计算机口语考试信度研究  

A Study of the Reliability of Computerized Oral Proficiency Interview

作　　者：戴朝晖[1] 

机构地区：[1]上海大学外国语学院

出　　处：《外语电化教学》2011年第2期45-50,共6页Technology Enhanced Foreign Language Education

摘　　要：本研究分析了信度概念及影响因素，并从成绩相关性、效度、评分维度及施测维度等诸信度要素分析了传统面试型口语考试（OPI）与计算机口语考试（COPI）。研究结果表明，OPI与COPI相关性较高，具有较高的效度和可替代性。OPI与COPI信度较高，测试结果不存在显著差异。评分维度上，两位评分者在OPI及COPI中的严厉程度没有显著性差异，评分信度较好。但即便是高相关的两种测试，COPI仍在项目及考生维度上较OPI产生了更多、更严重的评分偏差。施测维度上，考生心理上更倾向于OPI，COPI使考生紧张，且对施测环境COPI有较大干扰。尽管COPI施测手段更公正，但较OPI缺乏交互性。要提高COPI施测信度，除考前需做好充分准备外，更需关注评分信度，固化评分标准，定期培训评分员，实时评估COPI评分信度，同时充分考虑COPI施测局限性，合理设置题型。This paper first quotes the concepts of reliability and its affecting factors. It also reports a contrastive study of the reliability between two kinds of oral English tests, namely the traditional Oral Proficiency Interview （OPI） and the Computerized Oral Proficiency Interview （COPI） from the reliability affecting perspectives such as results correlation, validity, rating and other dimensions. Research shows that high co-efficiency exists between COPI and OPI results, hence COPI can replace OPI and both of them have high validity and reliability. There is no significant disprepancies in the severity of the two raters, and rating reliability is quite satisfactory. However, COPI biases more in the dimension of testing implementation than in OPI. The subjects view the COPI test as more objective and fair, but still favor the traditional OPI test, believing that the COPI test is more demanding and challenging, less interactive and personalized, and makes them more nervous. Finally, this paper also states that pre-examination preparations are indispensible to improve the reliability of COPI. Lastly, as COPI nowadays is particularly suitable for large-scale testing and lacks real oral communication interaction, testing items limitation for COPI should also be taken into consideration so as to make the test much more reliable.

关 键 词：信度 计算机口语考试 

分 类 号：H319[语言文字—英语]