Testing tasks: issues in task design and the group oral

Abstract

This article investigates some of the issues which surround the use of tasks in oral tests, with particular reference to the group discussion. This is done from the perspective of a group of students who were asked to attempt three oral tasks. Questionnaire techniques and retrospective reports were used to collect data from the students. The principle is that test-takers have a great deal to offer to the test researcher in making judgements about the value of the tests which they take (Brown, 1993). The issues surrounding task design and use are complex, and are currently being debated not only in language-testing circles but also in the fields of second language acquisition and discourse analysis. For this reason, this article will refer to discussions in all three areas to shed light on the selection of tasks for use in oral tests. Information from the statistical analysis of tests will also be presented. All views about tests and tasks used in tests, however much some authors might eschew theory or statistical analysis (Underhill, 1987), spring from inherent theoretical positions. These positions make predictions about test scores under particular conditions, and the results of analysis enable the researcher to assess whether a view can be supported by empirical evidence. Finally, the article will look at what is possibly one of the most problematic questions in proficiency testing: the generalizability of a test score given on one task to another task or tasks. It is arguably the case that, if this is not possible, there is no justification for proficiency testing.