| National Matriculation English Test(NMET)is one of China’s representative high stake and large-scale examinations,with more than 10 million candidates in the year 2021.Reading is an essential skill in second language learning,and reading comprehension also weights in language tests.Text complexity could reflect the linguistic difficulty of a text in a test.It also indicates the appropriacy of the text difficulty,which is a part of the validity work.Therefore,analysis on text complexity of NMET reading comprehension passages can verify the validity of NMET and provide strong validation evidence.Nevertheless,most studies on text complexity analysis put the foci on CET4 and CET6,IELTS and TOEFL,while few pay close attention to that of NMET reading comprehension texts qualitatively and quantitatively.This study aims to examine the text complexity of NMET reading comprehension passages and provides empirical evidence of text complexity analysis by concentrating on lexical,syntactic and discourse level text complexity.In this study,the online text analysis tool Coh-Metrix and the Eng-Editor are used to analyze the NMET reading comprehension texts and compare their text complexity.Based on the theoretical framework of factors influencing text complexity,this thesis elaborates on the factors influencing text complexity from three dimensions,lexical level text complexity,syntactic level text complexity and,discourse level text complexity.The present study will explore both synchronically and diachronically and answer the following questions: 1)Are there statistically significant differences in the text complexity among different sets of NMET reading comprehension passages in 2020 and2021 respectively?)Are there statistically significant differences in text complexity among the NMET reading comprehension passages from 2017 to 2021(sample data:NEEA Ⅰ set and Zhejiang set)? 3)What are the causation and implications of the differences and similarities in the synchronic and diachronic data? The author collected30 sets of 121 NMET reading comprehension texts from 2017 to 2021,extracted 33 indices according to the adapted framework,and conducted a statistical analysis of the result of the indices using R Studio.The detailed results and discussion were presented at the end.The significant findings of this study are summarized as follows.Synchronically: 1)The word count of Beijing set,Jiangsu set,and Tianjin set of 2020 are more than that in NEEA Ⅰ set,NEEA Ⅱ set and NEEA Ⅲ set respectively,according to the statistically significant difference result.Also,in 2020,the average length of words in Beijing set is longer than that in NEEA Ⅱ set,and the average length of words in NEEA Ⅰ set is higher than that in NEEA Ⅲ set.The 2021 Tianjin set has more words beyond the NMET syllabus than the NEEA Ⅲ set and NEEA Yi set.There are no statistically significant differences found in the other indices,indicating that lexical diversity and lexical density are reasonable in year 2020 and 2021.2)The proportion of argument overlap between adjacent sentences in the 2020 Beijing set is higher than that of the NEEA Ⅰ set and Tianjin set,and the proportion of argument overlap between adjacent sentences in the 2021 Beijing set is also higher than that of the NEEA Q-Ⅰ set and NEEA Yi set,which denotes that the cohesion of the adjacent sentences of Beijing set is better than that in NEEA Q-Ⅰ set,NEEA Ⅱ set,NEEA Ⅰ set and Tianjin set.3)In 2020,the Eng-Editor text difficulty level of NEEA Ⅰ set and NEEA Ⅱ set are lower than that of Jiangsu set,and in 2021,the Eng-Editor text difficulty level of Beijing set is lower than that of Tianjin set,and the Eng-Editor text difficulty level of NEEA Ⅱ set is lower than that of NEEA Ⅲ set together with Jiangsu set.Of 2020 and 2021,the distribution of NEMT reading comprehension texts is imbalanced in subject thematic context and genre,in which the distribution of human and society as well as expository are significantly higher than other categories.4)No statistically significant differences were found in text readability among the NMET reading passages of 2020 and 2021.5)Of the NMET reading comprehension passages of 2020 and 2021,the distribution of the thematic context and genre are uneven,and the proportion of human and society and expository texts are obviously higher than that of other categories.Diachronically: 1)There are no statistically significant differences in the lexical level text complexity,including indicators of lexical diversity and lexical density,among the NMET paper developed by NEEA and Zhejiang from 2017 to 2021.2)Regarding the indicators of syntactic level text complexity,the average sentence length of the 2019 NEEA developed Ⅰ set is shorter than that of 2019 NEEA set Ⅰ and 2021 NEEA Ⅰ set;the noun phrase density of 2017 NEEA set Ⅰ is lower than that of 2018 NEEA Ⅰ set,also,the noun phrase density of the 2018 Zhejiang set is lower than that of year 2020 Zhejiang set and 2021 Zhejiang set.At the same time,the modifiers before the noun phrases in the 2017 Zhejiang set are less than in 2018 Zhejiang set,indicating a lower density of textual information in 2018 Zhejiang set and an increasing use of noun phrases in the text.The LSA between adjacent sentences in the 2017 Zhejiang set is higher than in 2021 Zhejiang set,indicating better coherence of the NMET reading comprehension passages in 2017 Zhejiang set than in 2021 Zhejiang set.The 2018 NEEA Ⅰ set has a higher syntactic difficulty rating than in 2020 NEEA Ⅰ set,and the2019 Zhejiang set also has a higher syntactic difficulty rating than in 2020 Zhejiang set.3)No significant differences were found between the indicators of text abstractness and text readability in text difficulty,indicating that there is no significant change in the degree of text abstractness and readability between the two sets of NMET reading comprehension passages over the five years’ NEEA Ⅰ set and Zhejiang set.However,all four indicators of connectives in the NMET reading comprehension passages of Zhejiang set reveal significant differences.Specifically,the number of causal connectives in 2017 Zhejiang set are higher than that of 2019 Zhejiang set,and the number of logical connectives are higher in 2021 Zhejiang set than in 2018 Zhejiang set and 2020 Zhejiang set.Finally,the number of comparative and temporal connectives are both higher in 2019 Zhejiang set than in 2018 Zhejiang set.Overall,the coherence of NMET reading comprehension passages of Zhejiang set has decreased year by year since 2017.4)Again,the distribution of thematic context and genres in the NMET reading comprehension texts over the five years are similarly uneven in both NEEA Ⅰ set and Zhejiang set,with human and society and expository appearing more frequently than other categories.The causation of the differences found contributes to the no long development of NMET testing syllabus is no longer published and the admission requirements of different provinces vary since then(e.g.,choosing texts with more words or less words).Also,the preference of human and society and expository of some NMET reading comprehension passages inevitably results in the differences in syntactic and diachronic level text complexity.At the end,the adjustment of original source texts,such as adding or deleting could impact on the cohesion of the passages.The significance of this study is as listed.Firstly,it offers empirical evidence to verify the validity of the reading comprehension of the test,and delivers inferential evidence for the interpretation of the test scores.Then,it probes into the process of text selection and presentation of the NMET reading test,which could arouse the attention of the test administration and test developers and thus promote the enhancement of the quality and validity of this test.Thirdly,senior high school English teachers and students can refer to the results,understanding the characteristics of NMET reading comprehension texts and improve the teaching and learning thereafter.Finally,this study enriches the text analysis of NMET,with a view to providing methodology and paradigm for future research. |
