In terms of lexical complexity, the most frequently used measures are lexical density and type-token ratio。 However, many researchers (Malvern & Richards 2002; Vermeer 2000; Yu 2009) have found that type-token ratio is easily affected by the length of texts, and, therefore, is of low reliability。 In order to rule out the effect of text length on type-token ratio, some scholars developed new type-token ratios, for example, mean segmental type-token ratio, D measure, and MSR-D。 The present research employed MSR-D developed by Meara and Miralpeix to measure the lexical variation of the students’ compositions。 

Third, the compositions analyzed for the present study were written by students in writing contests。 As the compositions of the three grades and the high-scoring group were all produced in the examination context, variables concerned with different contexts were excluded。 Moreover, timed argumentative compositions in the contests also indicated that the students took the writing task seriously and the written products could reflect the best of their writing。 

Finally, this study reported the calculation and interpretation of effect sizes for the parametric and nonparametric statistical analyses employed。 Effect sizes measure the practical significance or magnitude of a treatment effect。 Unlike the p-value of statistical significance tests, effect sizes are independent, or relatively independent of sample size。 Considering the limitations associated with statistical significance testing and p-values as the sole criterion for interpreting the meaning of results (Field, 2005; Pu, 2014, 2016), the sixth edition of the APA Publication Manual (APA, 2010) requires effect sizes be reported for quantitative research in addition to statistical significance values (Ellis, 2010; Pu, 2014, 2016)。 However, effect sizes were rarely reported by domestic L2 researchers。 文献综述

In summary, the current study investigated the development features of both syntactic complexity and lexical complexity of compositions written by Chinese English majors in the context of writing contests。 In addition, it employed relatively new and more valid measures so as to reflect the development of syntactic complexity and lexical complexity more accurately。

1。3 Structure of the Thesis

The thesis consists of five sections。 The present section serves as an introduction, presenting the research background, significance of the study and the overall structure of the paper。 It is followed by a review of related theories and empirical studies on syntactic and lexical complexity。 Section Three is the methodology part, introducing the specific research questions, and describing the participants involved, the instrument adopted, and the ways the data were collected and analyzed。 The next section reports the findings obtained from data analyses and discusses the possible reasons that might explain these results。 The last section is a summary of the major findings, followed by a discussion of the limitations of the study。