| As a national teaching test,superintended by China's Ministry of Education,College English Test has been an important supplementary measure for level-based English teaching in College English Curriculum since the introduction of the reform and opening policy.It aims to measure university students' English actual capacity objectively and accurately and then serve college English teaching.Listening is an essential language skill,which is used far more than any other single language skill in normal daily life.A good test should measure what it is intended to measure and nothing else.if a test does this,it is said to be valid. Construct validity is the most important property of a test,and the general view of Construct validity is the extent to which the test measures the right constructs.Content validity is also considered very important.Up to now,the LC test of CET-6 accounts for a large proportion of the whole test.The present study explores the construct validity of the LC test of CET-6 from the following three aspects:1) test taking processes or strategies,2) test performances or scores,and 3) test item content,which are receivable and scientific.Participants for this research survey are students in Hebei Normal University and Shijiazhuang Vocational College of Posts and Telecommunications,who have just taken CET-6,June 21st,2008.The instruments are the CET-6 listening comprehension strategy questionnaire and the LC test paper of the CET-6.220 participants,with different academic backgrounds,do the questionnaire.The data obtained are analysized with the EFA in the SPSS 11.5 documents.Of the 220 participants,40 perform the LC test.6(selected from 35 test takers whose scores are valid) are trained to give the verbal reports.The major findings of this study are:The excessive use of the test-wisenesss strategies can lead to the failure of achieving the objective of the test.So,it has a negative effect on assuring the construct validity of the LC test of the CET-6.According to the content analysis, it can be said that,due to the larger coverage of the question type of Main idea and Details, the three question types—Main idea and Details,Inference,and Specific language features are not arranged in a proper proportion.In terms of the test performance analysis and the immediate retrospection data analysis,it is certain that when listeners perform questions of the three different types—Main idea and Details,Inference,and Specific language features, there is no significant difference of strategy use.The findings of this research are valuable for test designers to estimate construct validity from a new perspective,and for college teachers to instruct the students to employ language use strategies instead of test-wiseness strategies.
|
PDF Full Text Request