A precise and cost-effective way of assigning key word to digital documents is essential for their effective use in digital repositories. Keywords help find documents are play a key role in content based business modells. The focus of this thesis is the comparison of different methods of key word assisgnment, i.e. manual key word assignment by humans (online-editors) and automatic key word assignment via mathematical procedures and weighting methods (tf-idf and entropie). These methods were compared in a test run. In this test run ten documents were assigned key words by humans and through mathematical formulas. The resulting key word sets per document were then assessed by the humans involved as to how competent they felt the author of the key word was. They did not know the origin of the key word set. An analysis of the results of this test made clear that automized key word assignment with the methods utilized could not be used as a substitute for manual key word assignment through humans. Nonetheless, scenarios are conceivable where automated key word assignment can be used to augment the work of human content workers. The thesis ends with a critical assessment and description of possible future uses of automatic key word assignment