|
|
 |
|
Professor Barry O’Sullivan Plenary Abstract |
|
 |
|
|
It is now widely accepted that any claim made regarding the validity or suitability of a test for use in a particular context with particular learners must be supported by empirical evidence. However, while this concept is accepted, it is not understood in the same way by test developers and policy makers. This is at least partly due to the fact that most existing models of test validity are too vague when it comes to defining the nature and make-up of the sort of evidence required to support validity claims.
In this presentation, I will present a practical and theoretically sound model of test validation in which the test taker is clearly seen as the starting point. While presenting the model, I will demonstrate how key themes such as standards and localisation are central to the conceptualisation of validity which underpins the model presented. I will discuss the implications for test development of commonly held standards and, perhaps more critically, the need for language tests to be locally appropriate, where local is defined as the test use domain (i.e. where and what the test is used for). |
|
|
|
|
|
