Font Size: a A A

Validation Study On 2008-2009 National Matriculation English Test Papers Of Three Provinces

Posted on:2011-09-07Degree:MasterType:Thesis
Country:ChinaCandidate:J YanFull Text:PDF
GTID:2167360308468778Subject:Foreign Linguistics and Applied Linguistics
Abstract/Summary:PDF Full Text Request
National Matriculation Test (NMT) in our country is a large-scale high-stake examination. National Matriculation English Test is an important part of NMT. It is the largest norm-referenced standardized English proficiency test in China. Since 2004, apart from Beijing, Shanghai, all provinces have stepped toward implementing high middle school new curriculum reform and autonomous designs of NMT papers among provinces have begun. Up to 2007, altogether 18 provinces or cities have designed their own NMT papers with the number of English test papers up to 19 sets, including 2 national and 17 provincial English test papers (Hainan and Ningxia have shared one set of NMET paper). The quality of NMET papers and the consistency of test in each province or city over the years have caught great attention among the public and have become the key research topics in all fields, especially the education circle. However, the validity studies on NMET are merely some qualitative analysis of content validity, item analysis, etc. and occasionally, there is some quantitative information such as item difficulty and discrimination. In light of the particularity of NMT, the test takers' performance data are kept confidentially in examination institutes of each province or city. Rarely has the NMT data analysis been published to the public, which accounts for a huge limitation of NMT validation studies.This study is based on the NMET data collected from 2008 China National Philosophy Social Science Planning Project—Developing An English Score Equivalent Model for All Provincial College Entrance Examination. Three provinces have been selected, representing regions of high, medium and low level of senior middle school English teaching respectively.2008 and 2009 NMET papers of these three provinces have been chosen, specific data information of test takers'responses on the six papers are collected, and vertical and horizontal comparison and analysis of the selected papers are conducted by using both qualitative and quantitative analysis. The following conclusions are drawn:1. In terms of reliability, Cronbach internal consistency coefficient has been used. The results show good internal consistency of the test papers of three provinces over these two years. The reliability of these six papers is very high; 2. In terms of construct validity, correlation analysis and factor analysis have been conducted. It is found that each component of the test has high correlation with each other and only one factor is generated to explain the majority variance. Based on the Test Syllabus and Senior Middle School English Curriculum Standards, NMET should focus on the integrated ability of using English. Hence these six test papers have relatively good construct validity; 3. In term of content validity, the test items are compared with the Test Syllabus and Senior Middle School English Curriculum Standards. Evidence has been found to support the generally good content validity of the test papers for their wide coverage of test points; 4. In terms of the item quality, the item difficulty and discrimination analysis have been conducted. Except the province of high English teaching level, the difficulty indices of the other two provinces over two years are not stable. One year's items are relatively difficult and the other year pretty easy. The extremely difficult or easy items are also the main reasons for low item discrimination. With the detailed analysis of the problematic multiple-choice items, factors leading to these problems are explored.It is hoped that this paper would contribute to make up for the limitations of the NMET researches and shed some lights on further NMET paper designing. However, considering that the limited data only covers three provinces or cities over these two years, the results may not be comprehensive. In certain sense, further work need to be done in order to ensure the scientific and consistent NMET papers, hence to improve the quality of NMET item writing and to promote the development of basic English teaching in China.
Keywords/Search Tags:NMET Papers, Reliability, Content Validity, Construct Validity, Item Quality
PDF Full Text Request
Related items