Tests of language ability are based on a certain construct that defines this ability, and this blueprint determines what it is that will be measured. The University of Pretoria has, since 2000, annually administered a test of academic language proficiency to more than 6000 first-time students. The intention of this test is to identify those who are at risk academically as a result of too low a level of academic language proficiency. If their academic literacy levels are too low, students are required to enrol for a set of four courses in order to minimise their risk of failure. The Unit for Language Skills Development at the University of Pretoria has now embarked on a project to design an alternative test to the one used initially, specifically with a view to basing it on a new construct. The reason is that the construct of the current test has become contested over the last decade as a result of its dependence on an outdated concept of language, which equates language ability with knowledge of sound, vocabulary, form, and meaning. Present-day concepts emphasise a much richer view of language competence, and their focus has, moreover, shifted from discrete language skills to the attainment of academic literacy. In this paper the abilities encompassed by this view will be discussed in order to compare the construct of the current test with the proposed construct.
Key Words: language testing; academic literacy; test constructs; accountability
Journal for Language Teaching Vol.38(1) 2004: 1-13