Achievement Tests

Abstract

This chapter discusses the achievement tests. Achievement tests are used throughout school system around the world. They are used primarily for sorting the students and for encouraging them to study harder. Doctrinaire treatment of item writing and test construction is counterproductive. The validity of a proposed interpretation of achievement test scores rests ultimately on the quality of inference available at the item level. Aptitude and achievement were treated as interchangeable, it was widely supposed that these models would provide the most efficient and coherent means of estimating extent of achievement. Generalizability theory can work equally well or poorly with latent trait estimates which are after all just transformations of raw scores. Given the doubts about the appropriateness of latent trait models for representing achievement are considered which are barely assuaged by moving to three-parameter models it may be asked whether there is point in even discussing the possibility of integrating the two systems.