A Comparison of Four Methods of Obtaining Discrepancy Measures Based on Observed and Predicted School System Means on Achievement Tests

Standardized tests of academic achievement have long been used as a basis for measuring the effectiveness of educational systems. The typical approach has been to compare the average achievement test score of students in a given system with some sort of national average or norm and to assume that the discrepancy between the two averages constitutes a measure of the educational effectiveness of the system. There are a number of well-known fallacies in this type of discrepancy measure as a measure of system performance, not the least of which is that it takes no account either of the conditions in which the schools of a particular system must operate nor of the achievement level at which students enter any phase of the system. It has been suggested (Dyer, 1966) that a more rational discrepancy measure of school system effectiveness might be based on the deviations between the mean achievement scores actually found at any grade level and the mean achievement scores predicted from measures of previous student achievement and the hard-to-change conditions that presumably affect the teaching-learning process. Such discrepancy