Multidimensional analysis of Master thesis abstracts: a diachronic perspective

This study aims to investigate the diachronic change of Master thesis abstracts written by Chinese students in the applied linguistics. A corpus of 1000 English abstracts was built with 100 abstracts per each year during a 10-year period from 2009 to 2018. Based on the multidimensional analysis, both the textual and linguistic changes were investigated. To be specific, the Biber's (Variation across speech and writing, Cambridge University Press, Cambridge, 1988) six-dimension model was adopted to capture the dimensional and linguistic styles of abstracts in each year. Multidimensional analysis tagger (Nini in Multidimensional analysis tagger (version 1.3), 2015. ) was used to automatically extract the data including the z-scores of dimensions and linguistic features. Further analysis focusing on the linguistic features of each dimension was performed by using stepwise regression analysis. The results showed that there was a pattern of a 3-year cycle of abstract style and the textual feature of Dimension 1 and linguistic features of Dimension 1, 3 and 5 had significant differences during these years. Two reasons, internal and external, were suggested to interpret the diachronic evolution of English abstracts by Chinese students.