Abstract
Most language proficiency tests provide very little information about the tasks that test-takers can actually perform in real life. However, if we can find methods to relate test scores to a wide range of real-life tasks, we will then be able to empirically predict how well test-takers can perform outside the domain of particular tests. There seem to be two approaches for inferring test-takers' ability to perform real-life tasks based on their test scores. One is to examine the common characteristics of the test items that they answer correctly and show how these characteristics are related to the performance of real-life tasks; the other is to relate test scores to test-takers' performance of language-related tasks outside of test situations by using questionnaires and similar measurement tools. The former approach generates what we refer to as the Proficiency Guidelines in this project; the latter approach generates the CAN-DO Statements. This paper reports on the development process we used to create an English proficiency scale based on the GTEC for STUDENTS.
