addition, we discuss problems that arise when using ‘wrong’ or at least ‘incorrect’ versions of election manifestos by presenting replications of estimated policy positions of German parties. We show that the latter can result in biased estimates that may affect the outcome of theoretical models. On the basis of those findings, we present the idea of the open access archive polidoc.net to build up a common database for political texts.