A Comparability Study Of The Listening Sub-tests In CET-4 And CET-6

College English Test (CET), designed to assess the achievement in the Englishcourse of the non-English major undergraduates in China as well as to examine theirEnglish proficiency, is widely administered across the whole country. In accordancewith the National College English Teaching Syllabus (NCETS), CET consists of threetests: Band 4 (CET-4), Band 6 (CET-6), and the CET-Spoken English Test (CET-SET).CET-6 measures the English proficiency of test takers who have passed CET-4, and isthus higher in its designed difficulty level compared with CET-4. However, thequestion of how these two tests differ in difficulty level in reality on the listeningcomprehension (LC) sub-test has rarely been investigated. The present research aimsat addressing this question by analyzing the content of the LC sub-tests and a 9,461sample test takers' scores on the LC sub-test.Based on the TMF (Test Method Facets) framework proposed by Bachman,the author put forward a revised framework in the LC sub-test content analysis. Tentest papers, five for CET-4 and five for CET-6, administered from the year 2009 to2011 are analyzed from the perspective of test content construct, lexical difficulty,syntactic difficulty and characteristics of the response. The major findings of thecontent analysis are as follows: compared with CET-6, passages in CET-4 LC sub-testare shorter and the time interval between the tasks is longer indeed. Results also showthat the LC sub-test in CET-4 is easier than that in CET-6 in terms of the severalindexes we explored.9,461 test takers who sit in both CET-4 and CET-6 on 12/18/2010 are taken asour sample subjects. Results from the quantitative analysis show that when all the testtakers are taken as a whole, their mean scores on all the tasks in the LC sub-test andon the MCQ tasks of CET-4 are higher than those on CET-6, but lower on thedictation tasks. Furthermore, in order to explore how the difference on the difficultylevel of LC sub-test in CET-4 and CET-6 is manifested on different Englishproficiency test takers, the whole sample test takers are divided into three groups: thehigh, medium and low proficiency groups. The analysis indicates that medium and low proficiency test takers demonstrate the same result pattern on each part of the LCsub-test, however, high proficiency test takers' mean scores on each part of thelistening sub-test in CET-4 are higher than those in CET-6.This research is meaningful in the sense that it can provide a clearer view onthe test content and task characteristics of the LC sub-test in both CET-4 and CET-6,and on how the difference in the difficulty level of listening sub-test is manifested ontest takers' scores in reality, which will be helpful, to some limited extent, for the testdesigners to optimize the quality of tasks or the items so as to meet the requirementsstipulated in the testing syllabus on the difficulty difference between the two bands.