Nov 11, 1992
Indicators of mathematics and science education have become increasingly important to policymakers and the public as debate about the quality of education has intensified and widespread changes in educational policy and practice have occurred. This Note discusses two fundamental aspects of the quality of educational indicators — their reliability and validity. The threats to validity and reliability to which educational indicators are particularly prone stem from the characteristics of individual indicators, the characteristics of indicator systems, and the ways in which they are used. Indicators are used to support broad inferences, and indicator systems rely extensively on simple measures and on corruptible measures. In addition, some indicators are context dependent. Finally, indicator systems often include a sparse array of measures and can be narrower than the effects of the policies they are intended to monitor. The author suggests that these problems can be addressed in constructing indicator systems, in evaluating them, and in using them. In addition, users of indicator systems must recognize their inherent limitations and should avoid using them to support unwarranted inferences.