A longitudinal multidimensional analysis of EAP writing: Implications for pedagogy

Abstract

The proposed paper applies a multi-dimensional analytical framework (Biber, 1988; Biber, 1995; Biber & Conrad, 2009) to describe language variation on a longitudinal corpus of written EAP essays and reports collected from freshman undergraduates over their first semester of EAP training at a leading university in Hong Kong, known as the CAES-HKU learner corpus (Crosthwaite, 2015). While multidimensional analysis of L2 data is now increasingly prevalent (e.g. Weigle & Friginal, 2015; PerezParadez, 2015), there are very few studies using multidimensional analysis of longitudinal L2 data, which would allow researchers to quantifiably measure the impact of EAP instruction on L2 production along a range of dimensions of linguistic variation. Approximately 300 EAP essays and reports were taken at three data points during the semester (Week 1 pre-training, Week 9 post-training and Week 13 under exam conditions), totalling approximately 300,000 words. Multidimensional analysis was performed using the Multidimensional Analysis Tagger (Nini, 2014). The results show significant variation of linguistic features across the L2 cohort from the first data point to the third data point, suggesting a considerable effect of EAP instruction on L2 production, particularly on Biber’s dimension 3 (context independent vs. dependent discourse) where an increase in nominalisation is typical of scientific or learned exposition. However, there were few significant differences across data points for dimension 4 (overt expression of persuasion) suggesting that the author’s stance (a key provision of the EAP course in question) is as discernible pretraining as it is post-training, suggesting that improvements to existing EAP course materials are recommended.