A Preliminary Study On Testing Techniques Of Reading Comprehension In College English Test, Band Four

For years the validity of tests with multiple-choice questions (MC) as a major testing technique has caused heated discussion in academic circles, especially when it comes to large-scale, high-stake test such as College English Test, Band Four (CET 4), whose reformation commencing from March 2004 underlined the importance of appropriate use of MC. As reading comprehension is the ability most Chinese students should acquire for the sake of their work or study, the author focused the research on reading comprehension to try to find out whether MC is a valid means to test students' reading ability. To be more specific, the author hope to search out the answers to the following question: (1) Will a test have a higher validity if we use more than one testing technique like MC to measure the reading subskills listed in the CET 4 syllabus in a single test? (2) In testing certain subskills, Will one of the testing techniques, i.e. SAQ, TR, MC and Cloze, be more effective than the others?Both qualitative and quantitative studies are employed to test face validity and constructive validity of those four testing techniques listed above. For the purpose of this, an experiment was carried out among 187 sophomore students with the help of the following instruments: four reading comprehension tests, two questionnaires for gathering information, and SPSS-11 to dig out the correlation coefficient between testing techniques. The research findings show that among the four testing techniques under discussion, SAQ and TR enjoy higher face valid and constructive validity in testing the skill of gathering specific information. The statistics analysis and subjects' responds to the questionnaires also indicate that cloze test is valid in checking students' summarizing ability provided the text is long enough; and MC is good at assessing candidates' ability to distinguish general information from the specific one, or decide the relevant importance of arguments. All these findings lead to a safe conclusion that an appropriate and valid reading comprehension test should include more than one testing technique for assessing students' comprehension abilities. To conclude the thesis, the author suggests incorporating metacognition training into reading course, and proposes further research objectives as well.