基于多面Rasch模型的商务英语口语测试评分研究

A Study of the Scoring of Business English Oral Test Based on Many-facet Rasch Model

信度和效度研究较多考察通用英语测试,而对专门用途英语测试的效度微观研究则相对缺失。本文应用FACETS软件对某高校VECTOR商务英语主观题型测试进行分析,从微观上对评分者严厉程度、考生能力差异、试题难易度差异进行进一步分析。结果显示,与其他通用英语主观题型测试的经验性结论相比,商务英语话题任务难度差异大,评卷者严厉度存在显著差异。本研究的结论有助于解释商务英语口语测试效度,并且利用该模型对评分者进行培训,从而改善考试设计、控制评分质量,改进评分标准和提高考试效度,这对相关教学和测试意义重大。

Reliability and validity research mostly examined general English test, while the validity of the ESP test was relatively absent. In this research, the FACETS software is used to analyze the oral test of VECTOR Business English in a university. The severity of the score rater, the difference of examinee＇ s ability and test item difficulty are further analyzed from a micro perspective. The results show that, compared with the empirical conclusions of other general English speaking test studies, there are great differences in the difficulty of the topic of business English oral test items and there are significant differences in the severity of raters. The results of this study are helpful to explain the validity of business English oral test and train raters, so as to improve the test design, control the quality of the scoring, expand scoring standard and test validity, which is of great significance to the business English teaching and testing practice.