| Instructional measurement and evaluation are quite significant parts for the process of teaching as well as learning. Moreover, they can guarantee teaching quality and learning effect to a great extent. Working as a primary tool of instructional measurement and evaluation, exam plays an important part in such a procedure, especially its accuracy and reliability. At present, most of the tests apply the form of paper-and-pencil test. But this method can’t provide examines with corresponding items according to their different ability lever, which have different difficulty levels. All the examinees use the same items, of which some items are too easy or difficult for a certain user. While computerized adaptive testing bears the advantage that it can provide corresponding items matching every certain examinee’s ability level. As a result, compared to the traditional test method, it shortens the length of test, at the same time, it could also accurately measure examinee’s ability value.Through defining CET-6 words difficulty value, this paper aims to develop an adaptive system for CET-6 words testing. Through the analysis in many aspects, a new way of defining English words difficulty value is proposed in this paper and a relative consummate CET-6 words system is also built here, which achieves the aim of using the least items to measure learner’s ability accurately, combining the traditional words memorizing method and online self-testing effectively. This paper combines several study methods, such as the literature analysis, mathematical reasoning method, survey research method, statistical analysis method and experimental method.This system adopts the maximum information method to select the new item, Bayesion Maximum A Posterior estimation to deduce the mathematics formula of estimating learner’s ability and fixed-length way to terminate the test process. Thanks to the aid of ASP, a CET-6 vocabulary adaptive testing system is developed and evaluated. Besides, through the participants registering account as well as password and then giving answers to the questions automatically provided by the system, their vocabulary ability will be estimated accurately.This paper’s focus or achievement is the construction of item pool, which guarantees the realization of adaptive testing. During the process mentioned above, the key part is the definition of word difficulty. This article takes three parameters into consideration. They are word frequency, word length as well as the harmony of speech and writing. In addition to defining the computing method of every parameter, the three parameters are assigned their respective weight 0.44, 0.30 and 0.26.On the basis of building vocabulary, through the demand analysis and system function design, the vocabulary adaptive testing system is developed. The thesis presents the key codes, including the choice of initial capacity value, item selection algorithm, the ability value precise estimation, the end of the test and ASP database linking part code. In addition, it presents the key interface of the system, including the log-in screen, interface initial interface and testing process.This study uses SPSS software to analyze the correlation data between the vocabulary ability and test scores. The result shows the significant probability value is 0.027, which is less than 0.05, to be able to refuse the null hypothesis and prove the correlation is significant. Besides, the correlation coefficient is 0.812 which bears statistical significance. Through experiment and research, it is proved that this system can measure examinee’s CET-6 word ability precisely. And it can also predict learner’s score of CET-6 test to some extent. |
PDF Full Text Request