Research On Oral Proficiency Scales

College English Test-Spoken English Test(CET-SET)is a large-scale computerbased oral assessment designed for non-English major college students.Taking CET-SET as an example,this study uses empirical ways to explore improvement on the English speaking rating scale.The study first made reviews on definition of oral proficiency,methods of speaking tests,categories of rating scales,and construction of rating scales as well as background and construction characteristics about widely influential spoken language proficiency scales.This study picked out six oral performance samples,divided into three pairs.The scores of samples had been pre-assigned by the experts from the CET-SET committee.Nine college English teachers were invited as raters in the structured interview to report on the oral performance of two test takers in one sample pair under the guidance of the researcher.When analyzing data,the researcher summarized the rating criteria adopted by raters and the discrepancies between those criteria mentioned above and the current CETSET scale on the basis of summarizing the oral reports from raters.Besides,the researcher also made qualitative studies on analyzing the ordering of importance of different criteria and comparison of holistic and analytic rating.The findings of this study put forward suggestions on the improvement on the current CET-SET rating scale as below: integrating depth and richness,logic and coherence,communicative strategy,and task fulfillment into the current analytic rating criteria;assigning different weightings to different criteria in the above-mentioned rating scale as raters varied tremendously in the ordering of importance of criteria;rater training being essential in making raters' perception consistent with rating scale;making improvement on CET-SET level description scale to provide more detailed descriptions on oral performance of each proficiency level.