Theory-based Validity in Action

Abstract

In Part 1 we argued that approximation to the construct in a measurement instrument is essentially the result of the interactions between its context and theory-based elements. In Chapter 6 we looked at the elements of the validity framework that need to be considered under context validity. We now turn to the theory-based elements. In our present state of knowledge it is easier to treat context- and theory-based aspects of validity separately for descriptive purposes but we recognize that it is the interaction between them and the scoring criteria that lies at the heart of construct validity; see Chapter 11, Case studies 1–3 for ways we might start to address these relationships. Establishing the nature of these interactions is what will take forward our understanding of language testing and the constructs it attempts to measure (see Chaloub Deville 2003 for a programmatic discussion of the ‘abilities-in language users-in contexts’ participant metaphor perspective).