Revisiting the meaning of validity for language testing: The case of two tests of English language ability

Abstract

Validity is probably the most crucial of all concepts that govern all kinds of measurement. This is more so the case in educational and psychological testing where high stakes decisions often need to be taken about individuals and institutions. From the time it saw the light of day, however, the concept of validity has been a source of inconclusive contestation. Two schools of thought have arisen from this debate. In the main, the first of these, also known as the traditional view, regards validity as a property of a test while the second, also known as the unitary view, locates it in the way test scores are interpreted and used. This creates a challenge for test developers with regard to exactly what the object of test validation should be. The aim of this article was to determine which of these views is defensible particularly for language testing. Using two studies focusing on the validity of two tests of language ability as the basis, the article demonstrates that the unitary view of validity is problematic for these tests as it leaves them susceptible to the possibility of being used for what they are not designed for.

Keywords: validity, language ability, language testing, meaning, traditional view, unitary view