口试评分规范化与信度研究

口语考试的效度较高,信度却比较低。但没有信度,效度也不可能真正得到保证。因此,如何提高口试的信度,是很多测试研究者普遍关注的问题。本文通过描述清华大学英语水平考试中口试部分的评分规范化与评分员培训,对如何规范评分以提高口试信度这一问题进行讨论。

Standardization of Marking and Reliability of Oral Tests

Evaluating a productive skill of learners, oral tests have a comparatively high validity. Their reliability, on the other hand, tends to be low. How to improve the reliability of oral tests has been a major concern to many researchers. This paper addresses the issue by describing the mark scheme and rater training of the speaking section of Tsinghua English Proficiency Test. The research focuses on the design of the mark scheme and the rater training aimed at the standardization of marking. It provides some insight into the relationship between the validity and the reliability of oral tests and that between the test reliability and the rater reliability.