The Effect Of Error Scores Of English Multiple-choice Cloze On The Validity Of Testing In Senior High Schools

The term"cloze test"was invented by W.L. Taylor in an article named"cloze procedure"in 1953. It was originally constructed by deleting words from a selected material and the examinees were required to fill in the blanks. As a test type which is able to test English learners'English efficiency effectively, cloze test draws the great attention of English teaching and testing experts and educators in China. As a main sub-type of cloze test, Multiple-choice cloze, widely used in both the college entrance examination and ordinary high-school English tests, is becoming more and more significant and attracting more and more English teaching and testing experts and educators to do various researches about it. Although those experts and educators have got the conclusion that Multiple-choice cloze is easy to deal with, formally-organized, simple to evaluate, facilitates students'guess toward answers, lowers the difficulty of tests, and tends to produce error scores, no experts and educators can analyze the proportion of error scores and the effect it has on validity of examinations comprehensively. The following research questions are examined in this thesis consequently, including the proportion of error scores in high-school multiple-choice cloze tests, the relationship between error scores and English learners'English efficiency, and in what ways and to what degree do error scores influence the validity of English examinations.The researching process and method are listed as followed. For the propose of getting the results of the proportion of error scores in high-school multiple-choice cloze tests, we designed two tests with the same language points but different test types and compare examinees'scores in these tests afterwards. The first test consists of three multiple-choice cloze passages. The second test contains the same language points but different test types, including translation form Chinese to English, translation form English to Chinese, correct mistakes, and fill-in blanks. In order to guarantee the validity of scores and data and the reliability of those researches, 53 sophomores from an experimental class in a key senior high school were selected as examinees. After getting the final 100 valid test paper, we use some data analytic soft wares to analyze the data and gain the final results which are listed below.We have got several inspiring conclusions. First, error scores do exist in multiple-choice cloze tests. The result suggests multiple-choice cloze does lower the difficulty of tests and fail to measure learners'true English efficiency. Second, the proportion of error scores is in inverse proportion to learners'English efficiency. Third, we analyze how error scores influence the validity of examinations negatively in terms of content, concurrent and backwash validity.At last, we further points out several recommendations for English teaching and testing.