A new method is proposed to set multiple standards in performance tests. The method combines three sources of information coming from three different data collections. The first is an empirical definition of mastery of an item; the second consists of parameter estimates of the items in an Item Response Theory (IRT) model, and the third source is a collection of experts' judgments on the relation between item mastery and level of performance. These judgments are given as an answer to very simple questions. The method is not iterative, and the experts are not required to judge borderline persons. The standard setting procedure is simple and can be carried out without a computer.
|Name||OMD research report|
|Publisher||University of Twente, Faculty of Educational Science and Technology|