论文部分内容阅读
Computerized tests have become popular due to the highly developed technology and the great demand of the large quantities of the test takers of various kinds. Paper-and-pencil tests are not cost-effective and it takes a long time for test takers to get feedback. The test scores of the computerized tests have proved to be more objective and are less subjected to human bias. They are quite successful with objective tasks, such as multiple-choice questions.Computerized listening tests can serve as a diagnostic purpose to pinpoint strengths and weaknesses in the learning abilities of a student. It can be used to find whether a given student has a low or high level of listening comprehension. It can also serve as a placement test to identify a particular performance level of the student so as to place him or her at an appropriate level of instruction or online study level.The present research focuses on computer-automated scoring system for a full-text dictation test. It discusses the advantages of this system over human grading and its validity issues, reliability of the full-text dictation test and its correlation with a multiple-choice question listening test (MC) or a compound dictation test (CD).The research led to the following conclusions: a full-text dictation test is more significantly correlated with a learner’s listening comprehension or his overall language proficiency than a multiple-choice listening test or a CD test; computer-automated scoring system is capable of generating more valid, reliable and consistent scores than human graders within a much shorter time range.