Focuses on improving the reliability of standardized tests in the evaluation of schools and programs, since student scores as now used may lead to incorrect conclusions. The authors examine common inaccuracies in reporting and propose some remedies. A major weakness is the expression of scores in such relative terms as grade equivalents and percentiles, which, given the changes in the standardizing population, precludes valid comparisons over time. Until absolute measures are available for direct inferences about school performance, the authors propose indirect approaches that take account of the limitations of present data systems. They suggest, for example, reweighting test scores at later grades before averaging to make them reflect the school population at the entering grade, and using the distribution between high- and low-performing students over time as a measure of school effectiveness. They also supply a formula that will permit measuring the achievements of individual students as evidence of the equality of opportunity afforded by the school program.
Coleman, James Samuel and Nancy L. Karweit, Measures of School Performance. Santa Monica, CA: RAND Corporation, 1970. https://www.rand.org/pubs/reports/R0488.html. Also available in print form.
Coleman, James Samuel and Nancy L. Karweit, Measures of School Performance, Santa Monica, Calif.: RAND Corporation, R-488-RC, 1970. As of January 13, 2022: https://www.rand.org/pubs/reports/R0488.html