Computerised Summative Testing: One Step Forward or Two Steps Back?

The contribution aims to initiate principled investigation into a subject of pressing concern. In recent years computerised systems of testing have become much more commonplace, with big testing institutions offering a computerised option, or even an exclusively computer-based test. The advantages of this in terms of delivery and, potentially, in terms of reliability are assumed and occasionally advertised unequivocally, while there has been little serious research into the real nature of this new test experience for the language learner and candidate. Sometimes even ‘high-stakes’ examinations are computer-based (at least as regards delivery) but the effects on candidate performance are not really understood. Is there an intrinsic difference to the computer-based English test, and is this difference linguistic, pragmatic or merely personal? Are the results liable to variation because of the instruments being used? What is the washback effect of a computer-based test as compared to traditional, paper-and-pencil English language examinations? The paper hopes to address these issues using the theoretical framework of the language testing literature and examples from current computer-based English language tests.