Aim of this project is an automatic evaluation of texts that were written in different language competence levels. Using methods of natural language processing, we plan to find out which of the criteria that have to be met to conform to a certain language level can be evaluated automatically.