Psychometrics Revisited: Recapitulation of the Major Trends in TESOL

A test is a tool for making quantified value judgments and/or comparisons, and a good test is a bias-free gauge that does its value judgments and quantifications with precision. This requires that the test be at least reliable. In applied linguistics in general, and TESOL in specific, the question of test reliability has always been at the forefront of all test construction activities. As high-stakes gate-keeping tests gained more and more importance in a globalizing post-industrial world, the statistical procedures used to estimate their reliability indices, too, became more and more complex and precise. Classical Test Theory (CTT) is no longer preached, and test developers and testing agencies have resorted to Generalizability Theory (G-Theory) and Item Response Theory (IRT) as their main dishes; more recently, they have decided to spice up their activities with Differential Item Functioning (DIF). This paper seeks to provide the less-versed reader with a short and simple account of these topics. The aim of this paper is to turn the tumid prose describing complex mathematical and statistical topics in psychometrics and measurement into readable English so that students less versed in the field can make sense of them, and university professors can use the paper as a simple and informative source in their teaching activities.