The Development Of The Listening Ability Subscale Of China’s Standards Of English

There is a large English learning population and numerous English tests and exams coexisting in China.It’s an urgent call for the development of a national foreign language assessment system,in order to face the challenges brought by globalization and meet the needs of China’s education reform.Developing China’s Standards of English(CSE)is of top priority and is the most fundamental work in the project of the National Foreign Language Assessment System.This thesis introduces the process of developing the Listening Proficiency Subscale,which is an indivisible part of the CSE,including the theoretical and empirical foundation,the development of a descriptive scheme and a descriptors pool as well as the validation and scaling methodologies.An empirical example of descriptor calibration of the Level 5 and 6 descriptors is given,where many-facet Rasch model,one of the IRT models,is employed to analyze the currently available data from a nationwide survey.To begin with,a use-oriented model of listening ability was developed.Then listening ability descriptors were collected through documentation and sampling to build a descriptors pool,and went through several rounds of revision in terms of wording and classification.Finally three scaling methodologies,i.e.intuitive,qualitative and quantitative methods are combined to transform the statements into measuring instrument.Results from Rasch analysis of 239 descriptors from Level 5 and 6 indicate a high descriptor quality with good fit statistics and reasonable difficulty and discrimination estimates.However,some problems have also surfaced:for example,the difficulty estimates of some of the descriptors mismatch its predetermined level,which call for careful reconsideration and revision after all the data are collected and analyzed.