Uncertainties In Language Testing

This study sets out to explore the uncertainties in the cloze test of Guangdong MET in 2007. Specifically, the research focuses on the uncertainties as demonstrated by teachers and students in understanding, preparing, and taking the new cloze test of Guangdong MET. For this purpose, method triangulation was contrived, using think-aloud tests, classroom observation, interviews and questionnaires to collect data from 96 Senior III students and 35 Senior III English teachers in Guangdong. Three students participated the think-aloud tests, and one English class was observed. Then three students and three teachers accepted the interviews. Questionnaires were completed by 96 students and 35 teachers, including all the subjects of the previous instruments. Before the main study, a pilot study was conducted.Results show that teachers and students were comparatively certain about what to be tested in the new cloze test, but they were uncertain as to how to prepare and take the test. Teachers' and students' relative certainties about abilities to be tested in the cloze test are probably due to the information they got from the inspectors not long ago. Their uncertainties on how to prepare and take the cloze test may be attributed to the changeable, unreliable and insufficient information they got on the cloze test reform.Under uncertainty, teachers and students relied on some heuristics (Tversky & Kahneman 1974) and other uncertain strategies to help decision making. Availability heuristic is similar to the phenomenon that teachers emphasized the ability of analyzing sentence structure while students focused on memorizing new words and improving language perception. The law of small numbers, which belongs to representative heuristic, pertains to teachers' worry about the exact-word scoring method. Additionally, students made judgment by common sense because of anchoring and adjustment. Loss aversion, one factor of prospect theory, is related to the phenomenon of teachers and students reviewing the basic knowledge and doing mock exercises.Besides the irrational strategies mentioned above, other rational strategies were used by students and teachers according to the research results. It was partly because Guangdong MET is a high-stakes test in nature.This study indicates adequate information and training are necessary for Senior III teachers and students. It also suggests education policy makers be cautious about test reform and test constructors be open to the suggestions and feedback from the teachers and students, who are expected to master the orientation of the test reform and arrange their teaching and learning methods in time.