Constructing An English Exam Item Bank Of Final Achievement Test For Junior Middle School Students

Item bank is making a substantial impact on educational test development. It prompts large-scale, standard and scientific tests. The present research, which based on Classical Test Theory, constructed an English exam item bank of final achievement test for junior middle school students in Chenggong County. Specially, it aimed to identify the characteristics of the distributions for all the 2, 400 items in the 40 pretests and to determine the facility value and discrimination indices of those items with the item bank in this research. The dijBferences of facility value of items among different samples and different methods were also scrutinized. To achieve these goals, the research sampled 308 testees from three schools in different areas. A set of procedures and instruments were contrived, following the preparation of test specification, scoring criterion, test materials, a panel, the revision of test material and scorer training, the pretest and scoring, item analysis sheets, and the collection and analysis of data of the 40 pretests. The 40 pretests were administered in terms of regular exams and included the same test battery: 1.) Vocabulary, grammar and functional language, 2.) Reading comprehension, 3.) Writing expression.The results show that the reliability indices of the 40 pretests are all acceptable. The mean facility value index is 0.50 and therefore is ideal. 99.63 percent of the 2, 400 items have a range of facility value between 0.10 and 0.90 and 96.79 percent of the 2, 400 items have satisfactory discrimination indices of more than 0.20. Theoretically and practically, these research items with a widerange of facility value from 0.10 to 0.90 and of discrimination indices more than 0.20 are all satisfactory with the standards of an item bank. Thus a total of 2, 323 items that meet the requirements were stored with their parameters in the item bank for future use.On the basis of the results of the present research, it is found that there were obvious differences of facility value of items among different sample groups from the three schools selected owing to certain testees' characteristics, such as: geographical location, educational background, proficiency, language learning history and the like. While the results indicate that there exists no difference of facility value of items among test methods.The research led to construct an item bank on the basis of Classical Item Analysis for the first time in Chenggong County and broke a new ground in the study. It also explored a convenient way to construct an item bank in a grass-root English teaching section.