Taking the institutional context into consideration and ensuring the reliability of the instrument at each stage of the development of our text analysis instrument will hopefully provide enough information for the triangulation of data (as suggested by Hyland, 2000; Maykut & Morehouse, 1994). In recent research there are examples in discourse analysis (e.g., Todd et al., 2007) for triangulation by relying on several approaches or taxonomies to analyze the same set of data. The present analysis will use two complementary taxonomies, a cohesive and a non-cohesive one, to ensure the consistency of the data collection procedure.