The area of language assessment has been slow to catch up with research findings in other areas of language learning, largely because of the impracticability of applying concepts such as autonomy, collaboration and student-centered learning in the high-stakes context. Because of this, standardized tests around the world continue to use a summative, product-oriented, competitive format that has been found to be lacking in pedagogic terms. Given the importance of such tests to the lives of the language learners, it is inevitable that teaching in this situation becomes test-driven, motivation becomes extrinsic, and teachers instruct their students in „test-taking‟ strategies. Learning how to pass the test becomes more important than gaining mastery of the target language, and the development of higher-order thinking skills is sacrificed to memory-based acquisition of discrete grammatical constructions. The final „result‟ is beneficial to nobody except university administrators, who need to manage the intake of students each year (gate keeping). In view of this situation, this paper examines whether it is practical for high stakes testing to take on a different form, more closely suited to the findings of language-learning research. In particular, it investigates how current theories regarding student-centered learning might be applied to internal and external assessment in schools.
