Two main sets of 40 texts were compared: 20 Research Articles (RAs) and 20 Master’s theses (MA theses). The MA thesis corpus is further subdivided into two sets of 10 higher rated and 10 lower rated texts, based on the final mark they received. The corpus was accumulated gradually through the previous stages (2–4 described in Sections 6 to 8), yet, as the analysis at each stage focused on different aspects, a full analysis of the corpus has not been carried out yet. The selection criteria for research papers to be included in the corpus have been described at the earlier stages; briefly, both RAs and MA theses were empirical papers from the field of applied linguistics written after the year 2000. The whole RA corpus is 143 083 words long, the full MA corpus is 315 578 words. For this part of the empirical study, the whole corpus was re-analyzed from beginning to end using a uniform method (described in Section 9.3).