In this chapter, I bring together two studies addressing vocabulary in L2 English writing by Swedish students of English. The first study was based on written material by first-term students, while the second involved material from essays in the third- and fourth-term of English studies. In both studies a matched native group was used as a benchmark. I introduce two different ways of capturing frequent word combinations, one with a focus on collocations, applied in the first study, the other using the “lexical bundle” method (e.g., Biber & Barbieri, 2007; Cortes, 2004), applied in the second study. The results from both studies converge: the native groups exhibit a larger range of both collocations and lexical bundles compared to the non-native groups. The investigations involved quantitative analyses of the word combinations used, as well as a qualitative analysis of the functions they served. Finally, three current methodologies in the identification of multiword structures, including the two methodologies applied in the present study, are compared and evaluated in terms of qualitative aspects, such as size of material and control of task, topic, and discipline.
Cambridge: Cambridge Scholars Publishing , 2015. 177-205 p.