Toward a framework for achievement testing

The approach to achievement testing used in the schools today is, with few exceptions, identical to that used fifty or sixty years ago—one test, given to all students, containing items taken from a bailiwick guarded carefully by the teacher or test constructor and producing scores interpretable only by him. Although this ap-proach may have been adequate in the past, such is not the case now. What we need today is not a new refinement on what we have been doing, but a new approach to assessing student achievement in an instructional program. Achievement testing is a well-worn topic in the literature, and it is understandable that the reader might wonder "What could possibly be new here? " What is new here is an explicit conceptual-ization of the relationship between achievement tests and an in-structional program. Everyone knows that an achievement test must be related to the instructional program, but the nature of this relationship has always been ambiguous or mysterious, and the resultant problems have become unbearable. So sad is the current state of affairs that we can barely distinguish between a "good " and a "bad " achievement test and must rely upon expert opinion—when the difference should be obvious to every student in the instructional program. What follows here and constitutes the substance of this article is The opinions expressed herein are those of the author and do not necessaril