Setting a cut-off score for a placement test at tertiary level

Abstract

In many placement tests, cut-off scores are determined once results become available. Various methods have been used to determine these scores. They are often also established arbitrarily by, for example, taking teaching capacity into account. This article illustrates how a cut-off point can be established my means of an objective statistical method. A test of Academic Literacy, TAG, is used to illustrate the method, which is based on historical data, an adjustment of scores to conform to the same norm, and a Receiver-Operating Characteristic (ROC) curve to generate a sensitivity/specificity report. An optimal cut-off score is then fixed. We suggest that a statistical procedure such as the ROC method can play a role in setting standards and cut-off scores in a responsible and accountable manner. Keywords: academic literacy test; validation; cut-off score; test objectivity, test fairness