Test Purpose and Item Type

Abstract

The recent Popham-Ebel (1978) debate on criterion versus normreferenced measurement has tended to point up some of the less clear patches in the field of measurement and evaluation. Central to many of the current problems, it seems, lie the issues of test purpose and of item type. This paper attempts to examine the two basic types of item in relation to currently emphasized core curricula, criterion versus norm referenced measurement, and formative and summative evaluation. In essence, Popham argued in the debate that most educators are primarily concerned with populations of knowledges, understandings, and skills which are susceptible to fairly precise description. In that the elements of a given population are known, a sample of them can be obtained in some systematic manner for estimating the level of a student's achievement with respect to that population. Further, as comparisons among students are largely irrelevant to the educational process, evaluation of achievement should be only by direct reference to an established criterion level of performance on the sample of items selected. Ebel presented the contrasting view that in general for most educators the populations of important knowledges and skills and their interrelationships for the courses they teach are almost infinitely great, so that these populations defy complete description both qualitatively and quantitatively. As the elements of these populations cannot be completely described, systematic sampling is not possible and, this being the case, the