A Study on the Validity, Reliability, Difficulty and Discrimination of the Final Test of College English A Case Study in Hezhou University 英语专业毕业论文.doc

A Study on the Validity, Reliability, Difficulty and Discrimination of the Final Test of College English: A Case Study in Hezhou University Abstract College English is a basic course for non-English major students, which includes listening, speaking, reading, writing, translation and so on. Therefore, studying College English test is extremely important. This paper is an empirical study, laying emphasis on studying the College English final test on validity, reliability, difficulty and discrimination. The samples of this study are selected the third term and fourth term’s final test papers of a class of non-English major students from Hezhou University. And then getting some basic information from their test papers, such as item validity, item reliability, difficulty indices and discrimination indices etc. by using SPSS17.0 and Excel. What’s more, the paper uses documentation method, statistical method and comparative method to collect and analyze data, which meets the need of the study. This paper mainly studies the following aspects in detail: content validity, predictive validity, reliability, each item’s difficulty and discrimination. From the results of this study, it is obvious that firstly, the mean of total scores is 76.191, 47% students get scores ranging from 80 to 90, and only two students can’t pass exam. That indicates most of the students can not only pass exam but also do well in the test. Secondly, to some extent, the validity of test is reasonable and acceptable. But it also has to improve: adding up the part of speaking in the test and make the test difficulty equal to CET4 or CET6. Thirdly, the reliability of test is not enough, needing to reduce the influence of subjective factors and objective factors. It should add up more items, adjust the difficulty of the test, and put forward some fair and efficient methods to mark. Fourthly, on the whole, the difficulty of the test is not very large. In the test paper, only 2.4% items are difficult, and