A Study On The Oral Rating Scales For English Majors

Communicative approaches to language testing have led to the widespread use of oral language assessments to measure learners' performance on real-life spoken tasks. Among elements involved in derived scores, rating scales are considered essential.TEM (Test for English Majors)-Grade 4 Oral test examines students'spoken abilities from various components of language abilities by designing the corresponding rating scales. However, the rating scales for TEM-Grade 4 oral test are rather broad. There are only four levels. As a result, the ratings in the classroom speaking tests are not totally based on the existing rating scales and some rating standards are made by the teachers. The reliability of such ratings is doubtful. Classroom speaking tests lack reasonable rating mechanism and dependable rating scales.This thesis employed the EBB rating theory to create more detailed rating scales based on quantitative results of testees'oral corpus and qualitative results of raters'verbal reports. The constructed rating scales were more suitable for classroom speaking tests. The procedures were given as follows:selecting speech rate, pause, grammatical errors, vocabulary range and vocabulary richness as key linguistic features after analyzing the three aspects of linguistic performance:fluency, accuracy, and complexity; sixty spoken responses from SECCL and their corresponding transcribed texts were employed to measure the values of those linguistic variables; the correlations between the linguistic variables and spoken performance were investigated using Pearson correlation analysis, and it was found that the scores had a positive correlation with speech rate, vocabulary richness and vocabulary range, and had a negative correlation with grammatical errors; regression analysis was performed using the significantly correlated variables to find variables which significantly predicted examinees'spoken performance, and the results showed that grammatical errors had the strongest influence on the examinees'performance, followed by vocabulary range, vocabulary richness and speech rate; the verbal reports produced by raters when rating the spoken responses were transcribed and their themes were extracted; the qualitative results were combined with quantitative results to identify salient features that could distinguish between oral proficiency levels. These salient features were used to develop new rating scales according to the EBB rating theory.