A Corpus-based Study On The Correlation Between Lexical Features And Scores In English Majors’ Writings

The lexical research based on the corpus has always been the focus of the scholars. As one of the most important parts in English, applications of corpus in vocabulary usage in terms of writings also got an extensive attention from the researchers. This paper is designed to explore the lexical features of English majors’ compositions comprehensively, aiming to provide a reference for students’ improvement of writing ability and lexical teaching as well.This study selects 290 timed English majors’ compositions shared with the same topic from Spoken and Written English Corpus of Chinese Learners (Wen,2008), and ranks them according to the scores. The first 100 pieces of composition are high score group, and the last 100 pieces of composition are low score group. Then the author conducts a comprehensive lexical features analysis and comparison from lexical richness, lexical cohesive devices and formulaic sequences used in the two groups with the help of RANGE, AntConc and SPSS 19.0. Finally, the study samples the compositions in the two groups respectively to examine the correlation between lexical features and the scores. The research mainly answers the following questions:1. Is there any difference between the two groups in lexical richness, lexical sophistication and lexical density? What are the features of word frequency of each group?2. Is there any difference in the quantity and category of lexical cohesive devices between the two groups?3. What are the characteristics of formulaic use in high score group and low score group? Are they different?4. How about the correlation of lexical features and writing scores in two groups? Is there any difference between them?The results indicate that:(1) Compositions in high score group enjoy higher lexical variation, lexical sophistication and lexical density than those in low score group; the accuracy and appropriateness of personal pronouns and modal auxiliary verbs will still be problematic in both groups. (2) The two groups show the same trend in the use of lexical cohesive devices, but the number of lexical cohesive devices in high score group is nearly as twice as low score group. (3) The two groups show the similar trend in formulaic use, that is:the amount of formulaic sequence used in compositions decrease with increasing formulaic length, and formulaic use with 3 words is richer than 2 words clusters and 4 words clusters. (4) The two groups are different in the correlation of lexical features and the scores:in low score group, the quantity of TOKENS, TYPES, repetition and synonym is significantly related to the scores, while in high score group, the amount of TYPES, synonym and superordinate has statistical significance with the scores. Besides, in both groups, the amount of formulaic sequences is significantly related to the scores, as well as the 4 words formulaic use, but the correlation between 2 words clusters and the scores is not significant, as well as the correlation between 3 words clusters and the scores.According to the first research finding presented above, this research suggests that English majors should deepen and widen their writing vocabulary, enhance lexical richness and be aware of the use of personal nouns and modal verbs during writing. Besides, the second, third and fourth findings indicate that English majors should put their attention on the influence of lexical cohesive devices and formulaic sequences on the scores and accumulate lexical cohesive devices and formulaic sequences during study, especially synonym, superordinate and 4 words clusters. Moreover, teachers should cultivate students’habits of using lexical cohesive devices and formulaic sequences reasonably, and teach them during classes.