This exploratory study focuses on the emerging errors in an analysis of 10 higher and 10 lower rated MA theses (from an expansion of the previous 10-paper corpus). The corpus consists of approximately 300,000 words. The two subcorpora compared in the present research are differentiated on the basis of marks and language scores:

 

The research questions that guided Stage 4 are the following:

 

At this point, the focus will not be on the different reference item types or patterns, but on the identification of the typical error types and the ways in which they may be categorized. The density of errors in the different text parts will be used as a measure to identify the most problematic text parts for novice writers.